DG Contribution to network security_R3.pdf

Developing the P2/6 Methodology

Report Number: DG/CG/00023/REP URN 04/1065

This work was commissioned and managed by the DTI's Distributed Generation Programme in support of the Technical Steering Group (TSG) of the Distributed Generation Co-ordinating Group (DGCG). The DGCG is jointly chaired by DTI and Ofgem, and further information can

be found at www.distributed-generation.gov.uk

Contractor

UMIST

Prepared by Ron Allan, Goran Strbac, and Predrag Djapic (UMIST)

Keith Jarrett (Power Planning Associates)

The work described in this report was carried out under contract as part of the DTI New and Renewable Energy Programme, which is managed by Future Energy Solutions. The views and judgements expressed in this report are those of the contractor and do not necessarily reflect those of the DTI or Future Energy Solutions.

First Published 2004 © Crown Copyright 2004

FES Project DG/CG/00023/00/00

DEVELOPING THE P2/6 METHODOLOGY

_________________________

Final Report

29 April 2004

_____________________________

Ron Allan

Goran Strbac

Predrag Djapic

Keith Jarrett

UMIST

Final Report, 29 April 04 ________________________

CONTENTS

GLOSSARY OF ABBREVIATIONS ....................................................................................8

PREFACE.................................................................................................................................9 Terms of Reference..................................................................................................9

I. Project Status....................................................................................................9 II. Project Objectives ...........................................................................................9 III. Project Deliverables.......................................................................................9 IV. Constraints and Scope of the Project...........................................................10 V. Status of Final Report and Material ..............................................................10 VI. Presentation of Report .................................................................................10

EXECUTIVE SUMMARY ...................................................................................................11

Objectives ..............................................................................................................11 Technical Issues.....................................................................................................11 Implementation of Methodology ...........................................................................14 Computer Program Package ..................................................................................14 New Table(s) for Inclusion in P2/6 .......................................................................14

1. Introduction....................................................................................................................17

1.1. DTI Objectives of Project ........................................................................17 1.2. Objectives ................................................................................................17 1.3. Methodology............................................................................................18 1.4. Constraints and Restrictions ....................................................................19 1.5. Implementation ........................................................................................20 1.6. Specific Activities and Milestones...........................................................21

2. Technical Issues..............................................................................................................23

2.1. Generic Systems.......................................................................................23 2.2. Single Units and Materiality Issues .........................................................24

2.2.1. Specific Concerns of Single Units........................................................24 2.2.2. Consistency with P2/5 ..........................................................................25 2.2.3. Reliance on Single Units and Effect of Redundancy............................27 2.2.4. Materiality Testing................................................................................32 2.2.5. Other Considerations ............................................................................32 2.2.6. Addressing Concerns and Proposals from WS3...................................34

2.3. Materiality Considerations.......................................................................35 2.3.1. Impact in P2/5.......................................................................................35 2.3.2. Consideration of ACE Report 51..........................................................36 2.3.3. Simplification of Table A.1 of ACE Report 51....................................38 2.3.4. The Term Circuit Outage......................................................................38 2.3.5. A Way Forward ....................................................................................41

2.3.5.1. Assessment of Number of Units ....................................................42 2.3.5.2. Application of the Approach .........................................................45 2.3.5.3. Simplification of Tables ................................................................48

2.3.6. Application Examples...........................................................................49

Page 2 of 150


2.4. Disaggregation of Demand and Generation Output.................................53 2.5. De Minimus Levels of Capacity ..............................................................53 2.6. Intermittent Generation and Tm Assessment Issues .................................53

2.6.1. Background...........................................................................................53 2.6.2. Concept of Persistence and Effective Capacity ....................................53

2.6.2.1. Basic Methodology........................................................................53 2.6.2.2. Rationale for Determining Tm........................................................54 2.6.2.3. Minimum Persistence Times for Switching Actions .....................55 2.6.2.4. Minimum Persistence Times for Repair Activities .......................56 2.6.2.5. Minimum Persistence Times for Maintenance Outages................57 2.6.2.6. Summary........................................................................................57

2.6.3. Effect of Data Resolution or Granularity..............................................58 2.6.4. Correlation between Peak Demand and Generation Output.................58 2.6.5. Treatment of Multiple Wind Farms......................................................59

2.7. Ride Through Capability..........................................................................59 2.7.1. Impact of Generation Trips...................................................................59 2.7.2. Concluding Comments .........................................................................61

2.8. Risk to Loss of Supply.............................................................................61 2.9. Effect of Remote Generation ...................................................................64 2.10. Effect of Widespread Anticyclonic Weather ...........................................64 2.11. Impact of Changes in Demand Profiles ...................................................64 2.12. Synthesising Forecast Profiles .................................................................64

2.12.1. Introduction.........................................................................................64 2.12.2. Need for Forecast Profiles ..................................................................65 2.12.3. Plant Types Surveyed and Assessed...................................................65 2.12.4. Standard Profiles.................................................................................66

3. Analysis Package for Assessing Generation Capability .............................................68

3.1. Overview..................................................................................................68 3.2. Security Contribution by Generation Continuously Available................68

3.2.1. Input......................................................................................................68 3.2.1.1. Non-intermittent Generation..........................................................68 3.2.1.2. Intermittent Generation..................................................................69 3.2.1.3. Demand..........................................................................................70

3.2.2. Output ...................................................................................................71 3.2.3. Data Editing ..........................................................................................72

3.3. Contribution by Generation Not Continuously Operational....................76 3.4. Examples..................................................................................................76

3.4.1. Non-intermittent Generation.................................................................76 3.4.2. Intermittent Generation.........................................................................78 3.4.3. Non-intermittent and Intermittent Generation ......................................79

4. Sensitivity Studies ..........................................................................................................80

4.1. Introduction..............................................................................................80 4.2. Effect of Shape of Load Duration Curve .................................................80

4.2.1. Background...........................................................................................80 4.2.2. Non-intermittent Generation.................................................................80 4.2.3. Intermittent Generation.........................................................................88

4.2.3.1. Wind Farms ...................................................................................88 4.2.3.2. Small Hydro...................................................................................90

Page 3 of 150


4.3. Effect of Resolution of Wind Generation Profiles...................................92 4.4. Aggregation of Multiple Generating Groups...........................................98

5. Development of Draft of New “Table 2”....................................................................102

5.1. Possible Formats of Table 2...................................................................102 5.2. Generation Technologies and Data Considered.....................................103 5.3. Look-Up Table Approach ......................................................................105

5.3.1. Format of Base Table..........................................................................105 5.3.2. Format of New “Table 2” ...................................................................109

5.4. Graphical Approach ...............................................................................111 5.5. Computerised Package Approach ..........................................................114 5.6. Concluding Comments...........................................................................114

6. Applications and Illustrative Examples.....................................................................115

6.1. Introduction............................................................................................115 6.2. System Structure Considered.................................................................115 6.3. Application of New “Table 2” ...............................................................116

6.3.1. Single Source of Non-intermittent Generation ...................................116 6.3.2. Multiple Sources of Non-intermittent Generation..............................116

6.4. Estimating System Capability................................................................117 6.4.1. Non-intermittent Generation, Two Identical Units.............................117 6.4.2. Non-intermittent Generation, Four Identical Units ............................117 6.4.3. Intermittent Generation, Wind Farm ..................................................118

6.4.3.1. Security Contribution During Switching Process........................118 6.4.3.2. Security Contribution During Maintenance Event ......................119

6.4.4. Non-intermittent Generation, Four Non-identical Units ....................119 6.4.5. Non-intermittent and Intermittent Generation ....................................120 6.4.6. Additional Contribution from Two (<24hr) Units..............................121 6.4.7. Systems with (24hr) and (<24hr) Generation .....................................122

6.5. Impact of Materiality Criteria for Non-intermittent Generation............122 6.5.1. Materiality Test on System Studied....................................................122 6.5.2. Materiality Condition Imposed...........................................................123 6.5.3. Materiality Condition Not Imposed....................................................124

6.6. Systems Requiring Consideration of Second Circuit Outages ..............124 6.7. Times At Which Capabilities Become Available ..................................126 6.8. Concluding Comments...........................................................................127

7. Application Guide ........................................................................................................128

7.1. Introduction............................................................................................128 7.2. Approaches ............................................................................................128 7.3. Systems and Data...................................................................................128

7.3.1. Generation Capacities .........................................................................129 7.3.2. Generation Availabilities ....................................................................129 7.3.3. Generation De-minimus Criteria ........................................................130 7.3.4. Generation Winter-Time Operating Regime ......................................130 7.3.5. Intermittent Generation.......................................................................130 7.3.6. System Demand ..................................................................................131 7.3.7. Materiality Criteria .............................................................................131 7.3.8. First and Second Circuit Outages .......................................................131 7.3.9. Time Periods for Compliance Testing................................................132

Page 4 of 150


7.4. Approach to Use ....................................................................................132 7.5. Procedure Used For Approaches 1 And 2..............................................132

7.5.1. Approach 1..........................................................................................133 7.5.2. Approach 2..........................................................................................133

7.6. Procedure Used For Approach 3............................................................134

8. Conclusions...................................................................................................................145 8.1. Objectives ..............................................................................................145 8.2. Technical Issues .....................................................................................145 8.3. Implementation of Methodology ...........................................................148 8.4. Computer Program Package ..................................................................149 8.5. New Table(s) for Inclusion in P2/6........................................................149

8.5.1. Base Table ..........................................................................................149 8.5.2. Materiality Test...................................................................................150 8.5.3. New “Table 2” ....................................................................................150 8.5.4. Aggregating Units...............................................................................150

Page 5 of 150


List of Figures

Figure 1 - Systems with and without a group of identical generating units ............................23 Figure 2 - System with groups of non-identical generating units............................................24 Figure 3 - Reproduction of the results shown in Figure 16 of Report 2 ..................................26 Figure 4 – Two possible network structures ............................................................................45 Figure 5 - Group demand supplied by three generating units..................................................50 Figure 6 - Group demand supplied by two circuits and six generating units ..........................50 Figure 7 - Input data for non-intermittent generation ..............................................................69 Figure 8 – Input data for intermittent generation.....................................................................70 Figure 9 – Example of file containing intermittent generation profile ...................................70 Figure 10 - Input data for load duration curve.........................................................................71 Figure 11 – Output ...................................................................................................................72 Figure 12 - The dialog box of Contribution page ....................................................................73 Figure 13 - The dialog box of Non-intermittent Generation page ...........................................74 Figure 14 – The dialog box of Intermittent Generation page ..................................................75 Figure 15 - The dialog box of the Load page...........................................................................75 Figure 16 – Part of ShiftGeneration worksheet .......................................................................76 Figure 17 - Settings for example 1...........................................................................................77 Figure 18 - Additional settings for example 2 .........................................................................78 Figure 19 - Settings for example 3...........................................................................................79 Figure 20 – Normalised winter load duration curves...............................................................81 Figure 21 – Contribution by generating units with 90% availability.......................................82 Figure 22 – Contribution by generating units with 85% availability.......................................82 Figure 23 – Contribution by generating units with 80% availability.......................................83 Figure 24 – Contribution by generating units with 60% availability.......................................83 Figure 25 – Contribution by one generating unit.....................................................................84 Figure 26 – Contribution by two generating units ...................................................................84 Figure 27 – Contribution by three generating units .................................................................85 Figure 28 – Contribution by four generating units ..................................................................85 Figure 29 – Contribution by six generating units ....................................................................86 Figure 30 – Contribution by eight generating units .................................................................86 Figure 31 – Contribution by ten generating units ....................................................................87 Figure 32 – Contribution by wind farms..................................................................................90 Figure 33 - Contribution by small hydro plant ........................................................................92 Figure 34 – Contribution by wind farm at site A.....................................................................93 Figure 35 – Contribution by wind farm at site B.....................................................................93 Figure 36 – Contribution by wind farm at site C.....................................................................94 Figure 37 – Contribution by wind farm at site D.....................................................................94 Figure 38 – Contribution by wind farm at site E .....................................................................95 Figure 39 – Contribution by wind farm at site F .....................................................................95 Figure 40 – Contribution by wind farm at site G.....................................................................96 Figure 41 – Contribution by two wind farms (site F and site G).............................................97 Figure 42- Typical winter daily load curve............................................................................105 Figure 43 – Contribution of generation operating for less than 24h......................................106 Figure 44 – F factors in % as function of availability and number of generators..................112 Figure 45 – F factors in % as function of availability and number of generators..................113 Figure 46 – F factors in % as function of persistence Tm for wind farms ............................113 Figure 47 – F factors in % as function of persistence Tm for small hydro............................114 Figure 48 – System structure used for illustrative examples .................................................115

Page 6 of 150


Figure 49 – Modified system structure ..................................................................................125 Figure 51 – Process diagram for capping capacity of generating units .................................141 Figure 52 – Process diagram for assessing capability after first circuit outage.....................142 Figure 53 – Process diagram for assessing capability after second circuit outage ................143 Figure 54 – F factors in % as function of persistence Tm for wind farms ............................144 Figure 55 – F factors in % as function of persistence Tm for small hydro............................144

List of Tables

Table 1 – Schedule of activities ...............................................................................................21 Table 2 – Schedule of milestones ............................................................................................22 Table 3 - Percentage contribution and associated redundancies..............................................29 Table 4 - Contributions modified by level of redundancy.......................................................31 Table 5 - Comparison between contributions and capacity levels...........................................32 Table 6 – Part of Table A.1 of ACE Report 51 .......................................................................37 Table 7 - Part of Table A.2 of ACE Report 51 and Table 2 of P2/5........................................37 Table 8 - Capacity Outage Probability Tables.........................................................................43 Table 9 – Results of Example 4 ...............................................................................................52 Table 10 – Recommended values for Tm .................................................................................58 Table 11 – Proposed availability profiles for new “Table 2” of P2/6......................................67 Table 12 – Winter load duration curves...................................................................................81 Table 13 – Contribution and load factors of three individual wind farms...............................88 Table 14 - Percentage contributions by wind farms ................................................................89 Table 15 – Contribution and load factors of three individual small hydro plants ...................90 Table 16 - Percentage contributions by small hydro plant ......................................................91 Table 17 – Correction factors from 30min data for single wind farm.....................................96 Table 18 – F factors for intermittent generation for various time resolutions .........................97 Table 19 – Correction factors from 30min data for two wind farms .......................................98 Table 20 – Correction factors from 30min data for single and multiple sites .........................98 Table 21 – Aggregation of groups of units with same availabilities .....................................100 Table 22 – Aggregation of groups of units with different availabilities................................101 Table 23 – Generation technologies considered and data used .............................................104 Table 24 – Average winter load duration curve used ............................................................104 Table 25 – Format of new “Table A.1” .................................................................................107 Table 26 – “Table A.1a”, F factors in % for non-intermittent generation .............................108 Table 27 – “Table A.1b”, F factors in % for intermittent generation ....................................108 Table 28 – “Table A.1c”, Number of generators (N1) contributing to FCO..........................108 Table 29 - Format of new “Table A.2” ..................................................................................110 Table 30 – F factors in % as function of availability and number of generators ...................111 Table 31 – Number of generators (N1) contributing to FCO.................................................112 Table 32 - Recommended values for Tm................................................................................134 Table 33 - Format of new “Table A.2” ..................................................................................135 Table 34 – “Table A.1a”, F factors in % for non-intermittent generation .............................136 Table 35 - “Table A.1b”, F factors in % for intermittent generation.....................................136 Table 36 – “Table A.1c”, Number of generators (N1) contributing to FCO..........................136 Table 37 - Format of new “Table A.1” ..................................................................................137 Table 38 – F factors in % as function of availability and number of generators ...................138 Table 39 – Number of generators (N1) contributing to FCO.................................................139

Page 7 of 150


GLOSSARY OF ABBREVIATIONS

CCGT - Combined Cycle Gas Turbine CHP - Combined Heat and Power COPT - Capacity Outage Probability Table DCHP - Domestic Combined Heat and Power DG - Distributed Generation / Distributed Generator DNC - Declared Net Capability DNO - Distribution Network Operator DTI - Department of Trade and Industry EENS - Expected Energy Not Supplied EFOR - Equivalent Forced Outage Rate ENA - Electricity Networks Association ER - Engineering Recommendation FCO - First Circuit Outage GT - Gas Turbine IEEE - Institute of Electrical and Electronics Engineers IIP - Information and Incentives Project LDC - Load Duration Curve MS - Microsoft® NGT - National Grid Transco OFGEM - Office for Gas and Electricity Markets SCO - Second Circuit Outage TSG - Technical Steering Group VBA - Visual Basic for Applications® WS3 - Workstream 3

Page 8 of 150


PREFACE

Terms of Reference

I. Project Status This report describes studies that are an extension of previous UMIST projects

funded by FES/ETSU, namely: K/EL/00287: “Network Security Standards with Increasing Levels of Embedded Generation”, by R.N.Allan and G.Strbac of UMIST. Final Report dated 10 August 2002. This report considered security and associated standards in considerable detail as its scope and objectives were much more wide-ranging than the subsequent projects

•

•

• •

•

•

K/EL/00287 Extension: “Security Contribution from Distributed Generation”, by R.N.Allan and G.Strbac of UMIST, and K.Jarrett of PPA. Final Report dated 11 December 2002. This report developed the methodology for determining the capability contribution that generation could make to security using the concepts and criteria underpinning the existing Engineering Recommendation P2/5. This methodology is used in the present project.

II. Project Objectives The main objective of this project was to use the methodology developed in the

previous FES project (K/EL/00287 Extension) to assess the security capability of modern distributed generation in order to:

review Table 2 and related text of Engineering Recommendation P2/5 propose information and results that could be used to create a new P2/6 that takes into account: modern type of generating units unit numbers unit availabilities capacities

III. Project Deliverables The agreed deliverables of this project were:

a progress report outlining the initial progress in the project. This was issued on 8 December 2003 prototype software using the agreed methodology that could be used by relevant parties for evaluating the capability of specific and special generation systems. The first version was issued on 8 December 2003, with revisions send subsequently. A final version is issued with this report on 1 April 2004

Page 9 of 150


an interim report containing preliminary discussion and resolution of several related issues of particular concern to Workstream 3. This was issued on 27 January 2004

•

• a final report containing: the definitive outcome of the discussion and resolution of the issues of

particular concern to Workstream 3 the proposals for the development of a new security standard P2/6

including a proposed new “Table 2” and supporting text examples illustrating the approach and application this forms the present report

IV. Constraints and Scope of the Project

The constraints were set by Workstream 3 of the DTI/Ofgem Distributed Generation Co-ordinating Group and its Technical Steering Group, and reflected its objectives and timescales. Consequently, the main constraint consisted of:-

Using the methodology developed in the previous project funded by FES (K/EL/00287 Extension) because this satisfied the sub-constraints of:-

o the methodology permits simple and straightforward extensions to Table 2 o the approach is consistent with the concepts and analysis underpinning the

existing P2/5 o the results are implementable in the short term

V. Status of Final Report and Material This report is the Final Report of this project and therefore all the content, including

concepts and ideas, results, discussions and conclusions are the definitive findings of the studies.

VI. Presentation of Report This report was commissioned by FES and it is to this organisation that the report

has been officially submitted. However Workstream 3 of the DTI/Ofgem Distributed Generation Co-ordinating Group and its Technical Steering Group has overseen the activity. Therefore the report has also been submitted to this Workstream, and it is understood that it will provide a major input to the subsequent activities of this and other workstreams.

Page 10 of 150

EXECUTIVE SUMMARY

Objectives As part of its wide brief on the impact of distributed and modern forms of

generation, the DTI1 has mandated Workstream 3 to focus on several short-term network solutions, one of which relates to the immediate problem of how best to assess the contribution of distributed generation to network security. This project formed part of these developments and followed other related projects conducted by UMIST.

The main objective of this project was to use the methodology developed in the

previous Methodology project2 (also funded by FES) to assess the security capability of modern distributed generation in order to review Table 2 and related text of Engineering Recommendation P2/5, and to propose information and results that could be used to create a new P2/6 that takes into account: modern type of generating units; unit numbers; unit availabilities; and capacities.

The specification was to apply this methodology in a way that would reflect the

attributes of present-day generation but constrained in two very specific respects. The application had to be simple, easy to implement and achievable in the short term, and had to be consistent with that used to develop the generation contributions specified in the present P2/5. In addition a number of related technical issues had to be resolved. Consequently this final report contains the definitive outcome of the discussion and resolution of these issues, together with the proposals for developing a new security standard P2/6, a proposed new “Table 2” and supporting text, an application guide, examples illustrating the approach and application, and a description of the associated analysis package.

Technical Issues

During the previous Methodology project, a number of issues were raised but, as they were not considered to be material at that time, were postponed until this project. These are summarised in the following paragraphs.

(i) Treatment of single unit generation systems During the previous project, one issue raised was that a single unit should not be

relied on to provide security. A significant part of this project, and of the report itself, was spent on addressing this issue. From the analytical assessment, it was agreed that single units do have a security capability and should be given credit for doing so. It was also agreed that

1 In order to implement the recommendations of the DTI/Ofgem, a Distributed Generation Co-ordinating Group together with a supporting Technical Steering Group (TSG) has been established. A number of workstreams are being pursued; Workstream 3 being focussed on short-term network solutions. 2 “Security Contribution from Distributed Generation”, ETSU/FES Project K/EL/00287 Extension. Final Report by R.N.Allan, G.Strbac and K.Jarrett, UMIST, 11 December 2002


the methodology quantifies this capability in a way that is compatible with the consideration of all other units and with the philosophy of the existing P2.6. Consequently, single units are included as entities in the development of the proposals for a new “Table 2”.

(ii) Effect of shape of load duration curves (LDC)

During the Methodology project, it was observed that the shape of the LDC affected

the capability associated with the generation. It was decided to leave any further study until the present project in order to perform the sensitivity studies using real data for the generating plant and for the LDC. Nine typical LDCs were obtained by the Data Collection and Processing project3. These individual LDCs have been used separately in these sensitivity studies. It was observed that the shape of the LDC affects the contribution delivered by generation; the effect being greater with increasing unit availability up to about 90% and with decreasing number of units. Although this may seem to be an effect that should be included as a parameter in the development of P2/6, it should be noted that the variation between minimum and maximum is, in most cases, of the same order or less than the variation in contribution caused by practical ranges of unit availability. It can therefore be recommended that an average LDC should be used as well as an average value of unit availability in the development of a “Table 2” similar to that of Table 2 in P2/5. If this assumption is considered unacceptable in specific situations, then such situations should be considered special cases and assessed individually using the analysis-package approach. The average LDC obtained from these nine typical LDCs was used to obtain the proposed values for P2/6.

(iii) Persistence of intermittent generation, Tm

For generation to provide security, the output must remain at or above a certain required level for a minimum period of time, defined as Tm. This is generally only a problem with intermittent generation such as wind due to its significant variability. This persistence time has a considerable impact on the capability that can be associated with intermittent generation and is related to the duration of the system conditions for which such generation may be able to avoid or reduce customer disconnections. There are three distinct system conditions, each of which can be associated with different minimum persistence times. These are switching activities; repair activities; and maintenance activities. Proposals for required values of Tm, depending on system conditions, are given in the report.

(iv) Time resolution of intermittent generation output profiles

All intermittent generation output profiles are obtained by averaging over a specific

period, frequently ½ hr periods. One issue was whether this was acceptable or whether resolution times of 5 min or even 1 min would be necessary. Sensitivity studies showed that the security capability of intermittent generation decreased as the resolution time decreased, as expected. However, since the thermal time constant of overhead lines tends to be the most critical and these are generally of the order of several minutes, it is recommended that a 5 min resolution for quantifying the contribution of wind generation should be used. This value is used in the derivation of the “new Table 2”. In other cases, for instance with circuits composed of cables and/or transformers, it may be sufficient to use a 30 min resolution. Appropriate correction factors are proposed and included in the report.

3 “Data Collection for Revision of Engineering Recommendation P2/5”, FES Project K/EL/00303/05. Final report by K.Jarrett, PPA, January 2004

Page 12 of 150


(v) Ride-through capability Ride-through capability is of great importance in knowing how, or even whether,

distributed generation can contribute to system security. This capability does not affect the effective generation contribution of a particular generation site to be determined by the methodology and to be specified in a table similar to Table 2 of P2/5. What it does affect is whether this contribution is available immediately, after a certain time period, or not at all. This depends on whether the generation plant is, or is not, tripped following a circuit outage; and if tripped, how long it takes to restore its network connection. These scenarios in no way affect the methodology, nor the effective generation contributions that the methodology determines. It is only the way that these contributions are used that is dependent on the scenario. In some cases, the contribution becomes available immediately (similar to base loaded stations in the current P2/5) and sometimes the contribution only becomes available after a period.

(vi) Risk to loss of supply There is a view that the risk to supply (presumably meaning the impact on customer

reliability) could be degraded by using generation instead of circuits: this belief may exist because of the view that the reliability of generating units is less than that of overhead lines and underground cables. However the benefit of reinforcing a system using generation is clearly an alternative to constructing a new circuit or upgrading an existing one. Reinforcing with generation would undoubtedly increase reliability although perhaps by less than that due to an additional circuit. The discussion in the report identified several points. Firstly, a revised P2/5 and Table 2, i.e. a new P2/6, should not cause customer reliability to be less than that envisaged by P2/5, and therefore a distribution network operator’s (DNO) licence. Secondly, if generation exists but is ignored in determining customer reliability levels, then the actual customer reliability may be in excess, even considerably in excess, of the minimum set by P2/5. Finally, if generation exists in a network, either its contribution to security should be included to maintain consistent treatment of all customers throughout the system, or the principles and criteria used in P2/5 should be completely changed. The latter is not the objective of the present review of P2/5, which is solely intended to be an update of Table 2 of P2/5, all principles and criteria remaining otherwise unchanged. A more fundamental review of P2/5 is possible and suitable approaches have been published previously4 by two of the authors of this report.

(vii) Other issues Four other issues have also been considered in parallel with this project. These were:

disaggregation of demand and generation, de-minimus levels of capacity; effect of remote generation; and effect of widespread anti-cyclonic weather. Although part of the original specification of this project, these were transferred from this project and became the direct responsibility of Workstream 3 (WS3). However, relevant sections identified with these items are retained in this report for completeness, but no details are included.

4 “Network Security Standards with Increasing Levels of Embedded Generation”. ETSU Project K/EL/00287. Final Report by R.N.Allan and G.Strbac, UMIST, 10 August 2002.

Page 13 of 150


Implementation of Methodology There are three main ways of implementing the methodology. This report

recommends all three approaches; the one to be used depending upon circumstances, the data available and the need for precision in the output. The three approaches are:-

(i) look-up table(s) approach - this has the merits of simplicity and should be free of erroneous applications. However, it is restricted to specific generation types and typical or average availability data. In the case of non-identical units, the capabilities are approximate since generation contributions are assessed separately for each type and unit size and then added arithmetically. This gives a pessimistic result, but generally of acceptable precision. This approach can also incorporate the concept of capping of the security contribution of generation so that it does not violate the materiality conditions which require generation to be less significant than the circuits

(ii) graphical approach - this is an extension of the previous approach and is still reasonably simple. It accommodates a wider range of data and choices, but still requires approximate summations for non-identical units

(iii) computer program approach - this permits an unrestricted choice of data and mixed generation types. The prototype analysis package is implemented in MS Excel using VBA environment and a User Guide is provided for its easy application. The package calculates the security contributions only and these need to be assessed for compliance in the same way as performed with either of the two previous approaches.

An Application Guide has been written and included in this final report.

Computer Program Package One requirement of the project was to develop a software package that could

perform the required assessments using the agreed methodology. This was an essential step in order to analyse many generation scenarios and to perform a wide range of sensitivity studies. It was agreed that the package need only be of a prototype form, not of a commercial grade and not commercially supported by UMIST. Instead, the specification was to produce a simple spreadsheet application package that an engineer with a reasonably knowledge of the approach could use. This prototype package has been developed in conjunction with Workstream 3. Its objective is to calculate the capability contribution to security of supply from distributed generation connected to a particular demand group. The application has been developed using Microsoft Excel® and Visual Basic for Applications®.

New Table(s) for Inclusion in P2/6

(i) Base Table

The data used to create the “new” tables for inclusion in P2/6 were established from the Data Collection and Processing project5, also funded by FES. This separate project proposed availabilities to be used and these were agreed by WS3.


Page 14 of 150


Table 2 of P2/5 is a derived table. The base table is Table A.1 of ACE Report 51, although this may not be evident from P2/5 itself. Therefore the first requirement is to establish the base table. The types of generation identified by the Data Collection and Processing project and included specifically in the tables are: landfill gas; CCGT; CHP (sewerage using spark ignition, sewerage using gas turbines, and others); waste to energy; wind; and small hydro. These tables are easy to use since they only need knowledge of generation type and number of units in the case of non-intermittent generation, and generation type and the degree of required persistence (Tm) in the case of intermittent generation. If the data underpinning the security capability used for these types of generation plant or other types of generation are being considered, then a more general set of tables and figures are also included.

(ii) Materiality Test

According to P2/5, generation should not be dominant and consequently materiality

criteria were introduced to prevent this happening. A revised set of criteria are proposed that are compatible with the methodology and P2/5. These are:

The contribution of generation specified in Table 2 (of P2/6) is based on the assumptions that:-

(a) the cyclic rating of the largest distribution circuit is greater than F% of the total sent-out capacity of the N1 largest generating units and

(b) the cyclic rating of the two largest distribution circuits is greater than F% of the total sent-out capacity of the N2 largest generating units

(iii) New “Table 2” In the application of this revised set of criteria, the F factors are those specified in

“Tables A.1, A.1a and A.1b” and N1 and N2 largest generating units correspond to the number of generating units to remove for a FCO (N1) and a SCO (N2) given in “Table A.1c”. On this basis, the base table of “Table A.1” simplifies to that “Table A.2”, the equivalent of “Table 2” in P2/5. If the materiality test is not satisfied, then the generation would become the most significant FCO. If this is unacceptable for technical and/or commercial reasons, then P2/5 states that separate risk and economic studies may need to be undertaken. However, a more practical alternative would be to use a process of “capping” the generation, as proposed in the report. This can be achieved by evaluating the maximum capacity used to assess the security contribution of each generating unit such that the materiality criterion is satisfied.

(iv) Aggregating Units If more than one type of generation exists in a demand group then the capabilities

must be combined. The exact way is to aggregate them using the methodology. Alternatively the individual capabilities could be arithmetically summated. The summation results are always less than those given by the aggregation approach and therefore its use would always give results that err on the side of caution. In addition, the difference between the two values is generally small particularly as the number of units is increased. Since the aggregation approach increases the degree of precision and not necessarily the degree of accuracy, these results indicate that, provided the number of separate groups is relatively small, the simplicity of the summation approach justifies its use for evaluating the security capability of groups having non-identical units. It is therefore suggested that the summation approach be used for

Page 15 of 150


most case studies and that the aggregation approach be used only for those cases in which greater precision is deemed necessary, or the number and diversity of the groups give cause for concern about whether the summation approach is sufficiently precise.

Page 16 of 150

1. Introduction

1.1. DTI Objectives of Project As part of its wide brief on the impact of distributed and modern forms of

generation, the DTI6 has mandated Workstream 3 (WS3) to focus on several short-term network solutions, one of which relates to the immediate problem of how best to assess the contribution to network security from distributed generation. Further work on longer-term reviews of this issue is also in hand.

It was decided by WS3 that this shorter-term work assessment should proceed in a

series of stages: (a) the development of a methodology for assessing the security-support

capability of modern generation. UMIST, in collaboration with PPA, was contracted to do this development. The Terms of Reference were agreed in July 2002 and the Final Report submitted in December 2002. [“Security Contribution from Distributed Generation”, ETSU/FES Project K/EL/00287 Extension. Final Report by R.N.Allan, G.Strbac and K.Jarrett, UMIST, 11 December 2002]. (This is referred to as the “Methodology Project” or “Methodology Report” in this report).

(b) the collection, collation and processing of real system data needed for input into the developed methodology. PPA, in collaboration with UMIST, was contracted to do conduct this project. The Terms of Reference were agreed in June 2003 and the Final Report has been submitted. [“Data Collection for Revision of Engineering Recommendation P2/5”, FES Project K/EL/00303/05. Final report by K.Jarrett, PPA, January 2004]. (This is referred to as the “Data Collection and Processing Project” or “Data Collection and Processing Report” in this report).

(c) the implementation of the methodology developed in stage (a) using the data obtained in stage (b), to provide the necessary results needed to update the existing P2/5 into a new P2/6. UMIST, in collaboration with PPA, was contracted to do this development. The Terms of Reference were agreed in October 2003 with the Final Report due in March 2004.

The present report describes the outcome of this third stage.

1.2. Objectives This project was conducted in order to use the methodology developed previously to

assess the capability contribution of distributed generation to security of supply. The present standard P2/5 for assessing this security was written in the 1970s and clearly does not reflect 6 In order to implement the recommendations of the DTI/Ofgem, a Distributed Generation Co-ordinating Group together with a supporting Technical Steering Group (TSG) has been established. A number of workstreams are being pursued; Workstream 3 being focussed on short-term network solutions.


present-day generating units, their mode of operation nor the load profile. Therefore the specification was to apply this methodology in a way that would reflect the attributes of present-day generation but constrained in two very specific respects:-

(i) firstly the application had to be simple, easy to implement and achievable in the short term

(ii) secondly the application had to be consistent with that used to develop the generation contributions specified in the present P2/5.

1.3. Methodology The agreed methodology7 (see Methodology Report) determines the capacity of a

perfect circuit which, when substituted for the distributed generation, gives the same level of expected energy not supplied (EENS). This capacity is the effective contribution of the generation system. This approach is identical in concept with that used in developing the present P2/5, a conclusion confirmed by the results given in the previous report, which reproduce the 67% value specified in Table 2 of P2/5.

It should be noted that the operational characteristics and operational regimes of

modern generating plant differ significantly from those that existed in the 1970s. P2/5 assumes that embedded generating plant will be available when required subject only to plant unavailability and shifting regimes. The latter can be neglected with modern plant and the former is generally construed to mean the relevant plant may not be available due to plant failures and forced outages. It would be convenient if this was the current situation, but there are now other considerations that need to be taken into account. To derive energy output from a generator, the following conditions are required:

•

•

•

the generator must be in working state, i.e. it must not have failed. This aspect reflects the technical up and down states of the generating plant there must be a source of primary energy, e.g. gas for GTs, wind for wind generators, etc. If the primary source of energy to a generator is unrestricted then consideration of this source can be neglected. However if there are restrictions or the source is intermittent, then this may need to be considered. The most important area of concern is with wind plant and the intermittency of the wind regime. This would not matter if the wind and load were perfectly correlated because the variation in wind power would mirror the variation in load. However this is unlikely to be the case. it must be commercially advantageous to the generator owner/operator to run the plant. Present-day generating plant is privately owned and therefore its use for network support may be restricted for commercial reasons. Unless the “stick” approach is used where a generator is instructed to operate in emergencies, the “carrot” approach must be used, i.e. the generator must be given financial incentives to be available and to deliver when needed.

These conditions all contribute to the overall availability of a plant and are best

subdivided as:-

7 For details of this methodology, see “Security Contribution from Distributed Generation”, ETSU/FES Project K/EL/00287 Extension. Final Report by R.N.Allan, G.Strbac and K.Jarrett, UMIST, 11 December 2002

Page 18 of 150


• • •

• •

•

•

technical availability: relates to whether the plant is in a working state energy availability: relates to whether primary energy is available commercial availability: relates to whether it is commercially available

The methodology permits an extensive set of plant and system attributes to be

considered, including the above availability parameters, and reflects modern types of generating units and operational modes including conventional, CHP and renewable energy units. Specifically the methodology permits the following attributes to be assessed:-

unit attributes: number of units, capacity of units, technology of units system attributes: peak load, load profile, multiple generation sites, remote location of generation sites, units not available for 24hr in a day availability attributes: technical availability which relates to whether the plant is in a working state, i.e. it must not have failed: energy availability which relates to whether energy is available to drive the units: commercial availability which relates to whether it is commercially available materiality attributes: the methodology is applicable to all generation sites irrespective of number of units and their capacity.

1.4. Constraints and Restrictions The previous and present projects were subject to several specified constraints. The

most significant was the need to be consistent with the existing P2/5. This restricts the methodology to comparing the generation with the effective capacity of a perfect circuit and to use EENS as the reliability criterion. There are alternative approaches and alternative reliability measures against which the generation could be compared. These aspects have been discussed in a previous report8 written by two of the present authors.

In addition there are several aspects relating to constraints, restrictions and

applications associated with the output of the methodology. Some were addressed in the previous report but others were outside of the Terms of Reference and therefore the scope. These include the following considerations, most of which have been dealt with in the present project.

The values given by the methodology are similar in concept to the 67% value quoted

in P2/5. This value is essentially an average value representing the average behaviour of the generating system. In deciding whether a system complies with P2/5, this value is treated in a deterministic sense, i.e. effective capacities are summated and compared with the requirements specified in P2/5. There is, therefore, an implicit assumption that this level of capacity is available at all times of need. It must be recognised that the actual contribution can be greater or less than this assessed level and therefore P2/5 itself can not, and does not, ensure that a capability is deliverable at the time of need. It can also be recognised that this variability is generally greater with generating units than circuits, and may be greater with a small number of units than a large number of units. For this reason, one school of thought has suggested that sites with a small number of units should be treated differently. However it


Page 19 of 150


must be recognised that the approach underpinning the methodology treats all units irrespective of number and capacity in an absolutely objective manner. This is completely consistent with the concepts of P2/5, and permits the actual effective contribution to be calculated, unlike the present P2/5, which specifies a single value of 67% contribution for all unit sizes and numbers. This aspect was not dealt with in stage (a) of this sequence of projects because it was out of scope. However it remained an outstanding issue and consequently has formed a significant part of the present project under the general heading of “materiality”.

The methodology does not evaluate directly a level of risk as would be experienced

by customers. Instead it establishes a proxy to this by evaluating a capability level, which is perceived to be sufficient to minimise the duration of interruptions if they occur. Indeed this is the principle and philosophy of the present P2/5. It should be noted that the inherent risk is unaffected by the methodology. Therefore, given that EENS is the criterion for assessing the contribution of generation to network security, the inherent risk to loss of supply will be no greater than that assessed by the present P2/5. It is probably worth noting however that, if sections of the system, including generation and/or other transfer capacity, are ignored in determining whether the system is P2/5 compliant, then the actual capability of the system would be greater and in excess of P2/5 requirements, and the inherent risk would be lower. This is a consequence of the assessment procedure, not the methodology. This aspect has also been considered in more detail in this project.

In any practical situation, the protection and stability of the generation would need to

be taken into account. This was again outside the scope of the previous project. However, even if the generation is tripped following a fault, the developed methodology is still applicable for quantifying the security contribution made by that distributed generation. This may not be available instantaneously because of the time to restore the generation but could still be a contributing factor after a short period of time, such as 1min, 15min, 3hr etc. This is again consistent with the current P2/5, which permits generation to be considered in this way. This aspect has also been addressed in this project.

1.5. Implementation There are three main ways of implementing the methodology. The first option is a

look-up table in the form of the current Table 2 of P2/5. This would retain the simplistic and practical merits of the P2/5 approach, but it is likely to be slightly more complex and extensive in its application than the present table. The second option is based on families of graphs and/or figures. Here a larger range of system design parameters can be factored into the graphs to reduce the implicit approximations of the tabular approach. The third option is a computerised approach based on a spreadsheet environment. Each situation is then the subject of an individual assessment, but using a standardised approach to ensure equality of treatment whilst recognising many local or site-specific parameters. It is only this approach that can accurately assess all specific attributes pertaining to specific situations including the ability to assess different generation technologies on the same site and multiple generation sites feeding the same load group. No firm decision about which option to choose was made in the previous project, but clearly required a resolution during the present project. As described later, this report recommends all three approaches; the one to be used depending upon circumstances, the data available and the need for precision in the output.

Page 20 of 150


1.6. Specific Activities and Milestones The Terms of Reference of this project identified a number of activities and three

milestones. The scheduled activities included consideration of several technical issues as well as evaluating the security capabilities of different generation types and are shown in Table 1. The three milestones are shown in Table 2.

Table 1 – Schedule of activities

project activity number

activity

report section number

1

Application of the data to resolve the “Issues”

1.1 development of generic systems 2.1 1.2 development and testing of the analysis package 3 1.3 sensitivity studies 4 1.4 treatment of single unit and single protection point generation systems 2.2/2.3

1.5(WS3) disaggregation of demand and generation output 2.4 1.6(WS3) de minimus levels of capacity and contribution 2.5

1.7 principles for the calculation of Tm and sensitivity assessments 2.6 1.8 ride through capability and contribution 2.7 1.9 risk to loss of supply 2.8

1.10(WS3) effect of remote generation 2.9 1.11(WS3) effect of widespread anti-cyclonic weather 2.10

1.12 impact of changes in demand profiles 2.11 1.13 method for synthesising forecast profiles and impact of energy market 2.12

2

Development of a draft New “Table 2” and application guide

2.1 initial discussion and likely format 5.1 2.2 treatment of non intermittent DG 5.1-5.6 2.3 treatment of intermittent DG 5.1-5.6 2.4 further development of analysis package (from 1.2) 3 2.5 treatment in Table 2 of P2/5 5.1-5.6 2.6 application guide and process diagram 7 3

Application testing – practicability and sanity check

3.1 development of examples for sanity checks 6 3.2 carry out sanity checks and review the application process 6 4

Presentations and reporting

4.1 presentations n/a 4.2 reporting n/a 5 WS3 review n/a 6 Final report n/a 7 Seminar n/a 8 Consultation n/a 9 Final report and WS3 summary paper n/a

Note: the four items labelled (WS3) were transferred from this project and became the direct responsibility of WS3. Sections identified with these items are retained in this report for completeness but no details are included.

Page 21 of 150


Table 2 – Schedule of milestones

No.

Milestone deliverable

Milestone date

Issued date

Progress report

8 December 2003

8 December 2003

1

Analysis package

8 December 2003

version 1: 8 December 2003

version 2: 1 April 2004

2

Interim report

3 February 2004

27 January 2004

Final report – draft

19 March 2004

1 April 2004

3

Final report

30 April 2004

29 April 2004

Page 22 of 150


2. Technical Issues

2.1. Generic Systems There is a need for a number of generic systems. However it must be noted that the

purpose of this project is not to perform reliability assessments on transmission or distribution networks; this would be completely outside the scope: the objective is to perform studies similar in concept to those included in ACE Report 51. Essentially this means that studies will be centred on various group demands with one or two infeed circuits together with appropriate generation systems. These should be sufficient to represent the different group demands specified in P2/5. It is therefore best to describe them as systems rather than networks because the intra-networks within each group demand will not be represented, only the infeeds to these groups. This leads to the need to consider systems that have structures similar to those shown in Figures 1 and 2.

(a) (b) (c) (d) (e) (f) (g) (h)

Figure 1 - Systems with and without a group of identical generating units

Page 23 of 150


Figure 2 - System with groups of non-identical generating units The generating units included in the systems shown in Figure 1 are assumed to be a

group of identical units, including as separate cases for all non-intermittent units and for all intermittent units. The arrangement in Figure 1(b) provides no additional security over that in Figure 1(a) if it is assumed that generation cannot be operated in an islanded system. This is the normal practice at present but may change in the future. If that happens, then the arrangement in Figure 1(b) could provide additional security.

The generating units included in the system shown in Figure 2 are assumed to be

groups with different technologies, different availabilities, different capacities and/or different numbers, and the combination of non-intermittent, intermittent and non-intermittent/intermittent units. This therefore can demonstrate the process of combining different generating groups together.

These systems have been used to demonstrate the procedure for determining the

capability contribution of generating systems and to illustrate the type of values that could be found. The data was provided by the companion “Data Collection and Processing” project.

2.2. Single Units and Materiality Issues

2.2.1. Specific Concerns of Single Units During the methodology development stage (see Methodology Report9),

considerable discussion took place regarding the treatment of single unit generating systems. The concerns raised within Workstream 3 at that time, and confirmed at the commencement of this project, centred on three specific aspects:-

i) the need to be consistent with the assumptions underpinning the development

of P2/5 ii) the problem of relying on a single unit which, if failed, means that it provides

no security, i.e. there is no apparent redundancy iii) the difficulty of accepting that a single unit, or even a small number of units,

should be a dominating factor in supplying security. This has been defined previously as “materiality”.

Page 24 of 150

9 “Security Contribution from Distributed Generation”, ETSU/FES Project K/EL/00287 Extension. Final Report

by R.N.Allan, G.Strbac and K.Jarrett, UMIST, 11 December 2002


It follows that all these concerns needed to be addressed and resolved. At that time, it was decided to treat them identically to multi-unit systems and to leave any special treatment to the present project. This was sensible since it gave time for serious consultations between the various parties and allowed further quantitative assessments to be made.

The Methodology Report stated that the methodology could be applied consistently

to any number and any size of units. The consultation paper issued by Workstream 3 subsequently proposed that:-

a) a single set’s contribution will be recognised as a special sub-grouping by

whatever format of P2/6 is agreed upon b) the contribution level is likely to be lower than that from multiple sets of the

same aggregate capacity and that this will be specifically computed c) materiality rules, similar to those in paragraph 3.6 of P2/5 will be used, and

these will be formulated within Phase 2 (present project) after sample calculations have been executed. The methodology for formulating these rules will follow the logic outlined in ACE 51.

These concerns are addressed in the following sections.

2.2.2. Consistency with P2/5 The constraints set by Workstream 3 were threefold (see Preface and Introduction).

One of these was that the approach should be consistent with the concepts and analysis seemingly underpinning the existing P2/5. Therefore Concern (i) stated in Section 2.2.1 is of significant importance and must be satisfied if at all possible.

It should be noted that many of the concepts and much of the philosophy

underpinning the development of P2/5 are unavailable and lost in the passage of time. Additionally, the associated ACE Report 51 is not helpful in many respects. It should also be noted that any concept and philosophy that is shown to be flawed or inappropriate needs careful consideration before being carried forward into P2/6.

P2/5 does not specifically state that Table 2 of P2/5 is inapplicable to single unit

systems; it is perhaps implicit in paragraph 3.6 of the standard. In the materiality rules, it only mentions the removal of two and three “large” units. The implication is that Table 2 of P2/5 may not apply to single unit systems. However the following paragraph states that, “where these assumptions do not apply, detailed risk and economic studies may need to be undertaken, particularly if there are only one or two large generators in a group”, i.e. it does not insist on studies and also does not state studies are required if the units are small.

At this point it is worth revisiting some of the results presented in the Methodology

Report, in particular those shown in Figure 16 of that report. This is reproduced as Figure 3 in this report.

Page 25 of 150


relate made singleused, unit nu

reasonunits Howevalue form o

units possibsubsta

10 it shoor are bProcessTherefoThis do

85%

Figure 3 - Reproduction of the results shown in Figure 16 of Report 2 The results10 shown in Figure 3 are for the systems described in ACE Report 51 and

to units with availabilities of 86%. The results clearly indicate that the contribution by each system is very dependent on the number of units, varying between 50% for a unit to about 76% for 10 units. This implies that, unless significant bandwidths are each ‘number of units’ should form its own sub-group. If this were the case, then all mbers would be dealt with in a consistent manner.

If banding of unit numbers is considered desirable, then possible sub-groups with

able bandwidths for the results shown in Figure 3 are 1 unit, 2 units, (3-5) units, (6-8) and (10+) units. This concept conforms with WS3 proposal (a) in Section 2.2.1. ver although the basis of proposal (b) is correct, it may not be necessary to compute the for each specific case but simply specify a contribution for single units either in tabular r graphical form, in the same way as for all other unit numbers.

There are no explicit statements in P2/5 or ACE Report 51 to indicate why single

are not specifically mentioned or why materiality constraints were introduced. One le reason is that, as shown in Figure 3, the contribution made by a single unit is ntially less than for other units, particularly four or more. Therefore, since the

uld be noted that the results in this section (Section 2) are either extracted from the Methodology Report ased on the data used in that report, e.g. the LDC from ACE Report 51. The Data Collection and ing Report proposes new data, and the latter are used for the development of the new “Table 2”. re there are some differences between the results shown in Section 2 and latter sections of this report. es not affect any discussion or conclusions.

45%

50%

55%

60%

65%

70%

75%

80%

1 2 3 4 5 6 7 8 9 10

Number of generators

Con

trib

utio

n

Page 26 of 150


requirement was to have a single percentage contribution in Table 2 of P2/5 (i.e. 67%), it would have been inappropriate to have included single units, which Figure 3 suggests should have no more than a 50% contribution. This idea can easily be extended to consideration of two and three units. These are also seen from Figure 3 to contribute less than 67%. Again it would have been inappropriate for these to have been allocated the 67% level if they were dominant in the system. Therefore a materiality test as stated in paragraph 3.6 of P2/5 could be justified. None of these problems arise for unit numbers greater than three since all have contributions greater than 67% in Figure 3. If this philosophy is correct, then the objectivity of restricting the contributions of two and three unit systems becomes apparent. Such restrictions would not be needed however if one, two and three units were treated separately and their proper contributions were specified. This is one of the objectives of P2/6; so to follow the constraints apparently embedded in P2/5 would then not be justified objectively: even if it appears to maintain consistency, it would not actually be doing so.

2.2.3. Reliance on Single Units and Effect of Redundancy Intuitively, there can be a feeling that one should not rely on a single unit to provide

safety or security in any engineering application, which then creates Concern (ii) in Section 2.2.1. The reasoning is that, if anything goes wrong with that single unit, then complete breakdown occurs. It must be stressed that this is a subjective judgement which may or may not be correct. It is best visualised with examples:-

• if one was choosing an aircraft to fly in, there is a temptation to choose a

four-engined plane rather than a single-engined one. However if each uses engines having the same reliability and both require all engines to work, then there are more things to go wrong with the four-engined plane and the choice should be the single-engined one11. For example, if the engine availability is 0.95, the availability of the single-engined plane is 0.95, but that of the four-engined plane is only 0.81! In this case both of the systems have zero redundancy, and it is a fact that the risk increases with the number of units if the level of redundancy remains the same. This effect is clearly demonstrated in Table 3 which follows later in this section

• even if the degree of redundancy increases, this does not mean the risk will reduce significantly. Consider the same situation as above but the four-engined plane requires only three or more engines to work. In this case the availability increases to 0.986, still only slightly greater than the single-engined plane

• the situation however becomes worse if, say, the availability of each engine of the four-engined plane is only 0.9 and that of the single-engined plane is 0.99. In this case the availability decreases to 0.66 if all four engines are required, and increases to only 0.95 if three or more are required, significantly less than that of the single-engined plane in both cases.

The important point to be derived from these considerations is that it is not simply

the single parameter of number of units that ranks the risk order, it is a combination of several parameters including the number of units and their availability. For this reason, a small

11 In practice, aircraft do not usually require all engines for system success. However, the example as stated is correct.

Page 27 of 150


number of units should not be neglected without evaluating their effect and comparing them with systems having a greater number of units.

All of the above considerations should be put into context of the objectives of the

review and revision of P2/5. One of these is to take into account modern generation technologies. These new technologies are associated with a wide variety of operating characteristics and availabilities, from modern GTs to wind farms. The original P2/5 considered only a narrow range of generating units with similar characteristics. In such situations it was probably reasonable to assume that a single unit system would be less reliable than one with several units, as illustrated in Figure 3. However this assumption is not valid with modern plant and small numbers of units cannot be disregarded simply on the basis of the number. Otherwise the absurd situation of allowing, for instance, four units with availabilities of 0.5 to offer security, but disregarding single units with an availability of, say, 0.99. The likelihood of all units of the four-unit system being unavailable is significantly greater than that of the single unit system.

This situation is discussed further below. However before doing so it is necessary to

recall the objective and criterion associated with P2/5. The objective is to ensure that there is a reasonable likelihood that sufficient capacity exists in the network to supply customers given that a first or second outage (depending on group demand) has occurred. It fully recognises that this capacity may not always be available. The criterion used is binary, the system is either P2/5 compliant or it is not: there are no halfway measures nor degrees of compliance. Therefore a system is non-compliant whether a small deficiency or a large deficiency exists. This implies that it is irrelevant whether the non-compliance is due to a single unit being unavailable or due to one or more units of a multi-unit system are unavailable. This is important in judging the contributions associated with single and multi- unit systems.

It is also important to note that P2/5 is not concerned with allocating “contributions”.

Instead it is only concerned with system capability and states that the capacity of a network is the sum of all capacities available to the group demand. If this exceeds the group demand, then the system is P2/5 compliant, otherwise it is not. Unfortunately P2/5 uses two terms, capability and contribution seemingly in the same sense. The title of Section 3 of P2/5 is Capability of a Network to Meet Demand. The capability is the total sum, not a partial one. Not all this capability may be needed but it is available, and that is all P2/5 is concerned with. This of course does not solve the subsequent problem of allocating contributions. This is a commercial problem, which is not a function of P2/5, nor therefore of P2/6. It is also worth noting that P2/5 (and therefore P2/6) only considers the worst-case scenario. If it can withstand this scenario, i.e. the one providing the smallest capability under a first circuit outage (FCO), then all other FCOs will be satisfied. Consequently, generation for instance may not provide a contribution under one scenario (the one considered by P2/5 say), but could under another. As an example, generation not available for the complete 24hrs is neglected in assessing compliance with P2/5 if it does not generate during peak demand, but could be available and contribute to security during another FCO when it is generating. Noting the objective of retaining the concepts and philosophy of P2/5, this consideration of contribution and the associated commercial aspects require separate consideration outside of the present project.

It is a very valid argument whether the criteria, concepts and philosophy of P2/5 are

acceptable or sufficient. However this would require a complete and radical change to the

Page 28 of 150


philosophy of P2/5, which is not the objective of WS3 nor within the scope of this activity. Two of the present authors have considered a much wider set of criteria to cover these concerns and these were described and discussed in a previous FES report12.

The methodology to be used to develop P2/6 and approved by WS3 determines the

percentage contribution that can be allocated to a specific generating system. This contribution depends on the number of units and their availabilities. This conforms fully with P2/5 and can be applied irrespective of number of units. However a subsequent judgement can then be made whether this contribution is considered acceptable. Three possible approaches have been raised and discussed during the present project. These are:-

i) the approach used in P2/5, whereby two and three units have a materiality

test applied, and all other systems considered acceptable without further question. This approach has been discussed in Section 2.2.2 and no further substantial discussion is required at this point

ii) an approach based on a minimum level of redundancy iii) an approach based on the likelihood of delivering the contribution, i.e. a true

risk-based approach. In order to discuss the approaches identified in (ii) and (iii), it is useful to recall a

results table presented in Section 4.1.3 of the Methodology Report. This is shown below as Table 3. The evaluation assumes unit availabilities of 86% and the probabilities are determined from the respective capacity outage probability tables. The procedure for evaluating capacity outage probability tables is described in detail in the Methodology Report.

Table 3 - Percentage contribution and associated redundancies

number of units

% contribution from Figure 3

probability of delivering

contribution

probability of not delivering contribution

degree of redundancy, i.e. number of units in excess

of minimum required

1 49.9 86.0 14 0 2 60.7 74.0 26 0 3 65.6 94.7 5.3 1 4 68.0 90.3 9.7 1 5 69.6 85.3 14.7 1 6 71.1 80.0 20.0 1 7 72.5 74.4 25.6 1 8 73.5 91.1 8.9 2 9 74.6 88.0 12.0 2 10 76.1 84.5 15.4 2

Table 3 provides some very interesting results. Column 2 gives the percentage

contributions determined by the methodology and therefore these conform with P2/5. If the philosophy of P2/5 is applied without question, then it would seem:-

12 “Network Security Standards with Increasing Levels of Embedded Generation”. ETSU Project K/EL/00287.

Final Report by R.N.Allan and G.Strbac, UMIST, 10 August 2002.

Page 29 of 150


• that for a single unit, would not be considered • those for two and three units, would be conditioned by the materiality test

conforming with paragraph 3.6 of P2/5 • those for four units and greater, would be applied without any further

assessment. This aspect is discussed further in Section 2.3 under Materiality Considerations. Column 3 of Table 3 shows the probability that the generating system will be in a

state sufficient to deliver the contribution associated with it, and column 4 shows the complementary value, i.e. the probability of not delivering the contribution. This latter value can be defined as the associated risk, i.e. the likelihood of the generating system not delivering its percentage contribution.

It can be seen that the risk associated with a single unit is 14%, the unavailability of

the unit. Although this may seem high, it is important to note that the risks associated with 2, 5, 6, 7 and 10 units are even greater, that for 7 units being nearly twice that of a single unit. The reason for the increasing level of risk is that the degree of redundancy does not change during these increases, the risk only decreasing when the degree of redundancy increases. The dilemma at this point is that, although the apparent concern is with single units, and to a lesser extent two and three unit systems, there ought be more or as much concern with greater number of units. Otherwise it is tantamount to saying that one would rather fly in the previous four-engined plane rather than the safer single-engined plane.

It is true that with a single unit system, once the unit has failed (with a probability of

14% in the present case), zero capacity is available, whereas with a multi-unit system there is likely to be some states which, although less than the required capacity, can still provide some output. However, two points arise from this. Firstly, P2/5 has a binary criterion – it is either compliant or not, and there is no concept of being “nearly” compliant, or “not so nearly”. Secondly, the aeroplane analogy simply means that a single-engined plane will crash immediately whereas a multi-engined one may take a little longer, but both are unsafe and lack acceptable security.

At this stage, various alternative ways forward were identified and discussed with

Workstream 3. i) Do nothing and take the values given by the methodology without question.

This would be the most consistent with the methodology though possibly not with the current P2/5, if it is assumed that single unit systems should be associated with zero capability

ii) Take the values given by the methodology but include a constraint clause of the form of paragraph 3.6 in the current P2/5. This would seem to be the most consistent with the current P2/5 but would not ensure consistency between the treatment of different number of units and their availabilities, and between the effect of different technologies. Therefore, it would not be consistent with one of the key objectives of the review.

iii) Specify a minimum level of redundancy. This is a subjective judgement and cannot be supported quantitatively. If implemented, then Table 4 can be derived from the results in Table 3. In this case the values in “red” are used,

Page 30 of 150


otherwise the values shown in column 2 are used. It would seem unwise to specify a redundancy other than zero or one. This would not be based on true risk assessment, nor would it be consistent with the current P2/5 for two reasons. Firstly, the concept of redundancy is not mentioned or used in any sense in either P2/5 or ACE Report 51. Secondly, although not specifically stated, P2/5 does in fact permit zero redundancy: in the case of two units, the system capability of 67% can only be provided if both units are working, i.e. there is a redundancy of zero

iv) Specify a level of risk that should not be exceeded. This can be established from P2/5 since the specified 67% has an implicitly associated risk that can be calculated. If the contribution given by the methodology has an associated risk that is greater than this specified level, the contribution should be reduced until the risk is satisfied. This would ensure the greatest consistency between the treatment of different number of units and availabilities, but would not be consistent with the current P2/5, and is beyond the scope of the present project. The previous FES Report13 by two of the authors describes several alternative approaches for including risk assessments.

Following discussions with WS3, it was decided that alternatives (iii) and (iv) were

not appropriate and only alternatives (i) and (ii) were to be considered.

Table 4 - Contributions modified by level of redundancy

number of units

% contribution

using methodology

% contribution for a minimum redundancy of x units, where x =

unit availability = 0.86

redundancy ignored

1 2 3

1 2 3 4 5 6 7 8 9 10

49.9 60.7 65.6 68.0 69.6 71.1 72.5 73.5 74.6 76.1

49.9 60.7 65.6 68.0 69.6 71.1 72.5 73.5 74.6 76.1

0 50.0 66.7 75.0 80.0 83.3 85.7 87.5 88.9 90.0

0 0

33.3 50.0 60.0 66.7 71.4 75.0 77.8 80.0

0 0 0

25.0 40.0 50.0 57.1 62.5 66.7 70.0


Page 31 of 150


2.2.4. Materiality Testing Concern (iii) in Section 2.2.1 has to some extent been discussed simultaneously with

Concern (i) in Section 2.2.2. However a materiality test is included in P2/5 as paragraph 3.6 in order to ensure that generating units do not become dominant in the contributions associated with them. This materiality rule needs to be reviewed and updated in order to take into account variations in unit numbers and availabilities. In addition, the objectives for these materiality rules need to be studied and their relevance to an update of P2/5 needs full consideration and discussion. For these reasons it was considered desirable to deal with materiality separately and in more detail. This is included as Section 2.3.

2.2.5. Other Considerations In order to provide further input to these discussions it is interesting to note the

following additional observations, all of which are extracted from Table 5:-

Table 5 - Comparison between contributions and capacity levels number of units

capacity state %

state probab

ility

% contribution

from Table 3

minimum state

capacity providing

this contribution

probability of providing contribution

average capacity

of “good” states

%

ratio of contribution to average capacity of

“good” states

1 100 0

0.8600 0.1400

49.9 100 0.8600 100 0.50

2 100 50 0

0.7396 0.2408 0.0196

60.7 100 0.7396 100 0.61

3 100 67 33 0

0.6361 0.3106 0.0506 0.0027

65.6 67 0.9467 89.1 0.74

4 100 75 50 25 0

0.5470 0.3562 0.0870 0.0094 0.0004

68.0 75 0.9032 90.1 0.75

5 100 80 60 40 20 0

0.4704 0.3829 0.1247 0.0203 0.0017 0.0001

69.6 80 0.8533 91.0 0.76

6 100 83 67 50 33 17

0.4046 0.3952 0.1608 0.0349 0.0043 0.0003

71.1 83 0.7997 91.8 0.78

Page 32 of 150


number of units

capacity state %

state probab

ility

% contribution

from Table 3

minimum state

capacity providing

this

probability of providing contribution

average capacity

of “good” states

ratio of contribution to average capacity of

“good” contribution % states

7 100 86 71 57 43 29

0.3479 0.3965 0.1936 0.0525 0.0086 0.0008

72.5 86 0.7444 92.4 0.78

8 100 88 75 63 50 38 25

0.2992 0.3897 0.2220 0.0723 0.0147 0.0019 0.0002

73.5 75 0.9109 88.6 0.83

9 100 89 78 67 56 44 33

0.2573 0.3770 0.2455 0.0933 0.0228 0.0037 0.0004

74.6 78 0.8798 89.0 0.84

10 100 90 80 70 60 50 40 30

0.2213 0.3603 0.2639 0.1146 0.0326 0.0064 0.0009 0.0001

76.1 80 0.8455 89.5 0.85

column 2: bold indicates “good” states, i.e. those capable of providing the associated contribution column 3: state probabilities obtained from relevant capacity outage probability tables column 4: percentage contributions obtained from Figure 3 and Table 3 column 5: the least “good” capacity state column 6: summation of “good” state probabilities column 7: Σgood states(column 2 × column 3) ÷ column 6 column 8: column 4 ÷ column 7 • a single unit with an availability of 86% is associated with a contribution of

only 50% although, when available, it can deliver 100% of its capacity. Therefore it is given a significant penalty for being a single unit

• multi-unit systems are associated with greater contributions (76.1% for 10 units) although, when available, they may deliver less, perhaps much less, than their capacity (down to 80% for 10 units)

• it follows that the smaller number of units are given increasing penalties, i.e. the ratio between associated contributions and their actual capabilities. A measure for this penalty is given in column 8 of Table 5, which shows that for a single unit it is 50% whilst that for 10 units it is only 85%.

Page 33 of 150


2.2.6. Addressing Concerns and Proposals from WS3 As stated in Section 2.2.1, at the time the Methodology Report14 was issued,

Workstream 3 raised several concerns and made several proposals that addressed these concerns. These concerns and proposals have been considered in the current project and assessed and discussed in the previous sections. From these considerations, the following comments and conclusions can be drawn.

Concerns made by WS3 i) the need to be consistent with the assumptions underpinning the development

of P2/5 - the methodology achieves this, as far as the assumptions can be

discerned from P2/5 and ACE Report 51. The reasons for treating one, two and three units differently from greater numbers in P2/5 are very difficult to determine. The rationale discussed in previous sections of this report implies that the most likely reasons are no longer applicable if the contribution associated with each number of units is assessed separately as the methodology is intended to do. Therefore all unit numbers should be treated identically.

ii) the problem of relying on a single unit which, if failed, means that it provides

no security, i.e. there is no apparent redundancy - although this is true, the same problem exists for greater number of

units, depending on the degree of redundancy in the system. The concept of redundancy is therefore not an issue.

iii) the difficulty of accepting that a single unit, or even a small number of units,

should be a dominating factor in supplying security. This has been defined previously as “materiality”

- this seems to be a subjective judgement, which may not be supportable by objective assessments. The concept of “materiality” however is much more wide ranging than just related to single unit systems and is considered in greater detail in Section 2.3, in which appropriate materiality rules are developed.

Proposals made by WS3 a) a single set’s contribution will be recognised as a special sub-grouping by

whatever format of P2/6 is agreed upon - this is superseded by the concept of the methodology which deems

that the contribution is dependent on the number of units. Therefore the statement applies conceptually to all unit numbers

b) the contribution level is likely to be lower than that from multiple sets of the same aggregate capacity and that this will be specifically computed

14 “Security Contribution from Distributed Generation”, ETSU/FES Project K/EL/00287 Extension. Final Report by R.N.Allan, G.Strbac and K.Jarrett, UMIST, 11 December 2002

Page 34 of 150


- this is true and was shown in the Methodology Report and confirmed in Figure 3 and Table 3. In fact the contribution increases with the number of units

c) materiality rules, similar to those in paragraph 3.6 of P2/5 will be used, and these will be formulated within Phase 2 after sample calculations have been executed. The methodology for formulating these rules will follow the logic outlined in ACE 51

- appropriate materiality rules have been developed and discussed in later sections of this report.

2.3. Materiality Considerations

2.3.1. Impact in P2/5 P2/5 states that Table 2 of P2/5 gives the contribution of generation given the

conditions specified in paragraph 3.6, that is:- (a) the cyclic rating of the largest transmission or distribution circuit is greater

than two thirds of the total sent-out capacity of the two largest generation units and

(b) the cyclic rating of the two largest transmission or distribution circuits is greater than two thirds of the total sent-out capacity of the three largest generation units and

(c) the load pattern in the group is similar to the national load pattern. Condition (c) is not of significance in the present consideration of materiality,

however, conditions (a) and (b) are very important. The problem with P2/5 is that it does not include any reasoning why these conditions have been specified. This only becomes clear from consideration of ACE Report 51. Before discussing ACE Report 51 however, it is worth commenting on the effect of conditions (a) and (b) above.

Consider three scenarios15:- • Scenario 1: two identical circuits each with a cyclic rating of CT, and three

identical generating units each with a capacity of 9MW, i.e. a total of 27MW. condition (a) implies that CT must be greater than 12MW, i.e. it must be

greater than that of the largest generator condition (b) implies that CT must be greater than 9MW, i.e. it can be

equal to that of the largest generator in both cases the implication is that any single generating unit must have

a capacity less than or equal to that of any single circuit, i.e. it is not dominant

the total capability of the system after the FCO is 30MW (= 12 + 0.67 x 27).

15 In these scenarios, it is assumed that generators have 86% availability and hence the effective capacity of 67%, as in P2/5.

Page 35 of 150


it is worth noting that, in this case, the two lines cannot support a group demand of 30MW under normal operating conditions. Therefore the generation would have to make a contribution, not only after a FCO, but also under normal conditions. This can be achieved with conventional generation but not with intermittent sources. In this latter case the group demand could not exceed 24MW.

• Scenario 2: two identical circuits each with a cyclic rating of CT and three

generating units having output capacities of 12MW, 12MW and 3MW, i.e. the same total of 27MW. condition (a) implies that CT must be greater than 16MW, i.e. it must be

greater than that of the largest generating unit condition (b) implies that CT must be greater than 9MW, i.e. it can be

significantly less than that of the largest generating unit in the second case, the implication is that two generating units can be

dominant provided the third is significantly not so the total capability of the system is 34MW, greater than Scenario 1.

• Scenario 3: two identical circuits each with a cyclic rating of CT and three

generating units having output capacities of 21MW, 3MW and 3MW, i.e. the same total of 27MW. condition (a) implies that CT must be greater than 16MW, i.e. it can be

less than that of the largest generating unit condition (b) implies that CT must be greater than 9MW, i.e. it can be

significantly less than the largest generating unit in both cases the implication is that at least one generating unit can be

dominant provided the others are significantly not so the total capability of the system is 34MW, the same as Scenario 2.

The question at this point is: “What was really intended by conditions (a) and (b) in

paragraph 3.6 during the development of P2/5”? There is no obvious answer in P2/5. Anecdotal information suggests that the materiality rules were set to prevent generation being the dominant contributor. However the above scenarios suggest that the rules can permit one or more dominant generating units. In order to establish new materiality rules for modern type generation, it is necessary to revert to ACE Report 51.

2.3.2. Consideration of ACE Report 51 It should be remembered that there is no formal linkage between P2/5 and ACE

Report 51 since P2/5 does not mention or refer to ACE Report 51. WS3 have stressed that, although the legal obligations on DNOs in their licences only refer to P2/5, there is a general expectation that licensees are cognisant of ACE Report 51 and would be expected to defend their decisions on the application of P2/5 in the light of ACE Report 51.

ACE Report 51 includes the same table as Table 2 of P2/5. This is Table A.2 of

ACE Report 51. However it also includes a more detailed table from which Table A.2, and hence Table 2 of P2/5, is derived. The significant parts of both tables are included here as Tables 6 and 7.

Page 36 of 150


Table 6 – Part of Table A.1 of ACE Report 51

1 2 outage condition

system capacity including three-shift manned generation

no outage (a) total transmission/distribution capacity plus (b) maximum sent-out capacity of all generation units x 2/3 = SC0

first circuit outage SC0 minus the larger of (c) capacity of any transmission/distribution circuit or (d) maximum sent-out capacity of any two generation units x 2/3 = SC1

second circuit outage SC0 minus the larger of (e) the sum of (c) and (d) above or (f) capacity of any two transmission/distribution circuits or (g) maximum sent-out capacity of any three generation units x 2/3 = SC2

It is evident from Table 6 that conditions (a) and (b) in paragraph 3.6 of P2/5 and re-

stated in Section 2.3.1 are not an inherent part of the determination of system capacity with no outages nor with one or two outages. Instead, the effective capacity available to satisfy a group demand is determined by the remaining capacity after an outage involving transmission/distribution circuits or generating units or a combination in the case of a second order outage, whichever is the smallest.

Table 7 - Part of Table A.2 of ACE Report 51 and Table 2 of P2/5

type of generation

contribution after first circuit

outage classes of supply A-

E

contribution after second circuit

outage classes of supply D

and E only

notes

base load units

67% of DNC16 67% of DNC load factor 30% or above

gas turbines 67% of DNC 67% of DNC the contributions should be restricted to supplying that part of the demand which is not required to be supplied immediately following first or second circuit outages, and/or to reliving short term overloads of transmission or distribution circuits following such outages

16 Declared net capability: defined in P2/5 as “The declared gross capability less that proportion of the normal total works power consumption attributable to that set or unit (measured as MWSO), where SO is interpreted as “sent-out”

Page 37 of 150


Two situations arise from consideration of ACE Report 51; the simplification to Table 7 (Table A.1 of ACE Report 51 and Table 2 of P2/5), and the term circuit outage. These are discussed in the following sections.

2.3.3. Simplification of Table A.1 of ACE Report 51 There does not seem to be any fundamental reason for simplifying Table A.1 of

ACE Report 51 since this table can be used without introducing any simplifying conditions and applied to any situation. Therefore the only seemingly logical reason why this was done in the 1970s was to create as simple a table as possible, which could be applied consistently by all distribution companies. Introducing the conditions specified in paragraph 3.6 of P2/5 meant that the loss of capacity on circuit outage(s) would always be greater than the loss of “effective” generation. In so doing, Table 6 simplifies to Table 7.

It should be noted that ACE Report 51 continues with the statement; “If situations

are encountered where conditions 3.6a and b do not hold, the more-complex Table A.1 may be used”. This statement was not transferred from ACE Report 51 into P2/5, for reasons that remain unknown.

It should also be noted from Table 6 that it would seem condition 3.6a applies only

to the first circuit outage (FCO), which concerns all classes of supply, but that condition 3.6b applies only to the second circuit outage (SCO), which concerns only classes of supply D and E. Therefore it would seem that this condition need not be considered for any supply class below class D, and only condition 3.6a of P2/5 is then of relevance.

2.3.4. The Term Circuit Outage

P2/5 and ACE Report 51, including the relevant tables, describe the outage conditions to be considered as Circuit Outages. The immediate implication is that these consider outages of the network circuits only, thus neglecting generating unit outages. However Table A.1 of ACE Report 51 includes generating unit outages as discussed in Sections 2.3.2 and 2.3.3. Therefore there seems to be some misuse of terminology. Two possible reasons exist:-

i) Despite the fact that ACE 51 also refers to generating unit outages, the

outcome of ACE 51 and P2/5 is that, because of the conditions in paragraph 3.6 of P2/5, only actual circuits are considered as outage conditions to be met. This is possible but not entirely convincing.

ii) A more likely reason flows from a detailed consideration of the capacities associated with the generation states considered in Table A.1 of ACE 51. These are assessed as being equivalent to the circuits considered on outage in the first and second circuit outage conditions. A more detailed discussion of this follows since this consideration is fundamental to further development of the materiality rules.

Page 38 of 150


To repeat, condition (a) of paragraph 3.6 of P2/5 states that “the cyclic rating of the

largest transmission or distribution circuit is greater than two thirds of the total sent-out capacity of the two largest generation units”. For identical17 generating units, this translates into:-

2)32( ×> GT CC (1)

The factor 2/3 is the contribution that was deemed to be associated with a generating

unit having an availability of 86%. The value of this factor is one of the key objectives of this and previous studies and has been amended during the review of P2/5. The outcome of this review recommends a range of values depending on unit availability, number, technology, etc. The 2/3 factor is therefore not of any significance at this point. The principle of Equation (1) would still remain but with a revised factor, F, giving:-

2).( ×> GT CFC (2)

The remaining question is why is one circuit made equivalent to two generators?

The logic, according to ACE 51, is based on the likelihood of circuits failing compared with generating units. ACE 51 states: “It is also evident from simple probability theory that, for groups supplied by two or more generators, the probability of having coincident forced outages of two sets will be equal to or greater than 2% (= 0.142). This will usually exceed the probability of having a forced outage of any of the transmission (distribution) circuits supplying the group, thus the effect of losing the two largest generating units should be considered as an alternative to the loss of the largest transmission (distribution) circuit.” Although not explicitly stated at the same point of ACE 51, the implication is that a transmission circuit has an unavailability of less than 2%: a value of 0.5% is quoted previously.

Therefore the logic for a “first circuit outage” is either one actual circuit on outage

or a generation outage state having a similar level of probability. On the basis of a unit availability of 86%, the authors of ACE 51 seem to conclude that this would be one in which two generating units were on outage. It is worth stressing at this point that the FCO to be considered in assessing P2/5 is meant to be the most critical FCO condition. The assumptions in paragraph 3.6 exist to ensure this is achieved, and do not imply that single generator systems cannot make a contribution to security. Therefore it would be wrong to assume that at least two units are required to start considering the contribution of generation to security.

For this philosophy to continue following a review of P2/5, the number of units

and/or number of circuits must also be reviewed. Logically if a generating unit has the same

17 the following logic and associated equations assume identical units. This is for simplicity of explanation only and the logic is equally applicable to non-identical units.

Page 39 of 150


availability as a circuit, it seems reasonable that the ratio should become one-to-one rather than two-to-one. In this case Equation (2) further generalises to:-

GGT NCFC ×> ).( (3)

A similar set of logic also applies to consideration of second circuit outages and the

need to equate this to three generating units on outage (condition (b) of paragraph 3.6 of P2/5).

There were several questions and issues that had to be discussed and resolved before

this question of materiality could be finalised. These included:- • is a materiality rule necessary or can a table similar to Table A.1 of ACE 51

be used? • should a simplified table similar to Table 2 of P2/5 be required, or should a

more complete table similar to Table A.1 of ACE Report 51, or should both be included?

• the probability of two identical generating units on outage is only given by q2 if two units exist. The general relationship for identical units is nCr.p(n-r).qr where n is the number of units, r is the number on outage, p is the availability, q is the unavailability and nCr is the number of combinations. For four units and an unavailability of 14%, gives the probability of any two units on outage of 8.7%, increasing with the number of units. Should this effect be included in the relevant new “Table 2”?

• although failures of conventional generating units, such as GTs, are likely to be spread randomly throughout a year, those of overhead lines are not so. Instead these are more likely to fail during short periods of adverse weather involving gales, thunderstorms, snow and ice. The likelihood of an overlapping outage of two overhead lines is then very much greater than one involving two generators if the annual availability of the overhead line and the generator are similar. Should this consideration be included? In such cases, Equation (3) could become:

GGTT NCFNC ×> ).(. (4)

where NT may be greater than, equal to, or less than NG •

finally, it should be noted that the capability of a generating unit is determined by comparing it with a perfect line. This is not the case with a real circuit, which is assumed to have a capability equal to its rating. It can be questioned whether this puts circuits and generating units on a level playing field. If a generating unit and circuit have the same availability, should their capabilities be the same? If yes, then the real circuit should also be compared with that of a perfect line in exactly the same way that a generating unit is. This was considered and discussed in the Methodology Report18. If done, the most general case indicates that Equation (4) becomes:


Page 40 of 150


GGTTT NCFNCF ×> ).()..( (5)

Following detailed discussions with WS3, it was agreed that, to be consistent with

P2/5, both FT and NT should be set to unity, although this penalises generation but not distribution circuits. Also, it was agreed that the value of NG should be dependent on likelihood using a procedure similar to that used in P2/5. Therefore, the materiality rules are effectively determined on the basis of Equation (3) with the parameters F and NG being determined by the methodology underpinning the studies described in this report.

2.3.5. A Way Forward From the discussion in Section 2.3.4, it follows that ACE Report 51 and

consequently P2/5 used the principle of Equation 4 in which NT = 1 and NG = 2, on the basis that the probability of two generators failing was of the order of one line failing. As the above third bullet implies, this assumption is very crude and is only really true for the total failure of two units of a two-unit system. However it could be argued that it was a reasonable approximation for the generating systems used in the 1970s, which generally had 2-4 units with availabilities in winter of about 86% in each station. It becomes totally invalid for modern types of plant, where the number of units can be less than or much greater than this range with availabilities very different to the assumed 86%.

Before proceeding to discuss the agreed way forward, it is useful to review the steps

of the assessment procedure. These are discrete and independent:- i) determine the capability of the generating system using Table 2 of P2/5 or an

updated version of it ii) evaluate SC0, the system capability with no outages iii) determine the effective capacity of the significant contributor to the FCO. In

the case of the present P2/5, this is either the cyclic rating of the largest circuit, or the effective capacity of the two largest generating units

iv) determine SC1 by subtracting the capacity found in (iii) from SC0, or summating all capabilities except that found in (iii). The effect is the same.

v) this concept is then repeated for the SCO if applicable. The objective of this section is to address item (iii) of the above list. This in itself is

a multi-part consideration depending on whether a table of the form of Table A.1 or of Table A.2 of ACE Report 51 (i.e. Table 2 of P2/5) is used.

If a table similar to Table A.1 is to be used, then:- •

•

•

the value of NG19

in Equation 4 is determined (it is assumed that the base assumption of NT = 1 will remain fixed in the updated version of P2/5) the capacity of the largest circuit is compared with the effective capacity of the largest NG units, and the greatest chosen this is used to determine SC1 in step (iv) above

19 again it should be noted that this equation relates to identical units

Page 41 of 150


If a table similar to Table 2 of P2/5 is to be used, then:-

the value of NG in Equation 4 is again determined (it is assumed that the base assumption of NT = 1 will remain fixed in the updated version of P2/5)

•

•

•

a condition similar to that in paragraph 3.6 restricts the capability that can be associated with the generation the value of SC1 is determined by assuming the FCO is due to the largest circuit and the capability of the generation is restricted to that assessed using the condition similar to paragraph 3.6.

Both approaches can be used depending on the conditions (see Sections 5 and 7). In

both cases, it follows that the number of units “equivalent” to a single circuit is a separate exercise to that associated with “paragraph 3.6” of P2/5.

2.3.5.1. Assessment of Number of Units In P2/5 the number of equivalent units is assumed to be two. At best this can only

continue to be the number if unit availability remains at 86%. As the effect of this parameter is one of the key drivers of the P2/5 review, this assumption must be questioned, even for availabilities of 86%. This can only be done from knowledge of the associated capacity outage probability tables (COPTs). These are shown in Table 8 for 1-6 units having availabilities of 98%, 95%, 86% and 50%.

Page 42 of 150


Table 8 - Capacity Outage Probability Tables

number of units number of units

out 1 2 3 4 5 6

(a) availability = 98%, unavailability = 2%

0 0.98 0.9604 0.941192 0.92236816 0.903921 0.885842 1 0.02 0.0392 0.057624 0.07529536 0.092237 0.10847 2 0.0004 0.001176 0.00230496 0.003765 0.005534 3 8E-06 3.136E-05 7.68E-05 0.000151 4 1.6E-07 7.84E-07 2.3E-06 5 3.2E-09 1.88E-08 6 6.4E-11

(b) availability = 95%, unavailability = 5%

0 0.95 0.9025 0.857375 0.81450625 0.773781 0.735092 1 0.05 0.095 0.135375 0.171475 0.203627 0.232134 2 0.0025 0.007125 0.0135375 0.021434 0.030544 3 0.000125 0.000475 0.001128 0.002143 4 6.25E-06 2.97E-05 8.46E-05 5 3.13E-07 1.78E-06 6 1.56E-08

(c) availability = 86%, unavailability = 14%

0 0.86 0.7396 0.636056 0.54700816 0.470427 0.404567 1 0.14 0.2408 0.310632 0.35619136 0.382906 0.395159 2 0.0196 0.050568 0.08697696 0.124667 0.16082 3 0.002744 0.00943936 0.020295 0.034907 4 0.00038416 0.001652 0.004262 5 5.38E-05 0.000278 6 7.53E-06

(d) availability 50%, unavailability = 50%

0 0.5 0.25 0.125 0.0625 0.03125 0.015625 1 0.5 0.5 0.375 0.25 0.15625 0.09375 2 0.25 0.375 0.375 0.3125 0.234375 3 0.125 0.25 0.3125 0.3125 4 0.0625 0.15625 0.234375 5 0.03125 0.09375 6 0.015625

Page 43 of 150


From Table 8(a), it is clearly seen that the probability of losing one circuit of a two-

circuit network is about 4% (0.0392). It is against this value that other systems, including generation systems, should be compared since this represents approximately the likely worst outage scenario of a demand group fed by two independent circuits. In reality however, other state probability values will never be exactly equal. Therefore a range of probabilities must be considered as the basis of comparison. This is a matter of judgement and the appropriate values of this range were discussed with Workstream 3. Following these discussions, it was agreed that an upper level of about twice this base value would be acceptable, i.e. a value less than about 10%.

On this basis, the following comments can be deduced:- (a) P2/5 states that, for a unit availability of 86%, two units should be

considered as an outage state, i.e., NG = 2. However Table 8(c) shows that, although this is acceptable for 2-4 units, this number is insufficient for plants with more than four units. It is evident that an outage of three units should be considered for plants having five or more units, and that this number will increase further at some point beyond that shown

(b) from Table 8(a) for a unit availability of 98%, it is seen that, although an outage of one unit only needs to be considered in most cases, two units need to be considered if the plant consists of six or more units. The implication in ACE Report 51, and therefore in P2/5, is that an outage of only one unit needs to be considered in all situations

(c) from Table 8(b) for a unit availability of 95%, it is seen that an outage of only one unit needs to be considered for plants containing up to two units, but two should be considered for plants with more than two units

(d) from Table 8(d), it is seen that, although an outage of all units should be considered in systems with up to a five units, there is no need to consider more than five when the plant size exceeds five units

(e) whilst all of the above comments are readily deduced and easily understood, it is the concept underpinning the basis of the comments that needs to be also clarified. A simple identification of a state whose probability is equal to or nearly equal to the comparative value does not necessarily give the state of significance. For instance, the probability of one unit on outage of a six-unit system when the units availability is 50% in 0.09375, which is within the comparative range. However the significant state is when 5 units are on outage; this also having the same probability. The logic to be used is as follows. Commence with the state in which all units are on outage and proceed up the table until a state is reached in which the probability is greater than the minimum value of the range and less than the upper value of the range being used for the comparative value. Following the discussions with WS3, this range has been set to 3.9% - 10%, for reasons discussed above20

(f) this logic also leads to another important conclusion. This is that the process is equally applicable to all sizes of systems, including single unit systems.

•

The sense of this is most apparent considering systems in which the unit availability is 50%. In this instance, five units of a six-unit system

20 it should be noted that this may be a relatively complex evaluation process. A look-up table is the best approach to be used for this purpose.

Page 44 of 150


should be considered on outage, but all units of a five-, four- three-, two- and therefore logically a one-unit system. It should be noted that outaging all units of a system is a much more severe and significant test than outaging only some of the units. This logic therefore also applies to all other systems. In the case of those in which the unit availability is 86% (the P2/5 case), there is no difference in concept between outaging all units of a single unit system (one) and outaging all units (three) of a three-unit system in which the unit availability is 50%. In both cases the state probabilities are about the same (0.14 and 0.125 respectively).

•

It can be concluded from the above discussion that:-

i) the number of units to be considered on outage and equivalent to a single

circuit can be determined from the relevant COPT ii) this procedure applies equally to any number of units from one upwards iii) to conform with the existing P2/5, the comparison should be with the present

assumption that a circuit will have an unavailability of about 4% iv) a range of values are needed since state probabilities are discrete and very

lumpy in COPTs of systems with a small number of units v) a range from 3.9 – 10% has been recommended and agreed by WS3.

2.3.5.2. Application of the Approach (a) Illustrative Scenarios

In order to illustrate the application of the approach described in Section 2.3.5.1,

consider the arrangements shown in Figure 4.

(a) (b)

Figure 4 – Two possible network structures Let the demand be 50MW and consider three scenarios.

Scenario 1

Consider Figure 4(a) and let each circuit have a cyclic rating of T1. Then:- SC0 = T1 + T1 = 2 T1 A circuit is the FCO and hence:- SC1 = 2 T1 – T1 = T1 ≥ 50MW Therefore each circuit must be rated cyclically at 50MW.

Page 45 of 150


Scenario 2

Consider Figure 4(b) and one generator. Let each circuit have a cyclic rating of T2

and let the generator have an effective capacity of G1. Then:-

SC0 = T2 + T2 + G1 = 2 T2 + G1

if T2 > G1, then a circuit is the FCO and hence:-

SC1 = 2T2 + G1 – T2 = T2 + G1 ≥ 50MW Therefore, depending on the effective capacity of the generator, the cyclic rating of

each circuit can be significantly less than 50MW, and thus the generator can make a significant contribution to the system.

if T2 < G1, then the generator is the FCO and hence:-

SC1 = 2T2 + G1 – G1 = 2T2 ≥ 50MW Therefore, although the capacity of the generating unit does not appear in SC1 because

it causes the worst, i.e. most critical, FCO, the unit makes a significant contribution to system security in two respects. Firstly, the cyclic rating of each circuit need now be only 25MW. Secondly under the less severe FCO of a circuit, the generating unit can provide significant support of the group demand.

It should be noted that this scenario also applies to the case when more than one

generating unit exists but the complete plant must be considered on outage to ensure a similar outage probability to that of a single circuit. Scenario 3

Consider Figure 4(b) and four generators and that the outage probability of one unit is similar to that of one circuit. Therefore only one unit need be outaged to determine SC1. Let each circuit have a cyclic rating of T3 and let the generator have an effective capacity of G2. Then:-

SC0 = T3 + T3 + 4G2 = 2T3 + 4G2

if T3 > G2, then one circuit is the FCO and hence

SC1 = 2T3 + 4G2 – T3 = T3 + 4G2 ≥ 50MW Again, depending on the effective capacity of the generator, the cyclic rating of each

circuit can be significantly less than 50MW, and thus the generator can make a significant contribution to the system.

if T3 < G2, then one generator is the FCO and hence

SC1 = 2T3 + 4G2 – G2 = 2T3 + 3G2 ≥ 50MW

Page 46 of 150


Therefore, irrespective of the effective capacity of the generator, the cyclic rating of

each circuit need be less than 25MW, and thus the generator makes a significant contribution to the system.

It should be noted that this scenario is applicable in concept to all cases in which

some, but not all, units have to be considered on outage to ensure an outage probability similar to that of a single circuit.

(b) Consideration of Single Units

The previous scenarios included consideration of single units since these have now been accepted as systems that can contribute to security. However, one consideration that does not concern P2/5 is the variation in frequency and duration of outages, and this has been raised by WS3 as a source of difficulty. Two facts are important. Firstly a given unavailability is associated with an infinite range of frequencies and durations, provided the product of these two parameters remain constant. Secondly it is possible for the outage duration of a generating unit to be very much greater than that of a circuit. Two possibilities for dealing with these potential problems are as follows:-

the variation in availability/unavailability due to the distribution of frequencies and durations is ignored and only the average value of the availability/unavailability of a plant type is considered. This is directly equated to the assumptions underpinning P2/5, in which all units are assumed to have an availability of 86% and only this value is considered.

•

• the variation in availability/unavailability due to the distribution of frequencies and durations is considered. It would be unreasonable to use the longest possible duration or the greatest value of frequency, unless the risk was greater than some predetermined criterion. One approach would be to:- specify a risk level, i.e. the probability that the availability is less than

a certain acceptable level. This perhaps could be determined from knowledge of how the availability of circuits varies

determine or estimate the variation of availability of the units or plant under consideration

deduce the availability value given by the use of the above risk level use this value of availability instead of the average value considered

above and in P2/5 the main problem with this approach is limited data, and a compromise

between the two approaches may be more appropriate.

Following discussions with WS3, it was decided that the second possibility was outside the scope of these studies and inconsistent with P2/5, and hence with P2/6.

(c) Concluding Comments

It can be concluded from the above three scenarios that:-

Page 47 of 150


•

•

•

•

•

•

•

•

•

irrespective whether one or more generators exist, and irrespective whether the effective capacity of each is less than or greater than that of the cyclic rating of each circuit, the generating units contribute to system security

if the outage probability of the group of generating units to be considered on outage is similar to that of a single unit, then that causing the greatest loss of capacity should be considered as the FCO. This is the concept in Table A.1 of ACE Report 51, and provides a reasonable proxy to risk it should be noted that by ensuring the outaged generators have an outage probability equal or slightly greater than that of the single circuit, this means the generators are treated (marginally) more severely than the circuits whether the FCO is associated with the circuit or with one or more generating units, the generation still makes a contribution to system security it is also worth noting that the concept of P2/5, which is retained in this report, considers the FCO to be that causing the greatest loss of capacity. Even if the generation does not supply demand during this outage, it can still do so under other outage conditions. This is one of the main reasons why the P2/5 approach does not, and cannot, allocate contributions. The approach only considers one outage scenario, that causing the worst outage event. Therefore, allocation of contributions is quite separate.

2.3.5.3. Simplification of Tables The approach and methodology described in Sections 2.3.5.1 and 2.3.5.2 could be

used without any simplification. These provide all that is required to generate tables similar to that of Table A.1 of ACE Report 51. The only significant differences between the existing Table A.1 and a revised version would be:-

the / factor would change to the values evaluated by the methodology and defined later in this report

23

the removal of, for instance, two generation units would be replaced by the number found using the approach described in Section 2.3.5.1.

However, to simplify this table to the equivalent of Table A.2 of ACE Report 51 and

Table 2 of P2/5, then a set of conditions similar to those of paragraph 3.6 (a) and (b) of P2/5 are needed. The conditions in the existing P2/5 and proposed new ones are shown below.

the cyclic rating of each circuit can be reduced as the effective capacity of each unit and of the overall plant increases. This point is more applicable to the case of demand growth since generation can be considered to be a viable alternative to increasing circuit capacity one of the most important factors in the process is to determine the number of generating units to be considered on outage. The criterion used in P2/5 and in this report is that the outage probability of the group of units considered to be on outage is equal to or slightly greater than the outage probability of a single circuit. The probability range suggested in this report and agreed to by WS3 seems to satisfy this criterion

Page 48 of 150


Existing paragraph of P2/5

3.6 The contribution of generation specified in Table 2 (of P2/5) is based on the assumptions that:- (a) the cyclic rating of the largest transmission or distribution circuit is greater

than two thirds of the total sent-out capacity of the two largest generation units and

(b) the cyclic rating of the two largest transmission or distribution circuits is greater than two thirds of the total sent-out capacity of the three largest generating units

Suggested new paragraph for P2/6

3.x The contribution of generation specified in Table 2 (of P2/6) is based on the assumptions that:- (a) the cyclic rating of the largest 21distribution circuit is greater than F% of

the total sent-out capacity of the N1 largest generation units and (b) the cyclic rating of the two largest distribution circuits is greater than F% of

the total sent-out capacity of the N2 largest generating units

where F is given by the contribution methodology and detailed later. This parameter will not be fixed as in P2/5 but will be a variable depending on the characteristics of the units N1 and N2 is given by the approach described in Section 2.3.5.1 and will not be fixed as in P2/5 but will be variables depending on number of units and their availability.

2.3.6. Application Examples It is useful at this point to demonstrate the approach in more detail and to illustrate

its application and consequential effects. This is achieved using the following examples. Example 1: Consider three identical generating units each with a capacity of

9 MW (27 MW in total) and availability of 86 % connected to a specific group demand, as shown in Figure 5. Evaluate the total capability contribution.

The security contribution of three identical units is 65.6 %, giving a capability of

3 x 9 x 0.656 = 17.7 MW. This would be added to the remaining circuit capacity to determine the capability of the system under a first circuit outage.

21 “transmission” has been omitted from this suggested paragraph on the advice of WS3 since generally P2/5 is not formally used by NGT as its security standard

Page 49 of 150


G G G

D

G

Figure 5 - Group demand supplied by three generating units Example 2: Consider a fourth identical generating unit added to the existing three

units of Example 1, giving a total capacity of 36 MW. Evaluate the updated total capability contribution.

The security contribution of four identical units is 68.0 %, giving an updated

capability of 4 x 9 x 0.68 = 24.5 MW, and a capability increase of 6.8 MW. It should be noted that the capability of the three existing units is slightly increased by the addition of this fourth one. This is due to the percentage contribution increasing with number of units.

Example 3: A specific group demand is supplied by two circuits each with a

cyclic capacity of 32.5 MW and six generators each with a capacity of 9 MW and availability of 86 %, as shown in Figure 6. Evaluate the system capability following a first circuit outage.

G G G

D

G G G

Figure 6 - Group demand supplied by two circuits and six generating units This assessment depends on whether a circuit or a set of generating units is

considered as the FCO. Both are considered here taking three generating units as being equivalent of a single circuit outage as indicated in Table 8. The capability is then the smaller of the two assessed capabilities. If the FCO is:

(a) a circuit: the capability = 32.5 + 6 x 9 x 0.711 = 70.9 MW (b) three generators: the capability = 2 x 32.5 + 3 x 9 x 0.711 = 84.2 MW

meaning that the system capability following a FCO is a maximum of 70.9 MW. Example 4: This example illustrates the procedure that needs to be used to

estimate the system capability and the consequential group demand that can be met.

Page 50 of 150


A specific group demand is supplied by two circuits each having a capacity of 25 MW and a cyclic rating factor of 1.3. Furthermore, a range of generating units is considered to be connected to this group. Calculations similar to those in Examples 1-3 are performed in order to determine the group demand that can be supported with these circuits and generation plant.

The results are shown in Table 9. The following comments can be made:- •

•

•

in most cases the “first circuit outage, FCO” is dominated by an actual circuit outage, not a generation outage. In these cases, the group demand that can normally be satisfied is dependent on the remaining circuit and the effective generation capacity the only two exceptions to the above are when four generating units of 24MW exist with availabilities of 86% and 50%. In these cases the FCO is associated with two and four units on outage respectively (in the first case the generation dominance is very marginal), and the group demand that can normally be satisfied is dependent on both circuits and any remaining generating units the materiality rule of P2/5 was created to ensure that generation is not dominant and that the FCO is due to a circuit only. For this to be the case in these examples, it is necessary to limit, i.e. cap22, the capacity of the generation that can be credited towards security in the two cases that generation is dominant:-

o in the case of 86% availability, each unit must be capped at 23.9 MW, when the group demand that can be satisfied reduces to 97.5 MW

o in the case of 50% availability, each unit must be capped at 19.8 MW, when the group demand that can be satisfied reduces to 65.0 MW

22 this capping does not limit or cap the actual capacity of a unit nor its output, only the capacity that is considered for assessing its contribution to system security

Page 51 of 150


Table 9 – Results of Example 4

Input data

Number 2

Capacity (MW) 25 Circuits

Cyclic rating factor 1.3

Number of units 3 4 4

Capacity (MW) 8 8 24

Generation plant

Availability 95% 86% 50% 95% 86% 50% 95% 86% 50%

No outage

Contribution 76.4% 65.6% 39.3% 82.5% 68.0% 41.1% 82.5% 68.0% 41.1% Generation plant Effective

capacity (MW)

18.34 15.74 9.43 26.40 21.76 13.15 79.20 65.28 39.46

Group demand that can be met with no outage SC0 (MW)

83.34 80.74 74.43 91.40 86.76 78.15 144.20 130.28 104.46

“First circuit outage”

Circuit outage

Group demand that can be met (MW)

50.84 48.24 41.93 58.90 54.26 45.65 111.70 97.78 71.96

Number on outage23 1 2 3 1 2 4 1 2 4

Generation outage

Group demand that can be met (MW)

77.22 70.25 65.00 84.80 75.88 65.00 124.40 97.64 65.00

System capability

Group demand that can be met SC1 (MW)

50.84 48.24 41.93 58.90 54.26 45.65 111.70 97.64 65.00

23 Taken from Table 8. One unit is outaged for 95% availability, two units for 86% availability and all units for 50% availability.

Page 52 of 150


2.4. Disaggregation of Demand and Generation Output

The topic of disaggregation of demand and generation was included in the original

set of proposed activities. However it was subsequently decided that this topic would be discussed and dealt with directly by Workstream 3, and not by the authors of this report.

2.5. De Minimus Levels of Capacity

The topic of de minimus levels of capacity was included in the original set of proposed activities. However it was subsequently decided that this topic would be discussed and dealt with directly by Workstream 3, and not by the authors of this report.

2.6. Intermittent Generation and Tm Assessment Issues

2.6.1. Background The analysis carried out in the Methodology Report24 demonstrated that a number of

important factors influence the capability that could be attributed to intermittent generation such as wind. These are:-

i) Required minimum persistence Tm, ii) Resolution (granularity) of the intermittent generation data, iii) Correlation between intermittent generation output and peak demand, iv) Diversity between intermittent generation sites, e.g. wind farms (footprint) These factors are discussed in turn in the following sections.

2.6.2. Concept of Persistence and Effective Capacity

2.6.2.1. Basic Methodology

The concept of persistence is one of the key factors driving the security capability of intermittent generation. In contrast to non-intermittent generation, intermittent sources persist in generating at a particular output level for significantly shorter periods of time. This effect must be taken into account in the assessment process and the minimum persistence time for which generation is required has been defined as Tm. 24 “Security Contribution from Distributed Generation”, ETSU/FES Project K/EL/00287 Extension. Final Report by R.N.Allan, G.Strbac and K.Jarrett, UMIST, 11 December 2002

Page 53 of 150


The developed methodology, described in the Methodology Report, calculates the

effective capacity of intermittent generation as a function of Tm. In summary, this is the following two-stage process:

i) firstly, given the output profile over a specified period of time, e.g. the

winter season or a whole year, of an intermittent generation system such as a wind farm, and the required persistence level (Tm), the corresponding capacity outage probability table (COPT) is formed

ii) secondly, as in the case of non-intermittent generation, the appropriate load

duration curve (LDC) is superimposed on the COPT and the expected energy not supplied (EENS) is calculated. The effective generation contribution is then deduced as the transmission circuit capacity which, when substituted for the generating plant, gives the same value of EENS.

As demonstrated in the Methodology Report, the persistence level (Tm) has a

considerable impact on the capability that can be associated with intermittent generation. Using an annual wind profile and a Tm of ½ hr, the capability of the wind farm to network security was about 30%. Increasing the level of required persistence to a Tm of 24 hr, reduced the capability to about 15%. This stresses the importance of understanding the rationale for determining the level of Tm required in different systems and circumstances.

2.6.2.2. Rationale for Determining Tm

Tm is the period of time for which generation will need to operate continuously at or above a certain output level in order to support the demand and hence to provide system security. This period of time is therefore related to the duration of the system conditions for which such generation may be able to avoid or reduce customer disconnections.

There are three distinct system conditions, each of which can be associated with

different minimum persistence times. These are25-

(i) switching activities (ii) repair activities (iii) maintenance activities

Before entering into a discussion of these specific system conditions, it is important to

differentiate between Tm and system response times. P2/5 prescribes maximum times to restore supplies. For demands less that 1MW this can be the associated repair time of the system. For group demands up to 12MW, most of the demand should be restored within 3 hours and for group demands between 12MW and 60MW, much of the demand should be restored within 15 minutes. These times are therefore prescribed maxima for demand 25 Prolonged network development and replacement activities are not covered by P2/5. P2/5 is based on the concept of first circuit outages only (potentially superimposed on maintenance activities), while the probability of two circuits being simultaneously on forced outage is considered to be small and could be neglected. This is because repair activities of faulted circuits were assumed to be completed within sufficiently short periods that the probability of a second independent outage remained small.

Page 54 of 150


restoration times and should not be used as a proxy for Tm. These are allowable interruption times, i.e. the longest outage times before restoration must take place. Generation needs to become available within these times, i.e. they are the maximum values of start-up times or reconnection times. Generation needs to be persistent for a period of time after these response times during which other means of supply are not available or not fully adequate.

Two example scenarios can be considered:-

i) switching cannot be accomplished within the P2/5 response times. The

(distributed generation) DG must be capable of satisfying all demand from the latest response time in P2/5 until the switching is completed – generally a small Tm. This situation would occur for instance if the group demand increased above 12MW. Instead of automating, manual switching could be retained and intermittent DG relied on for the time period between 15 min and up to 3 hr. Therefore, the worst case scenario is to adopt a value of 3 hours for Tm.

ii) switching is accomplished within the P2/5 response times. The DG must be

capable of supporting any excess demand over that supplied by the switched network. This also applies to the second period of scenario (i), i.e. after switching. This probably involves long Tm, unless other alternative supplies are possible but network reconfiguration is more complex. In the case when the restoration process involves sequential switching (of which some could be remote and some manual) the required amount of generation may be different for each of the intervals of the sequence. The support that generation systems could make for each interval can be evaluated using the developed methodology.

2.6.2.3. Minimum Persistence Times for Switching Actions

Intermittent generation connected to the demand group could support this demand during the period in which transfer capacity is being switched in. In this circumstance, Tm can be conceived to be the period between the maximum restoration time specified in P2/5 for the Class of Supply being considered and the time point when switching of the transfer capacity is completed. The generation should be considered as contributing to security whilst this switching takes place. The associated value of Tm is likely to be relatively small. In the case of 11kV systems (Class B), manual switching could take up to 3 hours while for 33kV systems (Class C) designed to P2/5, remote switching may take up to 15 minutes.

A particular example where this consideration may be useful is when an increase in

demand in a specific group causes it to change from a Class B to a Class C group. In such a situation, the presence, and therefore capability, of a wind farm may allow manual switching to remain instead of replacing this with remote or automatic switching. Consider the case of a group demand growing from 12 MW to 15 MW. In this case, following a circuit outage, 3MW of this demand should be restored within 15 minutes, while the rest (12MW) should be restored within 3 hours. Hence, an intermittent generation system able to provide 3 MW of effective contribution for 3 hours would make the system P2/5 compliant, provided that the remaining network after being switched in can support the group demand in full (15 MW). In

Page 55 of 150


this case the manual switching could remain instead of being replaced by remote or automatic switching. In the worst case, Tm would need to be 3 hours.

Consequently, the capability of intermittent sources such as wind generation to

contribute towards system security will be greater in systems with automated switching compared with those in which manual switching is utilised.

Similar logic could be applied to other Classes of group demand.

In a similar way to the above, wind generation may be sufficient to provide immediate

support to a group following a non-damaged first circuit outage, given that the likely restoration time will be of the order of 15 - 30 minutes in most circumstances. In this case, the minimum persistence time Tm is equivalent to the restoration time of non-damaged faults.

The previous examples assume that the intermittent generation is used to support all

the group demand or less while the network is being reconfigured following a damaged fault or being restored following a non-damaged fault. In some situations, neither of these activities is completed in one step and a sequence of switching steps is required. In this case the intermittent generation can be used to support the group demand and prevent or reduce overloads that may occur during the switching sequence. This would allow earlier restoration of demand before switching is complete without encountering excessive overloads.

In all the above situations, it is evident that the value of Tm is likely to be in the

range of a few minutes to a few hours. The capability of intermittent generation is relatively significant in these time periods. It is important however to recognise that half-hour granularity of the output profile data will not indicate if generation can provide reliable support over 15 minute periods. Therefore, a shorter granularity may be needed for some situations. In order to preserve simplicity, the worst-case scenario situation could be considered, requiring Tm to take a value of perhaps 3 hours. However, it is important to bear in mind that this will be system configuration dependent and it may be appropriate to determine particular values for Tm and the corresponding contribution for the specific case in hand.

2.6.2.4. Minimum Persistence Times for Repair Activities

For situations in which, after a first circuit outage associated with a damaged fault, sustained support is required until the faulted circuit is repaired, the contribution of intermittent wind generation may be limited. This is also likely to be the case in the previous situations following reconfiguration and completion of the switching process. In such cases, the duration of repairs could last for a number of days, and this would then be the period for which security support is required. Therefore, for intermittent generation to be able to provide all or some of this support, the minimum persistence Tm will need to be sufficiently large to ensure that necessary repairs can be completed.

It is important to note that, in the above consideration, the peak demand is assumed

to last for the duration of the repair and that sustained generation support will be required during the entire repair activity. In reality, however, peak periods last for only a few hours and therefore the generation would need to sustain a certain output only during the peak period for several hours over one or more days. An analysis of such circumstances is beyond

Page 56 of 150


the scope of this project but it is worth noting that the capability attributed to the intermittent generation will be consequentially conservative.

2.6.2.5. Minimum Persistence Times for Maintenance Outages

P2/5 considers that a circuit outage may be due to a planned maintenance. In emergencies, these activities can frequently be closed down within a short period of time. Consider a circuit that is on a scheduled planned maintenance outage. This circuit is assumed to be in a group of circuits that, together with one or more intermittent sources, supply a group demand. The situation to be considered is the support of this group demand following a damaged fault on one of the remaining circuits. Under this situation, any intermittent generation connected to the group could contribute to the supply until the circuit on maintenance is put back into service, assuming this restoration time is less than the likely repair or restoration time of the failed circuit. Consequently the minimum persistence time Tm for which generation may need to provide support is related to the time it takes for an urgent return to service of the circuit undergoing a planned maintenance outage.

2.6.2.6. Summary

From the above discussion, it can be concluded that Tm can be defined in terms of :- i) switching times ii) non-damaged restoration times iii) repair times iv) return to service times (associated with network maintenance)

DNOs do not routinely collate these times, but expert judgement can be used to provide sufficient information for these purposes.

The values of Tm seem to logically fall into one of two groups, that associated with

switching processes and therefore perhaps up 3 hours, and that associated with repair actions and therefore be up to several days. In the first case the generation capability is likely to be significant but in the second case, the capability may be very small. It is also worth noting that, for cases when Tm is less than one half-hour, a resolution of the generation output profile data shorter than this may be the most appropriate to use.

Recommended26 indicative longest values for Tm are shown in Table 10. If, in any

particular compliance assessment case, these are viewed as unachievable, then a specific study would need to be performed.

26 These times were recommended by WS3, which is still looking to verify these times

Page 57 of 150


Table 10 – Recommended27 values for Tm

group demand

switching repair maintenance

A n/a n/a n/a B 24 hours 24 hours 2 hours C 3 hours 5 days 18 hours D 3 hours 15 days 24 hours E 3 hours 90 days 24 hours

2.6.3. Effect of Data Resolution or Granularity

The generation output of an intermittent generation plant, such as a wind farm, is likely to vary during each half hour period. This variation in output levels of the generation would need to be absorbed by the remaining circuits. For a short period of time, the generation output could drop significantly and hence the remaining circuits may become overloaded. This may not be of great significance if the periods are very short in duration, but could be quite the reverse if they are long. These times are relative to the system being considered and therefore the assessment may need to be system dependent. In the Methodology Report, preliminary studies were carried out to compare the capabilities associated with 30 min and 1 min resolutions of the output profiles and, as expected, the apparent capability of wind reduces if 1 min resolution of the data is used.

Furthermore, if the required persistence time Tm is less than one half hour, then clearly

a half hour resolution is not adequate and a shorter resolution may be more appropriate. From the sensitivity studies described in Section 4, the authors have recommended and WS3 has agreed to use a resolution of 5 min.

More detailed sensitivity studies were conducted during this project, and these are

described and discussed in Section 4, at which point the effect of 30, 5 and 1 min resolution is assessed and appropriate conclusions made.

2.6.4. Correlation between Peak Demand and Generation Output

It is expected in general that the system peak demand and the greatest output level of wind generation would both occur in the winter period, and to some extent be correlated. Using available wind data, the results included in the Methodology Report showed that the output of a wind farm during winter was considerably higher than during the summer months.

The results in that report also included the capability of wind generation to network

security using an annual output profile and also using profiles covering only the winter months. These results showed that the wind generation capability would be greater if winter months only were considered. For example, using a Tm of 24 hours, the capability increased


Page 58 of 150


from about 15% using an annual profile to over 21% when a three-month winter profile was used.

Since the peak load, the main criterion in P2/5, generally occurs in winter, then it is

recommended that winter profiles should be used for general application. However this does not preclude special studies for specific situations.

2.6.5. Treatment of Multiple Wind Farms

When a number of wind farms are connected to a particular demand group, a question arises as to how these should be treated when determining their contribution to security of supply. Given relatively close proximity of the wind farms, significant correlation may exist between their outputs. Therefore such wind farms could not be treated as independent. A practical analysis carried out by WS3 demonstrated this effect28. This analysis also showed that a number of factors may influence the performance of wind farms, including unit size, unit hub heights, unit locations, spatial array orientation of grouping relative to predominant regional wind pattern, regional wind patterns etc.

The developed methodology for intermittent generation cannot be used to incorporate

these factors directly. The model is essentially associated with the output profiles of wind farms. However these profiles can be used to assess indirectly the impact of these factors on the contribution to security of single or multiple farms. All that is required is recognition of the values of the parameters and to associate these with the relevant output profile. For multiple farms, the simultaneous output profiles can be obtained. The outputs would then be simply aggregated into a single profile, which can be used to assess the contribution of any given cluster of wind farms. This approach is simple, pragmatic and rigorous, given that the output profiles are the ultimate source of information regarding the contribution to security.

In case of new planned sites, some assumptions would need to be made, before actual

outputs can be obtained.

2.7. Ride Through Capability

2.7.1. Impact of Generation Trips Ride-through capability is of great importance in knowing how, or even whether,

distributed generation can contribute to system security. This capability does not affect the effective generation contribution of a particular generation site to be determined by the methodology and to be specified in a table similar to Table 2 of P2/5. What it does affect is whether this contribution is available immediately, after a certain time period, or not at all. This depends on whether the generation plant is, or is not, tripped following a circuit outage; and if tripped, how long it takes to restore its network connection.

28 Matthew Hays-Stimson, Data processing criticalities related to P2/5 class demand footprints when calculating support of dispersed multi-site wind generation.

Page 59 of 150


In order to describe these alternative ways of contributing, the various scenarios that could occur following a circuit outage29 need to be assessed.

(a) Generator sites connected to demand groups fed by a single infeed These relate to systems having a structure of the form shown in Figure 1(b) {see

Section 2.1}. In these cases, the generation site must be tripped following a circuit outage for technical and safety reasons. Under present codes of practice, the site must remain disconnected until the single infeed is restored to service. Consequently such generation sites cannot contribute to system security.

There is a move towards increased active management of distribution systems,

following which it may be possible for a load group to operate as an island. However, this is beyond the current review of P2/5 and therefore is not for consideration at this time.

It may also be possible for some special supply areas to operate as islands, but again

these are beyond the scope of this review and the normal application of P2/5.

(b) Generator sites that are designed not to trip following a circuit outage

These can only be considered for systems that have at least two permanently closed

infeeds, i.e. when the generation site is connected to a system having a structure of the form shown in Figure 1(f). If the generation site is designed so that the generator remains connected following a circuit outage, then it can be assumed that the generating units remain connected to the load group and continue to supply energy. In such cases they contribute significantly to system security, and their contribution can be considered to be available immediately. They are therefore treated in an identical way as three shift stations are treated in Table 2 of P2/5.

(c) Generator sites that are designed to always trip following a circuit outage These may exist in systems having a structure of the form shown in Figure 1(f) but

which always trip following a circuit outage, or in systems having the structure shown in Figure 1(d) because these have to be tripped and remain tripped until the network has been reconfigured and the transfer of load has been completed. In this case, several sub-cases exist.

If the load transfer can be completed and the generation site reconnected within one

minute30, then P2/5 defines31 this to be an immediate restoration of supply. In this case, the generation contribution becomes identical to that described in (b) above, and can be treated as if the contribution is continuously available.

If the generation site cannot be reconnected within one minute, then it cannot be

considered as a continuous contribution. Instead it must be treated in the same way as gas 29 This circuit outage may be the outage of an actual circuit, or the outage of other generating units 30 Although the IIP framework classifies three minutes as an immediate restoration, Table 1 of P2/5 classifies an immediate restoration as one lasting no longer than one minute 31 Despite the fact that an interruption has actually occurred.

Page 60 of 150


turbines are considered in the present P2/5, i.e. the contributions should be restricted to supplying that part of the demand which is not required to be supplied immediately following first or second circuit outages, and/or to relieving short term overloads of transmission or distribution circuits following such outages. The actual contribution however is that defined in Table 2 of P2/5 or any future contribution established by the methodology and included in any updated version of this table.

(d) Generators which may trip following a circuit outage These may exist in systems having a structure of the form shown in Figure 1(f) but,

instead of always tripping or “never” tripping, may or may not trip with some degree of uncertainty. The safest way of dealing with these is to treat them in the same way as those that always trip ((c) above). However if some risk is considered acceptable, provided it can be shown to be small, there may be some justification for allowing such generation sites to contribute to immediate system security.

2.7.2. Concluding Comments These scenarios cover all the situations that may occur. These in no way affect the

methodology, nor the effective generation contributions that the methodology determines. It is only the way that these contributions are used that is dependent on the scenario. In some cases, the contribution becomes available immediately (similar to base loaded stations in the current P2/5) and sometimes the contribution only becomes available after a period. The periods in which it can contribute depend on the Class of Supply:-

• available within 3 hours: can contribute without restriction to Supply Class B,

and to part (b) of Supply Classes C and D • available within 15 minutes: can contribute without restriction to Supply

Class C • available immediately: can contribute without restrictions to all supply

classes This process is captured in the Application Guide described in Section 7, so that one

step in the security assessment procedure determines the generation capability and a separate step determines when, if at all, this capability can contribute to security.

2.8. Risk to Loss of Supply All DNOs’ licences32 require them to “plan and develop (their) distribution

system(s) in accordance with a standard not less than that set out in Engineering Recommendation P2/5 (October 1978 revision) of the Electricity Council Chief Engineers’ Conference in so far as applicable to it or such other standard of planning as the licensee may, following consultation (where appropriate) with the transmission company and any 32 DTI - ELECTRICITY DISTRIBUTION LICENCE: STANDARD CONDITIONS, Condition 5, Distribution System Planning Standard and Quality of Service, taken from site:- http://www2.dti.gov.uk/energy/gas_and_electricity/regulation_policy/licences/sc2_elecdist.pdf

Page 61 of 150


other authorised electricity operator liable to be materially affected thereby and with the approval of the Authority, adopt from time to time.”

The sole objective of this planning standard, and presumably that of any revised

standard based on the same philosophy as P2/5 (namely the objective of the outcome of these studies), is to ensure there is an acceptable level of capacity needed to satisfy the group demand following first and maybe second order outages. This criterion is not directly related to reliability nor does it necessarily maintain specific levels of customer interruption indices. These indices can only be assessed by separate reliability studies, which in principle are outside of the scope of this project.

In reality the present overall performance of DNO distribution systems is

substantially in excess of the performance implied by the sole application of P2/5: this standard ensures a minimum security level for individual demand groups. Whilst any move to encourage DNOs to just meet P2/5 compliance could in principle result in a reduction in service to end customers, this is unlikely to happen because there are other regulatory and commercial mechanisms that will prevent a decline in system security. The recent IIP activities of OFGEM and DNOs illustrate how regulatory and commercial incentives are acting to improve system performance.

There is a view however, that the risk to supply (presumably meaning the impact on

customer reliability) could be degraded by using generation instead of circuits – this belief may exist because of the view that the reliability of generating units is less than that of overhead lines and underground cables. However the benefit of reinforcing a system using generation is clearly an alternative to constructing a new circuit or upgrading an existing one. Reinforcing with generation would undoubtedly increase reliability although perhaps by less than that due to an additional circuit. This relative benefit can be assessed by reliability studies.

Indeed, the existing P2/5 standard defines the contribution that conventional

generation makes to network security. It is important to bear in mind that there is no suggestion that applying the existing security standard to the network design, that includes conventional generation with characteristics as outlined in ACE 51, increases the risk of interruptions in supply to customers beyond acceptable levels. In other words, the risk of loss of supply that is inherent with the present security standard must be considered as acceptable and will be used as a reference.

Given that the proposed methodology, when applied to conventional generation

units with characteristics of those in the 1970s (availability of 86%), gives a contribution fully consistent with the existing standard (67%), the only question to be considered is whether the contributions associated with new generation technologies, as quantified by the proposed methodology, increase the risk of outages beyond that considered implicitly as acceptable by P2/5.

There is also a view that including generation into the assessment of security using

the principle of P2/5 would mean that the reliability would be less than that if generation exists but is not considered in determining whether the system is P2/5 compliant. This is essentially true but is somewhat misleading. Consider the following situations:-

Page 62 of 150


• two circuits exist but no generation: after a FCO the available capacity as per P2/5 is therefore equal to the cyclic rating of the remaining circuit. This is compared to that required by P2/5 (and presumably any revision of it)

• two circuits and some generation exist: after a FCO the available capacity as per P2/5 is therefore the summation of the cyclic rating of the remaining circuit and the contribution associated with the generation. This is compared to that required by P2/5 (and presumably any revision of it)

• two circuits and some generation exist but the latter is ignored: the available capacity is therefore again equal to that of the remaining circuit. This is compared to that required by P2/5 (and presumably any revision of it). In reality however, more capacity would be available because of the generation, meaning the reliability would be greater and the risk less. This generation then acts as insurance giving apparent increased comfort. This is true provided the generation company complies with providing security when demanded: without a contract to do so this may not be realisable. In any case it is misleading because it is tantamount to saying that the security levels specified by P2/5 are insufficient and greater capacity is required to provide the required degree of security. It also implies that group demands without generation can be allowed to have lower levels of security, which seems to be discriminatory.

It follows from the above discussion that:- •

•

•

•

a revised P2/5 and Table 2, i.e. a new P2/6, should not cause customer reliability to be less than that envisaged by P2/5, and therefore a DNO’s licence, as being acceptable however, if generation exists but is ignored in determining customer reliability levels, then the actual customer reliability may be in excess, even considerably in excess, of the minimum set by P2/5. If subsequently this generation is considered to have a security capability and used as a substitute to reinforcing the system with additional network capacity in a P2/5 compliance assessment, the level of customer reliability may well decrease, though still being in excess of P2/5 (or P2/6) requirements. This would not then preclude a DNO investing in the network to maintain existing reliability levels for IIP reasons, but would not be needed for compliance only if generation exists in a network, either its contribution to security should be included to maintain consistent treatment of all customers throughout the system, or the principles and criteria used in P2/5 should be completely changed. The latter is not the objective of the present review of P2/5, which is solely intended to be an update of Table 2 of P2/5, all principles and criteria remaining otherwise unchanged. A more fundamental review of P2/5 is possible and suitable approaches have been published previously33 by two of the authors of this report the replacement of P2/5 is intended to recognise the growth of modern types of generation, whilst maintaining the existing “back-stop” planning standards. Studies to assess the overall impact of adopting P2/6 on specific customer interruption indices were not part of the scope of this project,


Page 63 of 150


however it is not thought likely that the introduction of P2/6 will have any discernibly negative effect on aggregate distribution system performance.

2.9. Effect of Remote Generation

The topic of remote generation was included in the original set of proposed activities. However it was subsequently decided that this topic would be discussed and dealt with directly by Workstream 3, and not by the authors of this report.

2.10. Effect of Widespread Anticyclonic Weather

The topic of widespread anticyclonic weather was included in the original set of proposed activities. However it was subsequently decided that this topic would be discussed and dealt with directly by Workstream 3, and not by the authors of this report.

2.11. Impact of Changes in Demand Profiles During the Methodology project, it was observed that the shape of the LDC affected

the capability associated with the generation. It was decided at that time to leave any further study until the present project in order to perform the sensitivity studies using real data for the generating plant and for the LDC.

These studies are described in the section dealing with a complete range of

sensitivity studies, i.e. in Section 4.2.

2.12. Synthesising Forecast Profiles

2.12.1. Introduction

As discussed in Section 1.1, WS3 in association with FES decided to divide the overall activity into a number of stages. One stage related to the collection, collation and processing of real system data needed for input into the developed methodology. PPA was contracted to conduct this stage. The Final Report [“Data Collection for Revision of Engineering Recommendation P2/5”, FES Project K/EL/00303/05] was submitted in January 2004. This is referred to as the “Data Collection and Processing Project” or “Data Collection and Processing Report” in the present report.

The data, results and discussion included in the following sections are extracted from

this Data Collection and Processing Report and only provide a summary of its conclusions and recommendations. The data and information provided forms the basis of the development

Page 64 of 150


of the new tables for updating P2/5. For more detailed considerations, the reader of this report is referred to the Data Collection and Processing Report.

2.12.2. Need for Forecast Profiles

Table 2 of P2/5 applies to both established generation situations and to proposed generation situations. For both, it assumes that all generation units have individual availabilities of 86%, and that these units operate independently of one another. No seasonal variation in forced unavailability is assumed.

Recent reviews of the actual performance of embedded generation shows that

availabilities can deviate substantially from 86%, depending on the type and maturity of the plant, and sometimes its adopted operating regime; the latter being steered by the commercial market influences on operation, for example in the case of CHP and possible hydro plants.

The Data Collection and Processing Report considered the procedures that should be

defined for synthesising forecast profiles for proposed generation at the planning stage, and also the processes that might be prudent to use for the review of the security contributions of existing distributed generation (because of the possible changes in its operating and/or commercial environment). Recent considerations by the TSG of the commercial arrangements between distributed generators (DG) and DNOs also considered related issues of the effects of future electricity markets on the actual availability to be expected from DG at the planning and operating stages.

2.12.3. Plant Types Surveyed and Assessed

A considerable range of plant types was surveyed. The returns were substantial in some cases and poorer in others. It was deemed from the Data Collection and Processing project that sufficient data were obtained for initial assessments to be made of the following plant types:-

• landfill gas • CHP based on internal combustion engines • CHP based on gas turbines • CCGT plant • wind farms • small hydro plant • waste to energy plant.

At the outset of the Data Collection and Processing project, it was decided not to

consider biomass plant, microCHP or DCHP, tidal power, photovoltaic and diesel plant. Therefore these have not been considered separately in the development of the new “Table 2”. However, the generic tables suggested in Section 5 permit any of these types of plant to be assessed in terms of their security capability if an estimate of the unit availability is made.

Page 65 of 150


2.12.4. Standard Profiles The Data Collection and Processing project was able to conclude and recommend

the availability characteristics summarised in Table 11, with the proposal, agreed by WS3, that the values quoted should be used with the methodology to produce the new tables for P2/6.

It is pertinent to state that, because DNOs review their system security at least once

per year and report this implicitly with their Long Term Development Statements, the DNOs could simultaneously review the performance of any generation attached to various demand groups, and modify security assessments with time.

Following such reviews, the values shown in Table 11 may change over time as

plant types and technologies mature or emerge. This has some significance because it would be desirable for some of the current sample sizes to be increased and the central assumptions reviewed. In fact it would seem prudent to periodically review the continuing appropriateness of the assumptions underpinning P2/6.

Page 66 of 150


Table 11 – Proposed availability profiles for new “Table 2” of P2/6

Plant type Proposed availabilities to be used for P2/6

CCGT High technical availability for 132kV & 33 kV connected plant. Plant is likely to be operated as peaking /market venture plant and may not be imperative to operate at other times. It is proposed that an average availability of 90% could be used.

CHP Unlikely to be available to the system unless prices permit this – and recent experience is that this is not so. In the case of small CHP some spill is convenient, but generation is geared to heat load and spill output is variable, and usually lower at system peak times. It is proposed that:-

• for sewerage plant an average availability of 60% for spark ignition engine plant and 80% for gas turbine plant is used.

• for other established CHP plant an average availability of 80% could be used.

Landfill Gas Almost always operates at full capacity when technically available. It is proposed that an average availability of 90% is used.

Windpower Operates when energy available and no significant technical unavailability. Statistics suggest that there is no significant variation between small and large installations with respect to load factors. It is proposed that the profiles for three typical plants located throughout the UK are used and the output treated as that from intermittent generation.

Small Hydro Confidence in output depends mainly on water availability. Technical availability is very high. Most larger sites have water storage. Smaller hydro with no significant storage has a limited load factor, 20% being typical. It is proposed that the profiles for three typical plants located throughout the UK are used and the output treated as that from intermittent generation.

Waste to Energy

This plant is technically reliable when fully established, but there are numerous examples of poor early-years performance. Settled-down availabilities of 85% for the sample have been observed. This plant normally operates at full capacity under the current market situations. It is proposed that an average availability of 85% is used for established plant.

Photovoltaic, small DCHP, wave, tidal, and biomass power

These plants were not assessed.

Page 67 of 150


3. Analysis Package for Assessing Generation Capability

3.1. Overview One requirement of the project was to develop a software package that could

perform the required assessments using the agreed methodology. This was an essential step in order to analyse many generation scenarios and to perform a wide range of sensitivity studies. The output of these studies would provide the required input into the development of the new P2/6. In addition it was anticipated that there could be a need to perform special case studies in the future either to confirm previously tabulated values or to consider specific generation scenarios that were outside of the scope of the newly developed P2/6. P2/5 itself permits such special studies but does not provide the wherewithal of doing these. It was decided that the package need only be of a prototype form, not of a commercial grade and not commercially supported. Instead, the specification was to produce a simple spreadsheet application package that an engineer with a reasonably knowledge of the approach could use.

This prototype package has been developed in conjunction with Workstream 3. Its

objective is to calculate the capability contribution to security of supply from distributed generation connected to a particular demand group. This section of the report provides a short description of the package and a guide for using the package.

The application has been developed using Microsoft Excel® and Visual Basic for

Applications®. There are six worksheets: Contribution, Generators, Load, IntermediateResults, ShiftGeneration and About. The main output values are saved in the Contribution worksheet. There is also a Recalculate button that invokes a macro for calculating the contribution. The generation capacity outage probability table is formed during the calculation and saved in the IntermediateResults worksheet. The input data for non-intermittent and intermittent generation is entered in the Generators worksheet, while the load duration curve is inputted in the Load worksheet. The calculation of maximum additional contribution by generation not continuously available is done in the ShiftGeneration worksheet.

The calculation of the security contribution by generation continuously available is

explained in Section 3.2, and that by generation not continuously available in Section 3.3. Section 3.4 contains examples.

3.2. Security Contribution by Generation Continuously Available

3.2.1. Input

3.2.1.1. Non-intermittent Generation Figure 7 shows part of the Microsoft Excel worksheet that contains the required data

for non-intermittent generation. Generators are classified into groups of identical units. Two groups exist in this example. Each group has the following data:

Page 68 of 150


(i) an index or group number (ii) status of the group: on – the generator group is taken into account,

otherwise – not (iii) number of generators in the group (iv) number of states permitted in the capacity outage probability table,

maximum 10 (v) the capacity outage probability table of each unit in the group (the

capacity outage probability table of the complete group and of the system is evaluated by the analysis package)

Number of Generator Groups 2Index of Generator GroupStatus On / OffNumber of GeneratorsNumber of StatesState No Capacity Probability Capacity Probability

1 40 0.86 602 0 0.14 30 0.153 0 0.05456789

10

132

2Off

1

3On

Non-intermittent Generation

0.8

Edit Data

Figure 7 - Input data for non-intermittent generation

3.2.1.2. Intermittent Generation Figure 8 shows part of the worksheet that contains the data for intermittent

generation. This represents all the intermittent generation connected to a specific demand group. Persistence should be entered in three yellow fields (days, hours and minutes). In addition, values can be entered as a real number, i.e. value of 1.5 in field hours means 1 hour and 30 minutes. The persistence is shown as output values in the lower right hand box. If the value of Status field is On, then the intermittent generation is taken into account when the contribution is calculated, otherwise not. The Filename field represents the name of the file in which the chronological power generation profile is stored in the form of a column of values under the heading [Data]. An example is shown in Figure 9. The Installed Capacity field represents the total installed capacity of all intermittent generation connected to the demand group. The Interval field represents the time interval between two consecutive values. If these two fields are set to zero, the program uses the corresponding values supplied in [Header] part of the file shown in Figure 9. Finally, the Number of States field represents the number of states to be formed in the capacity outage probability table: a value of 101 was used for the case studies in Sections 4 and 5.

Page 69 of 150


Persistence Tm (days, hours, minutes) 1 days 6 hours 0 minutesStatus On / OffFilename Persistence TmNameplate Rating Value is given by file 1.25 daysInterval (min) Value is given by file 30 hoursNumber of States 1800 minutes

Intermittent Generation

10

OnWindFarm v2.txt

00

Edit Data

Figure 8 – Input data for intermittent generation

[Header] File Version = 2.0 Data Description = Intermittent Generation Test Data Nameplate rating = 100 MW Data Interval = 30 minute [Data] 20.320801 20.4331835 20.320801 21.0870454 …

Figure 9 – Example of file containing intermittent generation profile (the [Data] section is not complete, units are in MW)

3.2.1.3. Demand Figure 10 shows part of the Load worksheet which contains a piecewise linear

representation of the load duration curve. The peak demand is set equal to the total installed generation by the analysis package. The Duration field is the total duration of the period of interest. The value of 3624 hr represents a winter period of five months duration. The Number of Lines field represents the number of linear segments used to approximate the load duration curve. The maximum number of segments is nine. The normalised load duration profile is determined by this set of values. The first point of the load profile is always (0, 100%) and should not be changed. The following points are user-defined variables and should be set to ensure a best-fit to the load duration curve. The illustrated graph is deduced by the package.

Page 70 of 150


Peak Value 100Duration 3624Number of Lines 2

Duration Load Duration Load Energy0.00% 100.00% 0 100 1849502.76% 75.00% 100 75 1250

100.00% 25.00% 3624 25 931003624 0 90600

Energy is 184950

Normalised Load Profile

Load Duration Curve

0

20

40

60

80

100

120

0 500 1000 1500 2000 2500 3000 3500 4000

Duration (h)

Dem

and

(MW

)

Energy is 184950

Edit Data

Figure 10 - Input data for load duration curve

3.2.2. Output The evaluation results are saved in the Contribution worksheet shown in Figure 11.

In order to ensure consistency of results with the chosen input data it is strongly recommended that the Recalculate button is pressed. The Installed capacity and Peak demand fields represent the total installed capacity connected to the demand group and the peak demand respectively. The Contribution field is the evaluated generation contribution expressed in percentage and in MW. The Probability field of delivering contribution represents the cumulative probability of those states in the capacity outage probability table that are capable of supplying the evaluated contribution. The counterpart field, Probability of not delivering contribution, is the complementary value. The last field represents the ratio between the expected energy not supplied and the peak demand.

The generation capacity outage probability table of the system is shown in the

IntermediateResults worksheet and is not explained here. It should be noted that the maximum number of states in this table must not exceed 65000, which is unlikely to limit the number of generators.

Page 71 of 150


Installed capacity 220 MWPeak demand 220 MWContribution (%) 54.64%Contribution (MW) 120.21 MWProbability of delivering contribution 72.60%Probability of not delivering contribution 27.40%Expected energy not supplied per Peak demand 178.9140 MWh/MW

Edit DataRecalculate

Figure 11 – Output

3.2.3. Data Editing The data can be edited directly in the worksheets, but an alternative method is

explained in this section. The dialog box shown in Figure 12 is obtained by pressing the Edit button found on the Contribution worksheet. The text in this box explains the fields on the right hand side, which are the same as in the Contribution worksheet. If the check box next to the Recalculate button is selected, recalculation will occur automatically when the Update button is pressed. If the Recalculate button is pressed, before any updating, the original values in the fields on the right hand side are re-calculated. When the Update button is pressed all data inputted through the pages of the dialog box are saved in the appropriate worksheets.

Page 72 of 150


Figure 12 - The dialog box of Contribution page Figure 13 shows the dialog with the Non-intermittent Generation page. The data is

entered in the standard Windows way. Pressing Enter on the keyboard while editing data will have the same effect as pressing the Update button. It is important to note that a change in the Index field of a generator group automatically update the worksheet data for this group. The final field of the probability column is automatically calculated. If the value becomes negative the error message will appear.

Page 73 of 150


Figure 13 - The dialog box of Non-intermittent Generation page Figure 14 shows the dialog box of the Intermittent Generation page. The main data

are the same as in the worksheet. In addition the fields Description, Maximum, Minimum and Average generation output are shown. If the check boxes associated with the fields for Installed capacity and Interval are unchecked, the values for these parameters can be changed.

Figure 15 shows the dialog box of the Load page.

Page 74 of 150


Figure 14 – The dialog box of Intermittent Generation page

Figure 15 - The dialog box of the Load page

Page 75 of 150


3.3. Contribution by Generation Not Continuously Operational In order to reduce complexity, calculation of maximum additional contribution by

the generation not continuously available is done completely in the ShiftGeneration worksheet shown in Figure 16. The half-hourly typical load profile for a peak day is entered in the Demand (MW) column. The calculation can be done for two operating periods, e.g. 8hr and 12hr, simultaneously. The start time and duration is entered for each case in Time of Operation part of the worksheet respectively. If the value of Start time is Auto, the operational time is calculated automatically and infers the same demand at the beginning and the ending time of operation. The result of the calculation, obtained by pressing the Refresh button, is displayed in bold in Maximum Additional Contribution part of the worksheet. The Maximum contribution is given as a percentage of two different values. The first value is the percentage of the peak demand, and the second one is the percentage of the firm capacity; the latter being obtained by calculating the contribution from the network and the continuously available generation.

MW %0:30 35235.7 67.221% 11:00 35588.1 67.893% 21:30 35680.6 68.070%2:00 35296.7 67.337%2:30 35145.1 67.048%3:00 35515 67.754%3:30 34828.5 66.444%4:00 33973.8 64.814%4:30 33305.4 63.538%5:00 33040.5 63.033% Duration5:30 33120.2 63.185% hours Start Stop of peak of firm cap.6:00 33742 64.371% 8 12:00 20:00 11% 12%6:30 36227.4 69.113% 12 8:30 20:30 12% 14%

Typical Peak Load ProfileStartTime Demand Case

Time of OperationDuration

Maximum contributionTime

Please press the Refresh button before reading the results

Auto or hh:mmautoauto

812

hours

Maximum Additional Contribution

Refresh

Figure 16 – Part of ShiftGeneration worksheet

3.4. Examples

3.4.1. Non-intermittent Generation Example 1: Suppose that only non-intermittent generation is connected to the

demand group. There is a set of two generators with the following parameters: (i) rating capacity of each generator is 10 MW, and (ii) availability is 0.98.

Solution: Open Excel file Contribution and enable macros if asked so. Now the

ability to run macros should be tested by starting from menu Tools Macro Security. If the security level is high, the macros supplied in this file will not work. Choose medium security level to have ability to accept macros if you want to and close and reopen file. Do not forget to return your preferred settings of security level after finishing work on the file. Activate Generation worksheet and press Edit Data button next to the title Non-intermittent

Page 76 of 150


generation. It will activate dialog shown in Figure 13. In the Number of generators groups field enter 1. In the Index of generator group field enter also 1. Set Status to On. In the Number of generators field enter 2, and also in the Number of reliability states field. In the Capacity column enter numbers 10 and 0 respectively, and in the Probability column number 0.98. How it should look is shown in Figure 17 on the left. Choose Intermittent Generation page and in the Status set off. The other fields become irrelevant. Now press the Update button and after the Recalculate button. The contribution is 76.715 % for the load duration curve shown in Figure 15.

Figure 17 - Settings for example 1 Example 2: Suppose that additional generator is connected to the demand group

with the following states 10, 5 and 0 MW and probability of states 0.8, 0.15 and 0.05, respectively.

Solution: In addition to the solution of example 1 enter data as shown in Figure

18. Please ensure that the first pair of generators is on. The resulting contribution is 80.580 %.

Page 77 of 150


Figure 18 - Additional settings for example 2

3.4.2. Intermittent Generation Example 3: Suppose that only intermittent generation supplies the demand. The

total installed capacity is 100 MW, and average half hourly power output for a year is supplied in the file ‘WindFarm v2.txt’. The contribution is calculated for a persistence of 6 hr.

Solution: Select the Non-intermittent Generation page and in the Number of

generators group enter 0. After that, ensuring that the file ‘WindFarm v2.txt’ is saved in the same folder as Excel file Contribution.xls, in the Intermittent Generation page enter data as shown in Figure 19. Press Update and Recalculate buttons, respectively. The contribution is 28.623 %. Now check the calculation of contribution with different number of states. For 20 states contribution is 28.694 %. If the number of states is higher the calculation of the contribution becomes more precise.

Page 78 of 150


Figure 19 - Settings for example 3

3.4.3. Non-intermittent and Intermittent Generation Example 4: Suppose that the demand is supplied with all generators from the

previous examples. Solution: In the Non-intermittent Generation page return Number of generator

groups to 2. The contribution is now 42.519 %.

Page 79 of 150


4. Sensitivity Studies

4.1. Introduction A significant part of this project centred on performing sensitivity studies in order to

satisfy two objectives:- i) to obtain the relevant values needed to update P2/5 into proposals for P2/6 ii) to resolve a number of issues that were outside the scope of the previous

project but nevertheless were considered important. These studies were performed using the analysis package described in Section 3.

The results relating to objective (i) are included in Section 5, which presents the proposals for updating P2/5 into P2/6. The results relating to objective (ii) are included in this section.

Two main issues were studied: that associated with the effect of the shape of the

LDC, and that associated with the degree of resolution of the output profiles of the intermittent generating plant.

4.2. Effect of Shape of Load Duration Curve

4.2.1. Background During the Methodology project, it was observed that the shape of the LDC affected

the capability associated with the generation. It was decided to leave any further study until the present project in order to perform the sensitivity studies using real data for the generating plant and for the LDC.

Typical LDCs were obtained during the Data Collection and Processing project. The

average LDC has been used to obtain the proposed values for P2/6 as described in Section 5. This average was derived from nine individual LDCs. These individual LDCs have been used separately in these sensitivity studies, and are shown in Table 12 and Figure 20.

4.2.2. Non-intermittent Generation The LDCs shown in Table 12 and Figure 20 were used together with non-

intermittent generation having a variable number of units ranging from one to ten, and variable availabilities ranging from 5 to 98%. The results are shown as a function of number of units in Figures 21-24 and as a function of availability in Figures 25-31. The maximum value is the greatest value found for any one of the nine LDCs, the minimum is the least value found, and the average is the mean of the nine individual values. The LDC associated with the maximum (and similarly with the minimum) value is not necessarily the same for all situations of number of units and unit availability. The results however clearly indicate the range or scatter that could be expected in the values of contribution.

Page 80 of 150


Table 12 – Winter load duration curves

peak demand point knee demand point minimum demand point Winter34 LDC demand

% time point

% demand

% time point

% demand

% time point

% A 100 0 86.1 2.1 53.8 100 B 100 0 75.9 3.0 42.6 100 C 100 0 88.9 2.9 57.9 100 D 100 0 89.3 2.0 40.7 100 E 100 0 76.6 1.6 41.2 100 G 100 0 73.5 2.1 55.1 100 H 100 0 89.8 1.5 35.1 100 J 100 0 92.5 1.7 45.2 100 K 100 0 80.7 2.9 50.7 100

35%

40%

45%

50%

55%

60%

65%

70%

75%

80%

85%

90%

95%

100%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Duration (%)

Dem

and

(%)

A B C D E G H K L

Figure 20 – Normalised winter load duration curves

34 these correspond with the LDCs included in the Data Collection and Processing Project Report

Page 81 of 150


55%

60%

65%

70%

75%

80%

85%

Number of units

Con

trib

utio

n (%

)

Maximum 68% 73% 77% 79% 81% 82% 83% 83% 84% 84%

Average 63% 69% 73% 75% 76% 78% 79% 80% 81% 81%

Minimum 57% 63% 68% 69% 70% 71% 73% 74% 77% 78%

1 2 3 4 5 6 7 8 9 10

Figure 21 – Contribution by generating units with 90% availability

50%

55%

60%

65%

70%

75%

80%

Number of units

Con

trib

utio

n (%

)

Maximum 63% 68% 72% 74% 76% 77% 78% 79% 79% 80%

Average 58% 65% 69% 71% 72% 74% 74% 75% 76% 76%

Minimum 53% 60% 65% 66% 67% 68% 69% 70% 70% 71%

1 2 3 4 5 6 7 8 9 10


Page 82 of 150


45%

50%

55%

60%

65%

70%

75%

80%

Number of units

Con

trib

utio

n (%

)

Maximum 59% 64% 68% 70% 72% 73% 73% 74% 75% 75%

Average 54% 61% 65% 67% 69% 70% 71% 71% 72% 72%

Minimum 49% 57% 62% 63% 64% 65% 66% 67% 67% 68%

1 2 3 4 5 6 7 8 9 10


35%

40%

45%

50%

55%

60%

Number of units

Con

trib

utio

n (%

)

Maximum 44% 51% 54% 55% 56% 57% 57% 57% 58% 58%

Average 40% 48% 51% 52% 54% 54% 55% 55% 56% 56%

Minimum 36% 45% 48% 50% 51% 52% 53% 54% 54% 55%

1 2 3 4 5 6 7 8 9 10


Page 83 of 150


0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 20% 40% 60% 80% 100%

Availability (%)

Con

trib

utio

n (%

)

Average of Contribution Min of Contribution Max of Contribution

Number of units: 1

Figure 25 – Contribution by one generating unit

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 20% 40% 60% 80% 100%

Availability (%)

Con

trib

utio

n (%

)


Number of units: 2

Figure 26 – Contribution by two generating units

Page 84 of 150


0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 20% 40% 60% 80% 100%

Availability (%)

Con

trib

utio

n (%

)


Number of units: 3

Figure 27 – Contribution by three generating units

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 20% 40% 60% 80% 100%

Availability (%)

Con

trib

utio

n (%

)


Number of units: 4

Figure 28 – Contribution by four generating units

Page 85 of 150


0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 20% 40% 60% 80% 100%

Availability (%)

Con

trib

utio

n (%

)


Number of units: 6

Figure 29 – Contribution by six generating units

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 20% 40% 60% 80% 100%

Availability (%)

Con

trib

utio

n (%

)


Number of units: 8

Figure 30 – Contribution by eight generating units

Page 86 of 150


0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 20% 40% 60% 80% 100%

Availability (%)

Con

trib

utio

n (%

)


Number of units: 10

Figure 31 – Contribution by ten generating units It can be observed from Figures 21-31 that the shape of the LDC affects the

contribution delivered by generation; the effect being greater with increasing unit availability up to about 90% and with decreasing number of units. Although this may seem to be an effect that should be included as a parameter in the development of P2/6, it should be noted that the variation between minimum and maximum is, in most cases, of the same order or less than the variation in contribution caused by practical ranges of unit availability; this effect also being indicated clearly in Figures 21-31. It should also be noted that the average contribution, evaluated as the mean of the nine individual contributions, is almost the same as the contributions given by using the average LDC (see Section 5).

The reason for the scatter to be greater for a smaller number of units is that, in such

cases, the loss of a small number of units represents a much greater change in percentage capacity and therefore a greater drop down the LDC. The unit availability has a direct impact on the likelihood of this loss and therefore this parameter also affects the degree of change and therefore the scatter.

It can be recommended from all the above observations that an average LDC should

be used together with an average value of unit availability in the development of a “Table 2” similar to that of Table 2 in P2/5. This is done in Section 5 of this report. If this assumption is considered unacceptable in specific situations, then such situations should be considered special cases and assessed individually using the graphical or analysis-package approaches also discussed in Section 5.

Page 87 of 150


4.2.3. Intermittent Generation

4.2.3.1. Wind Farms The LDCs shown previously in Table 12 and Figure 20 were used as the basis of

these studies. The average LDC obtained from these individual nine LDCs was also evaluated. This average is used subsequently in the determination of the “new Table 2” and is shown in Table 24 of Section 5.

This average LDC was used together with the three wind profiles supplied by the

Data Collection and Processing project, to give the individual wind farm security contributions and load factors shown in Table 13.

Table 13 – Contribution and load factors of three individual wind farms

Tm, hr wind farm

0.5 2 3 18 24 120 360 > 360

load factor

Eastern 29 27 26 18 15 2 0 0 33

Western 30 28 27 18 15 1 0 0 33

Northern 27 25 24 14 11 0 0 0 34

average35 29 27 26 17 13 1 0 0 The percentage contributions given by these three wind farms and the nine

individual LDCs were evaluated. The results are shown in tabular form in Table 14 and in graphical form in Figure 32. As in the case of non-intermittent generation, the maximum value is the greatest value found for any one of the nine LDCs, the minimum is the least value found, and the average is the mean of the nine individual values. Also, the LDC associated with the maximum (and similarly with the minimum) value is not necessarily the same for all situations. The results again clearly indicate the range or scatter that could be expected in the values of contribution. In the case of wind generation, this scatter is seen to be small, as indicated by the small values of standard deviation.


be used together with the average contribution of the three wind profiles in the development of a “Table 2” similar to that of Table 2 in P2/5. This is done in Section 5 of this report. If this assumption is considered unacceptable in specific situations, then such situations should be considered special cases and assessed individually using the graphical or analysis-package approaches also discussed in Section 5.

35 The calculation of average value is done using CPU representation of the decimal numbers not their integer representation given in the Table.

Page 88 of 150


Table 14 - Percentage contributions by wind farms

Tm hr

maximum %

average %

minimum %

standard deviation %

0.5 30 28 27 1.02 1 30 28 27 0.98 2 28 27 26 0.91 3 27 26 25 0.86 4 26 25 24 0.80 5 26 24 23 0.74 6 25 24 23 0.69 8 23 22 21 0.60

10 22 21 20 0.53 12 20 20 19 0.47 18 17 17 16 0.32 24 14 13 13 0.21 36 10 10 10 0.12 48 7 7 7 0.05 72 4 4 4 0.00 96 2 2 2 0.00

120 1 1 1 0.00 144 1 1 1 0.00 168 0 0 0 0.00

Page 89 of 150


0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

30.0%

0 24 48 72 96 120 144 168 192

Persistence Tm (h)

Con

trib

utio

n (%

)

Max Average Min

Figure 32 – Contribution by wind farms

4.2.3.2. Small Hydro The average LDC was used together with the three small hydro plant profiles

supplied by the Data Collection and Processing project. The individual plant security contributions and the load factors are shown in Table 15.

Table 15 – Contribution and load factors of three individual small hydro plants

Tm, hr small

hydro 0.5 2 3 18 24 120 360 > 360 load

factor

hydro A 22 22 22 18 17 7 0 0 32

hydro B 34 34 33 33 32 20 7 0 51

hydro C 54 54 54 52 52 48 32 0 70

average 37 36 36 34 34 25 13 0

Page 90 of 150


The percentage contributions given by these three small hydro plants and the nine individual LDCs were evaluated. The results are shown in tabular form in Table 16 and in graphical form in Figure 33. The maximum, minimum and average values are determined as before. The results again clearly indicate the range or scatter that could be expected in the values of contribution. In the case of small hydro generation, this scatter is again seen to be small, as indicated by the small values of standard deviation.

Table 16 - Percentage contributions by small hydro plant

Tm hr

maximum %

average %

minimum %

standard deviation %

0.5 40 37 34 2.10 1 40 37 34 2.10 2 40 37 34 2.10 3 40 36 34 2.10 4 40 36 33 2.07 5 39 36 33 2.05 6 39 35 33 2.04 8 39 35 33 2.04

10 39 35 32 2.04 12 38 35 32 2.03 18 38 34 32 2.00 24 37 34 31 1.96 36 35 32 30 1.85 48 34 31 29 1.80 72 33 30 28 1.75 96 30 28 25 1.60

120 27 25 23 1.44 144 26 23 22 1.35 168 23 21 19 1.18 192 22 20 19 1.19 216 22 20 18 1.18 240 21 19 17 1.11 264 19 18 16 1.00 288 16 15 14 0.80 312 14 13 12 0.65 336 14 13 12 0.65 360 14 13 12 0.59 384 13 12 11 0.45


be used together with the average contribution of the three small hydro profiles in the

Page 91 of 150


development of a “Table 2” similar to that of Table 2 in P2/5. This is done in Section 5 of this report. If this assumption is considered unacceptable in specific situations, then such situations should be considered special cases and assessed individually using the graphical or analysis-package approaches also discussed in Section 5.

10.0%

15.0%

20.0%

25.0%

30.0%

35.0%

40.0%

0 24 48 72 96 120 144 168 192 216 240 264 288 312 336 360 384

Persistence Tm (h)

Con

trib

utio

n (%

)

Max Average Min

4.3. Effect of Resolution of Wind Generation Profiles

As discussed in Section 2.6.3, the output of an intermittent generation plant, such as a wind farm, is likely to vary during each half hour period. This variation in generation output levels would need to be absorbed by the remaining circuits. For a short period of time, the generation output could drop significantly and hence the remaining circuits may become overloaded. This may not be of great significance if the periods are very short in duration, but could be quite the reverse if they are long.

These times are relative to the system being considered and therefore the assessment

may need to be system dependent. For example, the diversity in output of wind farms with a large number of wind turbines will be greater that those with a small number of wind turbines. Similarly, the positive effects of increased diversity will be prominent for a group demand being supplied by multiple wind farms.

Figure 33 - Contribution by small hydro plant

Sensitivity studies were carried out for all seven wind farms for which data was available in order to compare the capabilities associated with 30 min, 5 min and 1 min

Page 92 of 150


resolutions of the output profiles. The results for each individual site are presented graphically in Figures 34 to 40.

0%

5%

10%

15%

20%

25%

30%

35%

40%

0 12 24 36 48 60 72 84 96 108 120 132 144 156 168

Persistence Tm (h)

Con

trib

utio

n (%

)

1 minute profile 5 minute profile 30 minute profile

Figure 34 – Contribution by wind farm at site A

0%

5%

10%

15%

20%

25%

30%

0 12 24 36 48 60 72 84 96 108 120 132 144 156 168

Persistence Tm (h)

Con

trib

utio

n (%

)


Figure 35 – Contribution by wind farm at site B

Page 93 of 150


0%

5%

10%

15%

20%

25%

30%

0 12 24 36 48 60 72 84 96 108 120 132 144 156 168

Persistence Tm (h)

Con

trib

utio

n (%

)


Figure 36 – Contribution by wind farm at site C

0%

5%

10%

15%

20%

25%

30%

35%

0 12 24 36 48 60 72 84 96 108 120 132 144 156 168

Persistence Tm (h)

Con

trib

utio

n (%

)


Figure 37 – Contribution by wind farm at site D

Page 94 of 150


0%

5%

10%

15%

20%

25%

30%

35%

40%

0 12 24 36 48 60 72 84 96 108 120 132 144 156 168

Persistence Tm (h)

Con

trib

utio

n (%

)


Figure 38 – Contribution by wind farm at site E

0%

5%

10%

15%

20%

25%

30%

0 12 24 36 48 60 72 84 96 108 120 132 144

Persistence Tm (h)

Con

trib

utio

n (%

)


Figure 39 – Contribution by wind farm at site F

Page 95 of 150


0%

5%

10%

15%

20%

25%

30%

35%

0 12 24 36 48 60 72 84 96 108 120 132 144

Persistence Tm (h)

Con

trib

utio

n (%

)


Figure 40 – Contribution by wind farm at site G As 30 min resolution data, rather than 5 min or 1 min, is expected to be readily

available, it is proposed to use the 30 min data to calculate capability factors and then use appropriate correction factors to scale these values into the capabilities that would correspond to 5 min and 1 min data. The factors needed to convert the 30 min resolution data into 5 min or 1 min capabilities were calculated. The average percentage correction factors for various values of Tm are presented in Table 17.

Table 17 – Correction factors from 30min data for single wind farm

correction factors in % for Tm (hr) of: wind profile resolution ½ 2 3 18 24 120

5 min 97 93 92 85 84 42

1 min 93 88 87 75 72 31

Furthermore, the average capabilities given in Table 13 can be multiplied by these

correction factors to give corresponding capabilities for 5 and 1 min resolutions. These results are summarised in Table 18. As expected, the contribution reduces as the time resolution becomes more refined, and the differences in contributions become relatively larger as Tm is increased.

Page 96 of 150


Table 18 – F factors for intermittent generation for various time resolutions

The effect of increased diversity has been evaluated by considering two wind farms that feed into a single demand group. This is shown in Figure 41 which gives the contribution for resolutions of 30 min, 5min and 1 min for the combined sites given individually in Figures 39 and 40; the data being scaled before combining in order to maintain the same ratings. The factors needed to multiply the average capabilities of these two sites for 5 min and 1 min resolutions and various Tm in order to obtain the group capabilities were calculated. These factors were then multiplied by the corresponding correction factors in Table 17 and the new factors are presented in Table 19. As expected, the correction factors for a single wind farm are lower than the combined values.

0%

5%

10%

15%

20%

25%

30%

35%

0 12 24 36 48 60 72 84 96 108 120 132 144

Persistence Tm (h)

Con

trib

utio

n (%

)


Figure 41 – Contribution by two wind farms (site F and site G)

Tm, hr wind profile resolution ½ 2 3 18 24 120 360 >360

30 min 29 27 26 17 13 1 0 0

5 min 28 25 24 14 11 0 0 0

1 min 27 24 23 12 10 0 0 0

Page 97 of 150


Table 19 – Correction factors from 30min data for two wind farms

correction factors in % for Tm (hr) of: wind profile resolution ½ 2 3 18 24 120

5 minute 97 95 94 88 87 -36

1 minute 95 91 90 79 76 -

Since the thermal time constants of overhead lines are in the order of several minutes, it is proposed to use 5 minute resolution for quantifying the contribution of wind generation, and this approach is used in the derivation of the “new Table 2” in Section 5 of this report. For circuits composed of cables and, in particular transformers, it may be sufficient to use 30 min resolution. The correction factors being proposed are shown in Table 20. Finally, the average values in Table 13 are multiplied by the factors in Table 20 and these values are those used in the “new Table 2” (see Tables 27 and 29 of Section 5).

Table 20 – Correction factors from 30min data for single and multiple sites

correction factors in % for Tm (hr) of: wind profile

½ 2 3 18 24 120

one site 5 min 97 93 92 85 84 42

multi- sites 5 min 97 95 94 88 87 4237

It should be noted that the application of correction factors is not appropriate if the

analysis package is being used to evaluate the contribution of generation systems that include a combination of various technologies together with generation. For such analysis, the actual 5 min data will need to be available. However, the set of factors presented in Table 20 can readily be used if a table or/and graphical approach is used.

The methodology for determining the security capability of a group of units strictly

means that all units of that group should be considered simultaneously, because the capability of a group depends on the number of units in the group as well as the availability of the units. This causes no problem if all the units in a group are identical because there is only one capability value for a given number of units and unit availability. Therefore a simple two-dimensional table can be constructed for all combinations of unit numbers and unit availabilities. This however is not possible if the units are non-identical since this then

4.4. Aggregation of Multiple Generating Groups

36 There is insufficient data to determine values of these fields 37 There is insufficient data to determine value of this field, hence the value for the single site is used.

Page 98 of 150


involves an infinite (or extremely large) number of possible combinations. Two alternatives are possible:-

•

•

to obtain a precise result: treat the situation as a special study and use the analysis package to obtain an approximate result: summate the capabilities of each unit, or each identical group of units, using tabulated values.

The first approach clearly will give precise values but requires the use of the analysis

package in each case. This approach can be defined as the aggregation approach since it aggregates the effect of all units simultaneously. The second approach is much simpler but clearly will only give approximate values since the aggregation process is a non-linear summation. This approach can be defined as the summation approach.

The degree of approximation was assessed by comparing the two approaches for a

range of case studies. The results are shown in Tables 21 and 22. Each unit in all cases had a capacity of 10MW. Table 21 shows the results using aggregation and summation for two groups with all units in both groups having the same value of availability. Table 22 shows the results for two groups with the units in each group having different values of availability.

It can be observed that the summation results are always less than those given by the

aggregation approach and therefore its use would always give results that err on the side of caution. In addition, the difference between the two values is generally small particularly as the number of units is increased. Since the aggregation approach increases the degree of precision and not necessarily the degree of accuracy, these results indicate that, provided the number of separate groups is relatively small, the simplicity of the summation approach justifies its use for evaluating the security capability of groups having non-identical units.

It is therefore suggested that the summation approach be used for most case studies

and that the aggregation approach be used only for those cases in which greater precision is deemed necessary, or the number and diversity of the groups give cause for concern about whether the summation approach is sufficiently precise.

Page 99 of 150


Table 21 – Aggregation of groups of units with same availabilities

summation aggregation difference first group second group

total no. of

units

avail’y % no. of

units

contrib-ution

%

firm capacity

MW

no. of units

contrib-ution

%

firm capacity

MW

total firm capacity

MW

contrib-ution %

firm capacity

MW MW %

90 1 63 6.3 1 63 6.3 12.6 69 13.8 -1.2 -8.70 2 60 1 40 4.0 1 40 4.0 8.0 48 9.6 -1.6 -16.67 90 1 63 6.3 2 69 13.8 20.1 73 21.9 -1.8 -8.22 3 60 1 40 4.0 2 48 9.6 13.6 51 15.3 -1.7 -11.11

1 63 6.3 3 73 21.9 28.2 75 30.0 -1.8 -6.00 90 2 69 13.8 2 69 13.8 27.6 75 30.0 -2.4 -8.00 1 40 4.0 3 51 15.3 19.3 52 20.8 -1.5 -7.21

4 60

2 48 9.6 2 48 9.6 19.2 52 20.8 -1.6 -7.69 1 63 6.3 4 75 30.0 36.3 77 38.5 -2.2 -5.71 90 2 69 13.8 3 73 21.9 35.7 77 38.5 -2.8 -7.27 1 40 4.0 4 52 20.8 24.8 53 26.5 -1.7 -6.42

5 60

2 48 9.6 3 51 15.3 24.9 53 26.5 -1.6 -6.04 1 63 6.3 9 80 72.0 78.3 80 80.0 -1.7 -2.13 2 69 13.8 8 79 63.2 77.0 80 80.0 -3.0 -3.75 3 73 21.9 7 79 55.3 77.2 80 80.0 -2.8 -3.50 4 75 30.0 6 78 46.8 76.8 80 80.0 -3.2 -4.00

90

5 77 38.5 5 77 38.5 77.0 80 80.0 -3.0 -3.75 1 40 4.0 9 56 50.4 54.4 56 56.0 -1.6 -2.86 2 48 9.6 8 55 44.0 53.6 56 56.0 -2.4 -4.29 3 51 15.3 7 55 38.5 53.8 56 56.0 -2.2 -3.93 4 52 20.8 6 54 32.4 53.2 56 56.0 -2.8 -5.00

10

60

5 53 26.5 5 53 26.5 53.0 56 56.0 -3.0 -5.36 1 63 6.3 10 80 80.0 86.3 80 88.0 -1.7 -1.93 2 69 13.8 9 80 72.0 85.8 80 88.0 -2.2 -2.50 3 73 21.9 8 79 63.2 85.1 80 88.0 -2.9 -3.30 4 75 30.0 7 79 55.3 85.3 80 88.0 -2.7 -3.07

90

5 77 38.5 6 78 46.8 85.3 80 88.0 -2.7 -3.07 1 40 4.0 10 56 56.0 60.0 56 61.6 -1.6 -2.60 2 48 9.6 9 56 50.4 60.0 56 61.6 -1.6 -2.60 3 51 15.3 8 55 44.0 59.3 56 61.6 -2.3 -3.73 4 52 20.8 7 55 38.5 59.3 56 61.6 -2.3 -3.73

11

60

5 53 26.5 6 54 32.4 58.9 56 61.6 -2.7 -4.38

Page 100 of 150


Table 22 – Aggregation of groups of units with different availabilities

summation aggregation difference first group

availability 90% second group

availability 60% total

no. of units no.

of units

contri-bution

%

firm capacity

MW

no. of

units

contri-bution

%

firm capacity

MW

total firm

capacity MW

contri-bution38

%

firm capacity

MW MW %

2 1 63 6.3 1 40 4 10.3 58.3 11.66 -1.36 -12 1 63 6.3 2 48 9.6 15.9 58.3 17.49 -1.59 -9 3 2 69 13.8 1 40 4 17.8 66.1 19.83 -2.03 -10

38 The values are calculated by the software program.

Page 101 of 150


5. Development of Draft of New “Table 2”

5.1. Possible Formats of Table 2 During the development of the methodology, and discussed in Section 1.5 of this

report, three main approaches have been considered for the implementation of a new “Table 2”. These are: a look-up table(s) approach in a form similar to that of the current Table 2 in P2/5; a graphical approach featuring a family of graphs/figures; and a computerised approach based on a spreadsheet environment. Each of these has merits and demerits:-

look-up table(s) approach: o merits: simplicity, free of erroneous application, restricted choices o demerits: restricted choices, approximate for non-identical generating

units, typical or average data and generation types graphical approach:

o merits: reasonable simplicity, wider range of data and choices o demerits: more complex to use, increased uncertainty in application,

approximate for non-identical generating units computerised approach:

o merits: unrestricted choice of data, mixed generation types o demerits: most complex to use, results possibly dependent on user.

At the time of the Methodology project39, it was decided to leave the decision about which approach to use until the implementation stage, i.e. to this project. It now seems preferable that all three approaches should be implemented in some form, as follows:-

look-up table(s) approach. Two forms of this tabular approach are required. A complete version similar to Table A.1 of ACE Report 51 in which either the circuit or the generation may be dominant, and a simpler version similar to Table A.2 of ACE Report 51 (identical to Table 2 of P2/5) in which the circuits are dominant. The latter requires materiality conditions similar to those in paragraph 3.6 of P2/5. Each row in each of these two tables will correspond to a particular generation type. These tables could be used as follows:

o the second table together with appropriate materiality conditions should be part of the main text of P2/5. This table could be expected to be sufficient for the majority of cases, particularly those for which full data sets of specific situations are not available, either because the system is not yet in an operational state or the operating life is too short to allow confidence in the statistical data of the plant

o the first table would be applicable if it was decided that the materiality test was not to be imposed in a particular situation. This may be the case when confidence in the statistics of the operational behaviour of the plant is high, or when the system is not compliant

•

•

•

•


Page 102 of 150


with the materiality test but is compliant if the test is removed and consequently could allow extended derogation. Whether this second table should be attached as an appendix to P2/6 or only inserted into supporting documents (presumably retained/published by Ofgem/DTI/ENA40) is outside the scope of this project and should be decided by WS3 or other appropriate body

•

•

graphical approach: o this will be the outcome of extended sensitivity studies and will

consist of a number of graphs and/or tables. These will still be look-up in concept, but will be generic and will apply to any generation type: for instance, the data required for non-intermittent generation will be the number of units and unit availability, rather than generation type

o whether these should be attached as appendices to P2/6 or retained in supporting documents again remains the responsibility of WS3 or other appropriate body

computerised approach: o a computerised package will exist which can be used for special cases

not considered in either of the above two approaches. This is particularly of importance when mixed generation types exist, or new technologies are being designed

o this computerised package requires central control to ensure that all users continue to use the same version and to receive any updates. This is best done by either Ofgem or ENA, but the decision remains that of WS3 or other appropriate body.

5.2. Generation Technologies and Data Considered

The only generation technologies that could be considered specifically in this project were those that the Data Collection and Processing Report identified as potential security sources and proposed sufficient data to make the assessments credible. This does not mean that other sources cannot be considered: instead such sources would need to be treated as special cases. This can be achieved in one of two ways. Either use the data provided in look-up tables for specific types of generating plant if it is known both have similar characteristics, or use one of the two generic approaches together with some flexibility of input data if numerical values are uncertain.

All the following technologies and data are those proposed in the Data Collection and Processing Report and accepted and agreed by WS3. The specific technologies considered are shown in Table 23. The winter load duration curve used is shown in tabular form in Table 24; this being the average curve of the nine individual LDCs used in Section 4. The values of Tm used correspond with the values discussed in Section 2.6.2.6 and shown previously in Table 10. These are also shown in Table 23.

40 ENA (Electricity Networks Association) formerly EA (Electricity Association)

Page 103 of 150


Table 23 – Generation technologies considered and data used

non-intermittent generation average availability, %

landfill gas 90

CCGT 90

sewerage: spark ignition 60

sewerage: GT 80 CHP

other CHP 80

waste to energy 85

intermittent generation output profile

wind average 6-month winter profile for three sites ½ hr and 1 min resolutions

small hydro average 6-month winter profile for three sites ½ hr resolution

values of Tm used ½, 2, 3, 18 and 24 hr 5 days, and more than 5 days

Table 24 – Average winter load duration curve used

load level, % time point, %

100 0

83.9 2.2

46.9 100 The typical winter daily load curve used to assess the contribution made by

generating units operating for less than 24 hr per day is shown in Figure 42. It may be questioned whether this is typical for most situations and the curve previously used in ACE Report 51, and hence in P2/5, was also considered. However it should be noted that the curve shown in Figure 42 gives smaller percentage contributions than those in P2/5, and hence, since it is best to err on the side of caution, it is judged by the authors of this report and by WS3 that the curve in Figure 42 is the most reasonable to use.

Page 104 of 150


60%

65%

70%

75%

80%

85%

90%

95%

100%

0 2 4 6 8 10 12 14 16 18 20 22 24Time (h)

Load

dem

and

(%)

Figure 42- Typical winter daily load curve

5.3. Look-Up Table Approach

5.3.1. Format of Base Table As in the case of P2/5, “Table 2” is a derived table. The base table is Table A.1 of

ACE Report 51, although this may not be evident from P2/5 itself. Therefore in any revision of P2/5, the first requirement is to establish the base table. In the development of P2/5, only one value of availability was considered and the effect of number of units and plant technology on capability was neglected. Consequently, only one base table was required. This is not the case in the present case because unit technology, unit availability and unit numbers all affect the capability. Consequently a series of tables is required. These tables are shown below.

The factors F in these tables are determined from the methodology using the analysis

package described in Section 3 and described in detail in the Methodology Report. The procedure for evaluating the “additional generation” contributions shown in columns 3 and 4 is also described in the Methodology Report. For the values shown, it is assumed that the associated plant runs for an operating period such that the load level at start-up is the same or almost the same as at shut-down. As shown in Figure 43 for an operating period of Top, this load level is x% of peak demand. If this assumption is not valid, the contribution made by such plant will be smaller and a special-case assessment should be made.

Page 105 of 150


5.1

daily load curve

load

time, h

Top

x % of peak demand

Figure 43 – Contribution of generation operating for less than 24h Two sets of values are shown in columns 3 and 4 of Table 25. The first value of each

set represents the contribution expressed as a percentage of peak demand; this corresponds with the values quoted in P2/5. The value shown in parentheses is the same contribution but expressed as a percentage of the total remaining circuit capacity plus the effective capacity of any generation operating for 24hr per day (defined as “firm capacity” in ACE Report 51); this value corresponds with the value shown in Table A.1 of ACE Report 51. The two values of each set are related as follows.

The capacity of any remaining circuits plus the effective capacity of any generation

operating for 24 hr per day must be sufficient to meet the demand at the times when the (<24hr) generation is not operating, i.e. it must meet the x% demand level shown in Figure 42. The firm capacity equals [x% of peak demand]. The first value of each set shown in columns 3 and 4 of Table 42 is equal to [(100-x)% of peak demand]. Therefore contribution of additional contribution is given by either:-

(100-x)% of peak demand or (100-x)/x % of firm capacity. Although P2/5 only includes the first of these values, the second one is easier to

determine since it depends on the network capacities (part of the assessment process) rather than the peak demand (against which the result is to be compared). The authors therefore believe both values should be included. However, if only one is judged to be necessary, then it would be best to include the second value (as in ACE Report 51) rather than the first (as in P2/5).

Page 106 of 150


The number of generators to outage in a FCO is determined from the probability criterion described in Section 2.3.5. The number of generators to outage in a SCO is, as in ACE Report 51, equal to one more than the number for a FCO, i.e. a maintenance outage in addition to the FCO.

Table 25 – Format of new “Table A.1” 1 2 3 4

MAXIMUM ADDITIONAL CONTRIBUTION IN % BY

GENERATION NOT CONTINUOUSLY AVAILABLE

OUTAGE CONDITION

SYSTEM CAPACITY INCLUDING GENERATION AVAILABLE FOR 24H

for 8hr for 12hr

no outage

(a) total transmission/distribution cyclic capacity plus (b) maximum sent-out capacity of all generation units x F = SC0

11 (12)

12 (14)

first circuit outage

SC0 minus the larger of (c) cyclic capacity of any transmission/distribution circuit or (d) maximum sent-out capacity of any N1 generation units x F = SC1

11 (12)

12 (14)

second circuit outage

SC0 minus the largest of (e) the sum of (c) and (d) above or (f) cyclic capacity of any two transmission/distribution circuits or (g) maximum sent-out capacity of any N2 generation units x F = SC2

11 (12)

12 (14)

where:- F = appropriate value from Table A.1a or Table A.1b N1 = appropriate value from Table A.1c N2 = N1 + 1 the percentages in columns 3 and 4 represent the % of group demand and, in

parentheses, the % of firm capacity; the latter being the total effective capability of the remaining circuit capacity plus the effective capacity of any generation operating for 24hr per day. These two values are equivalent.

the percentages in columns 3 and 4 are the maximum contributions because it is assumed that the operating period is such that operation spans the peak demand, and the demand at start-up is the same as the demand at shut-down, i.e. operation is symmetrically placed on the daily load curve. If these conditions do not apply, the contribution will be less

Page 107 of 150


Table 26 – “Table A.1a”, F factors in % for non-intermittent generation

number of units type of generation

1 2 3 4 5 6 7 8 9 10+

landfill gas 63 69 73 75 77 78 79 79 80 80

CCGT 63 69 73 75 77 78 79 79 80 80

sewerage: spark ignition 40 48 51 52 53 54 55 55 56 56

sewerage: GT 53 61 65 67 69 70 71 71 72 73 CHP

other CHP 53 61 65 67 69 70 71 71 72 73

waste to energy 58 64 69 71 73 74 75 75 76 77

Table 27 – “Table A.1b”, F factors in % for intermittent generation

Tm, hr type of generation

½ 2 3 18 24 120 360 >360

single site 28 25 24 14 11 0 0 wind multiple

sites 28 25 24 15 12 0

small hydro 37 36 36 34 34 25 13 0

0

Table 28 – “Table A.1c”, Number of generators (N1) contributing to FCO

number of units

type of generation 1 2 3 4 5 6 7 8 9 10+

landfill gas 1 2 2 2 2 2 3 3 3 3

CCGT 1 2 2 2 2 2 3 3 3 3


sewerage: GT 1 2 2 3 3 3 4 4 4 4 CHP

other CHP 1 2 2 3 3 3 4 4 4 4

waste to energy 1 2 2 2 3 3 3 3 4 4

wind all, i.e. complete wind farm

small hydro all, i.e. complete site

Page 108 of 150


5.3.2. Format of New “Table 2” According to P2/5, generation should not be dominant and consequently materiality

criteria were introduced to prevent this happening. These criteria have been discussed in Section 2.3.1 and a revised set compatible with the methodology was introduced in Section 2.3.5.3. These are:

3.x The contribution of generation specified in Table 2 (of P2/6) is based on the

assumptions that:- (a) the cyclic rating of the largest distribution circuit is greater than F% of the

total sent-out capacity of the N1 largest generating units and (b) the cyclic rating of the two largest distribution circuits is greater than F% of

the total sent-out capacity of the N2 largest generating units

In the application of this revised set of criteria, the F factors are those specified in Tables 25-27 (“Tables A.1, A.1a and A.1b”) and N1 and N2 largest generating units correspond to the number of generating units to remove for a FCO (N1) and a SCO (N2) given in Table 28 (“Table A.1c”). On this basis, the base table shown in Table 25 (“Table A.1”) simplifies to that shown in Table 29 (“Table A.2”).

If the “materiality” test is not satisfied, then the generation would become the most

significant FCO. If this is unacceptable for technical and/or commercial reasons, then P2/5 states that separate risk and economic studies may need to be undertaken. However, a more practical alternative would be to use a process of “capping” the generation. This can be achieved by evaluating the maximum capacity that each generating unit could have such that the materiality criterion is satisfied.

For instance in the case of identical generating units, if the cyclic rating of the

largest distribution circuit is Cc, the number of units to consider is N1 and the contribution factor is F, then the sent-out capacity of each generating unit Cg should not exceed:-

from first materiality condition Cg < Cc/F.N1 from second materiality condition Cg < 2.Cc/F.(N1+1)

and the smaller value of Cg should be used to determine the security contribution of the generating units. It should be noted that this does not restrict the actual capacity of the units, nor the amount of energy they may output into the system, it only restricts the amount of capacity that is credited to them for the purposes of security assessment.

Page 109 of 150


Table 29 - Format of new “Table A.2”

type of generation

contribution after first circuit outage classes of supply A-E

contribution after second circuit outage classes of supply D and E only

notes

landfill gas F % of DNC F % of DNC

CCGT F % of DNC F % of DNC

sewerage: spark ignition F % of DNC F % of DNC

sewerage: GT F % of DNC F % of DNC CHP

other CHP F % of DNC F % of DNC

waste to energy F % of DNC F % of DNC

wind F % of nameplate rating

F % of nameplate rating

small hydro F % of nameplate rating


the values in these two rows apply to the complete site irrespective of number of units

plant operating for 8 hr

SMALLER OF above value OR 11 % of group demand (12 % of firm capacity)



SMALLER OF above value OR 12 % of group demand (14 % of firm capacity41)


the values in these two rows assume that the operating period is such that operation spans the peak demand, and the demand at start-up is the same as the demand at shut-down, i.e. operation is symmetrically placed on the daily load curve. If these conditions do not apply, the contribution could be optimistic, in which case the generation ought to be treated as a special case. At one extreme, the contribution would be zero if the operating period did not span the peak demand at all.

NOTES all the values in this table apply from the point of time when the generation is connected or reconnected to the demand group. This may be immediately if the generation does not trip at the time of the outage or is not tripped for other technical reasons. Otherwise it will be from the point of time when the generation is reconnected

where:- F = appropriate value from Table A.1a or Table A.1b

41 this being the total effective capability of the remaining circuit capacity plus the effective capacity of any generation operating for more than 24hr per day.

Page 110 of 150


5.4. Graphical Approach Relevant graphs/tables were obtained from a series of sensitivity assessments using

the analysis package described in Section 3. These are presented separately for non-intermittent generation and for intermittent generation. There is no attempt to identify generation technology with any of these characteristics. Instead these are plotted and tabulated only as a function of unit availability and of unit numbers. In this way, these characteristics can be used for any generation type and any plant numbers.

The results are shown in Figures 44-47 and Tables 30 and 31.

Table 30 – F factors in % as function of availability and number of generators

number of units availability (%) 1 2 3 4 5 6 7 8 9 10

5 3 5 5 5 5 5 5 5 5 5

10 7 10 10 10 10 10 10 10 10 10

15 10 14 15 15 15 15 15 15 15 15

20 13 19 19 20 20 20 20 20 20 20

25 16 23 24 24 25 25 25 25 25 25

30 20 27 28 29 29 29 30 30 30 30

35 23 31 32 33 34 34 34 34 35 35

40 26 34 36 37 38 38 39 39 39 39

45 30 38 40 41 42 43 43 43 43 44

50 33 41 44 45 46 47 47 47 48 48

55 36 45 47 49 50 50 51 51 52 52

60 40 48 51 52 53 54 55 55 56 56

65 43 51 54 56 57 58 59 59 60 60

70 46 54 58 60 61 62 63 63 64 64

75 50 57 61 63 65 66 67 68 68 69

80 53 61 65 67 69 70 71 71 72 73

85 58 64 69 71 73 74 75 75 76 77

90 63 69 73 75 77 78 79 79 80 80

95 69 74 78 80 82 83 84 85 87 88

98 75 79 82 85 89 92 92 93 94 94

Page 111 of 150


Table 31 – Number of generators (N1) contributing to FCO

number of units availabil

ity (%) 1 2 3 4 5 6 7 8 9 10

30

35 9

40 7 8 9

45 6 7 8 8

50 5 6 7 7 8

55 all units

5 6 6 7 7

60 4 5 5 6 6 7

65 4 4 5 5 6 6

70 3 4 4 4 5 5 6

75 3 3 4 4 4 5 5

80 2 3 3 3 4 4 4 4

85 2 2 3 3 3 3 4 4

90 2 2 2 2 3 3 3 3

95 1 2 2 2 2 2 2 2 2

98 1 1 1 1 2 2 2 2 2

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%Availability (%)

Con

trib

utio

n (%

)

1 2 3 4 5 6 7 8 9 10Number of units

Figure 44 – F factors in % as function of availability and number of generators

Page 112 of 150


40%

50%

60%

70%

80%

90%

100%

60% 65% 70% 75% 80% 85% 90% 95% 100%Availability (%)

Con

trib

utio

n (%

)

1 2 3 4 5 6 7 8 9 10Number of units

Figure 45 – F factors in % as function of availability and number of generators

0%

5%

10%

15%

20%

25%

30%

0 12 24 36 48 60 72 84 96 108 120

Persistence Tm (h)

Con

trib

utio

n (%

)

Single site Multiple sites

Figure 46 – F factors in % as function of persistence Tm for wind farms

Page 113 of 150


10%

15%

20%

25%

30%

35%

40%

0 24 48 72 96 120 144 168 192 216 240 264 288 312 336 360 384

Persistence Tm (h)

Con

trib

utio

n (%

)

Figure 47 – F factors in % as function of persistence Tm for small hydro

5.5. Computerised Package Approach This analysis package has been described in Section 3 and has been used to obtain

all the results presented in the tables and figures of Section 4. Any user of this package should first establish that the same results can be obtained using the same data. Any differences should be reported to ENA, the organisation responsible for maintaining this analysis package.

5.6. Concluding Comments The results, tables and graphs presented in this section are the most important output

of this project since these are being proposed as the basis of the revised P2/5, i.e. the basis of the new security standard P2/6.

Page 114 of 150


6. Applications and Illustrative Examples

6.1. Introduction A necessary part of this project is to illustrate the application of the tables specifying

the capabilities of generation to system security. Two objectives exist:- i) to ensure that the results are considered reasonable from a practical point of

view ii) to demonstrate the process of estimating system capability to cover a FCO

and, if appropriate, a SCO for a range of typical examples. Both objectives can be achieved using the same set of examples.

6.2. System Structure Considered The examples chosen are similar in structure to those already included in ACE

Report 51. The merit of this is that the existing examples are widely known and fully understood. Consequently, the examples included in this report are extensions rather than completely new. Several adaptations were needed to illustrate the effect of number of units and availability, and the impact of intermittency. All examples are based on the system structure shown in Figure 48; this being based on that shown in Appendix A.10 of ACE Report 51.

load

2 x 45 MVA transformers 1.3 cyclic rating factor 0.95 power factor Ng generators

Figure 48 – System structure used for illustrative examples

The following examples consist of three groups:- (i) the example in Section 6.3 illustrates the application of the new “Table 2” for

a specific type of generation plant, namely landfill gas generation (ii) the examples in Section 6.4 illustrate general assessment principles and

effects for a range of generation, using the generic factors obtained from the graphical approach. In these examples, the effect of different number and type of generating units are considered together with the effect of whether the second transformer is connected in parallel with the first transformer as

Page 115 of 150


shown in Figure 48 or connected to another source through a normally open point. The generation units in all these examples conform with the materiality condition of “paragraph 3.6”. Consequently the circuits form the most significant outage, and there is no need to consider capping of the generating unit capacities used to assess security contribution.

(iii) the example in Section 6.5 illustrates the situation when the generating units do not conform with the materiality condition of “paragraph 3.6”. Two cases are considered. Firstly when the unit capacities are capped in assessing their security contribution so that the circuits remain the significant outage, and secondly when the unit capacities are not capped and the generating units form the significant outage.

All the examples in Sections 6.3-6.5 are concerned with the evaluation of the system

capability in each case. The only significant difference in the application of these capabilities is at what point in time does the support become available, i.e. either immediately or delayed in time following some switching action(s). This aspect is considered separately in Section 6.6.

6.3. Application of New “Table 2”

Data: type of generation = landfill gas

DNC of each unit = 1 MW

From “Table 2”: unit capability = 75 %

This section is not concerned with estimating the total system capability (this is the objective of Section 6.4), only with illustrating the use of “Table 2” in determining the capability contribution made by the specific generation. Consider the following two examples of non-intermittent generation. A similar procedure would be used for intermittent generation

6.3.1. Single Source of Non-intermittent Generation

number of units = 4

Contribution to system capability: = 3.0 MW

6.3.2. Multiple Sources of Non-intermittent Generation

The following approach evaluates the contribution of more than one source of generation by evaluating the capability contributions of the two sources separately and then summating these arithmetically. This is an approximation and gives a pessimistic value. A more accurate approach would be to treat it as a special case and use the analysis package.

Data: first type of generation = landfill gas number of units = 4

Page 116 of 150


DNC of each unit = 1 MW second type of generation = waste to energy number of units = 3 DNC of each unit = 0.5 MW

From “Table 2”:

capability of landfill gas = 75 % capability of waste to energy = 69 %

Contribution to system capability: = (4 x 1 x 0.75) + (3 x 0.5 x 0.69) = 3.0 + 1.0

= 83.2 MW

6.4.2. Non-intermittent Generation, Four Identical Units

= 4.0 MW

6.4. Estimating System Capability

6.4.1. Non-intermittent Generation, Two Identical Units Data: DNC of each unit = 20 MW unit availability = 90% capability = 69% Step 1: Examine FCO conditions (a) remaining circuit capacity following outage of the most critical circuit

= 1 x 45 x 1.3 x 0.95 = 55.6 MW

(b) effective capability of generation = 0.69 x (2 x 20) = 27.6 MW

(c) capacity to meet demand after FCO = 55.6 + 27.6

Step 2: Examine SCO conditions From Table 1 of P2/5, no SCO condition applies for demands less than

100 MW Step 3: Conclusion The system could usually be expected to support a maximum demand of

83.2 MW

Data: DNC of each unit = 10 MW unit availability = 90% capability = 75%

Page 117 of 150


Step 1: Examine FCO conditions (a) remaining circuit capacity following outage of the most critical circuit

= 1 x 45 x 1.3 x 0.95 = 55.6 MW


(c) capacity to meet demand after FCO = 55.6 + 30.0 = 85.6 MW

6.4.3. Intermittent Generation, Wind Farm

(c) capacity to meet demand after FCO



85.6 MW

6.4.3.1. Security Contribution During Switching Process Although intermittent generation such as wind farms make little contribution for

extended periods, these sources could provide significant support for short periods. One instance is when manual switching may take a period of say 3 hr after which the network is sufficient to comply with P2/6. In the period up to 3hr however the remaining circuit together with a wind farm may be able to provide sufficient security to satisfy P2/6. Consider this 3 hr period.

Data: nameplate rating = 10 MW capability = 24 % for Class C during switching, Tm = 3 hr Step 1: Examine FCO conditions (a) remaining circuit capacity following outage of the most critical circuit

= 1 x 45 x 1.3 x 0.95 = 55.6 MW

(b) effective capability of generation = 0.24 x 10 = 2.4 MW

= 55.6 + 2.4 = 58.0 MW


100 MW

Page 118 of 150


Step 3: Conclusion The system could usually be expected to support a maximum demand of

58.0 MW

6.4.3.2. Security Contribution During Maintenance Event With extended periods, wind sources make less contribution. However even for

relatively long periods, the additional contribution from a wind farm may be sufficient to ensure compliance with P2/6. Consider an 18hr maintenance period.

Data: nameplate rating = 10 MW

capability = 14 % for Class C during maintenance Tm=18 h Step 1: Examine FCO conditions (a) remaining circuit capacity following outage of the most critical circuit

= 1 x 45 x 1.3 x 0.95 = 55.6 MW

(b) effective capability of generation = 0.14 x 10 = 1.4 MW




57.0 MW

6.4.4. Non-intermittent Generation, Four Non-identical Units The following approach is an approximation using the tabulated values and

summating. The result is a pessimistic value. A more accurate approach would be to treat it as a special case and use the analysis package.

Data: DNC of Type 1 unit = 10 MW number of units = 2

capability = 69% unit availability = 90%

DNC of Type 2 unit = 15 MW number of units = 2

Page 119 of 150


unit availability = 60% capability = 48%

(a) remaining circuit capacity following outage of the most critical circuit

(b) effective capability of generation

= 55.6 + 28.2

The system could usually be expected to support a maximum demand of 83.8 MW

6.4.5. Non-intermittent and Intermittent Generation

Step 1: Examine FCO conditions

= 1 x 45 x 1.3 x 0.95 = 55.6 MW

= 0.69 x (2 x 10) + 0.48 x (2 x 15) = 28.2 MW

(c) capacity to meet demand after FCO

= 83.8 MW


100 MW

Step 3: Conclusion

42

The following approach is again an approximation using the tabulated values and

summating. The result is a pessimistic value. A more accurate approach would be to treat it as a special case and use the analysis package.

Data: two non-intermittent units

DNC of each unit = 20 MW unit availability = 90% capability = 69% one wind farm nameplate rating = 10 MW

capability = 24 % for Tm = 3 hours (switching event) Step 1: Examine FCO conditions (a) remaining circuit capacity following outage of the most critical circuit

= 1 x 45 x 1.3 x 0.95 = 55.6 MW

(b) effective capability of generation = 0.69 x (2 x 20) + 0.24 x 10 = 30.0 MW

42 If the capability of the system of four units is calculated simultaneously using the analysis package, the result will be greater. The aggregated capability of four units is 61.0% which means that the units could secure 0.61 x (2 x 10 + 2 x 15) = 30.5 MW. Therefore, in total the system could secure 55.6 + 30.5 = 86.1 MW.

Page 120 of 150




100 MW

capability (24hr) = 69%

Step 3: Conclusion The system could usually be expected to support a maximum demand of

85.6 MW

6.4.6. Additional Contribution from Two (<24hr) Units This example considers the case of non-intermittent generation. Data: DNC of each unit = 20MW unit availability = 90%

operating time = 12hr per day capability (<24hr) = 12% of group demand or 14% of “firm” capacity43 Step 1: Examine FCO conditions (a) remaining circuit capacity following outage of the most critical circuit

= 1 x 45 x 1.3 x 0.95 = 55.6 MW

(b) effective capability of generation is smaller of (i) 0.69 x (2 x 20) = 27.6 MW

and (ii) 0.14 x 55.6 = 7.8 MW




63.4 MW

43 this being the total effective capability of the remaining circuit capacity plus the effective capacity of any generation operating 24hr per day

Page 121 of 150


6.4.7. Systems with (24hr) and (<24hr) Generation This example considers the case of non-intermittent generation. Data: 2 units operate for 24hr per day 2 units operate for 12hr per day

DNC of each unit = 10MW unit availability = 90% capability (24hr) = 69%

Data: number of units = 6

capability (12hr) = 12% of group demand or 14% of “firm” capacity Step 1: Examine FCO conditions (a) remaining circuit capacity following outage of the most critical circuit

= 1 x 45 x 1.3 x 0.95 = 55.6 MW

(b) effective capability of 24hr per day generation = 0.69 x (2 x 10) = 13.8 MW

(c) effective capability of remaining circuit and 24hr per day generation = 55.6 + 13.8 = 69.4 MW

(d) effective capability of 12hr per day generation is smaller of (i) 0.69 x (2 x 10) = 13.8 MW

and (ii) 0.14 x 69.4 = 9.7 MW

(e) capacity to meet demand after FCO = 69.4 + 9.7 = 79.1 MW


100MW Step 3: Conclusion The system could usually be expected to support a maximum demand of

79.1 MW

6.5. Impact of Materiality Criteria for Non-intermittent Generation

6.5.1. Materiality Test on System Studied The materiality criteria (“paragraph 3.6”) are intended to ensure that generation is

not the significant outage condition and that the circuits dominate in supporting the demand. In order to illustrate the effect of this test, consider the following system example:-

DNC of each unit = 10 MW

Page 122 of 150


unit availability = 85 % unit capability = 74 % number of circuits = 3 capacity = 15 MW cyclic rating factor = 1.3 circuit capability = 19.5 MW

As:-

Condition (a) of paragraph 3.6 requires the cyclic rating of the largest distribution circuit to be greater than factor F % (74% from Table 30) of the N1 largest generation units (three from Table 31).

cyclic rating of the largest distribution circuit = 19.5 MW 74 % of the three largest generation units = 22.2 MW

the materiality condition (a) is not satisfied.

Condition (b) of paragraph 3.6 requires the cyclic rating of the two largest distribution circuits to be greater than factor F % (74% from Table 30) of the (N1+1) largest generation units (four from Table 31).

As:- cyclic rating of the two largest distribution circuits = 39.0 MW 74 % of the four largest generation units = 29.6 MW

the materiality condition (b) is satisfied. Therefore, because condition (a) is not satisfied, either the generation must be

capped in assessing its security contribution so that the condition is satisfied, or the capability is evaluated using the complete base table (“Table A.1”) rather than the simplified “Table A.2”/”Table 2”.

In order to ensure the materiality condition is satisfied, the capped capacity44 of each generating unit should not exceed an equivalent value of 8.8 MW in assessing the security contribution of the generation.. The following two examples illustrate the effect of this capping.

6.5.2. Materiality Condition Imposed Data: capped capacity of each unit = 8.8 MW unit availability = 85 % capability = 74 % Materiality test shows that the “FCO” is that of one circuit. Therefore:- Step 1: Examine FCO conditions

(a) remaining circuit capacity following outage of the most critical circuit = 2 x 15 x 1.3 = 39.0 MW

44 This process does not restrict the actual unit capacities, which would remain at 10MW in this case, and does not imply any suggestion of allocation between units

Page 123 of 150


(b) effective capability of generation = 0.74 x (6 x 8.8) = 39.1 MW




78.1 MW

6.5.3. Materiality Condition Not Imposed

Materiality test shows that the “FCO” is that of three generators. Therefore:-

Data: DNC of each unit = 10 MW unit availability = 85 % capability = 74 %

Step 1: Examine FCO conditions

(a) circuit capacity = 3 x 15 x 1.3 = 58.5 MW





80.7 MW.

6.6. Systems Requiring Consideration of Second Circuit Outages All the examples in Sections 6.4 and 6.5 support group demands less than 100 MW

and therefore SCO conditions do not apply. This approach conforms with most of the examples given in ACE Report 51. However ACE Report 51 does give one example for which the group demand that can be supported is greater than 100 MW. An example based on

Page 124 of 150


that shown in Figure 48 is included here. This example considers non-intermittent generation. A similar procedure would be used for intermittent generation.

In this example, a third circuit is added to the previous two giving the system shown

in Figure 49.

load

3 x 45 MVA transformers 1.3 cyclic rating factor 0.95 power factor 4 generators

Figure 49 – Modified system structure Data: number of circuits = 3

number of units = 4 DNC of each unit = 10 MW

unit availability = 90% capability = 75%

= 30.0 MW (c) capacity to meet demand after FCO

Step 1: Examine FCO conditions (a) remaining circuit capacity following outage of the most critical circuit

= 2 x 45 x 1.3 x 0.95 = 111.2 MW

(b) effective capability of generation = 0.75 x (4 x 10)

= 111.2 + 30.0 = 141.2 MW

Step 2: Examine SCO conditions From Table 1 of P2/5, the smaller of (1/3 of group demand) and (group demand

minus 100 MW) must be supplied within 3 hr of a double circuit outage if group demand exceeds 100 MW. If group demand were to be 141.2 MW (as in (c) above), then the remaining circuit and the generation would be required to supply the smaller of (1/3 of group demand = 47.1 MW) and (141.2 – 100 = 41.2 MW) within three hours. (Note that no contribution from plant operating for less than 24 hr per day should be allowed). The effective contribution of the remaining circuit and the generation is 85.6 MW and is thus well able to meet the SCO requirement.

Page 125 of 150


Step 3: Conclusion

This situation would occur if the circuits are operated in parallel, but all or some of the generation is either not running at the time of the outage or trip following the system fault. In this case the capabilities of the remaining circuit and of any generation that is running and does not trip is available immediately. However the capability of the remaining generation will not be available until the units are operating and/or those that are tripped are reconnected. Generally these would become available within the 15 min requirement specified in P2/5 for Class C demands, and therefore also for Class B demands and the second stage of Class D demands.

The system could usually be expected to support a maximum demand of 141.2 MW.

It should be noted that this level of maximum demand exceeds the continuous rating

capability of the three circuits. This is compatible with both ACE Report 51 and P2/5. However, although this means the system is P2/5 (and P2/6) compliant for a group demand of 141.2 MW, other system technical features may cause the actual demand to be supplied to be less than this value. This is outside the scope of this project and also outside the scope of P2/5.

6.7. Times At Which Capabilities Become Available

System capabilities are evaluated independently of the point in time that they become available to support group demand. Consider the following separate cases:-

(a) All sources available immediately

This situation would occur if the circuits are operated in parallel, and all the generating units are running at the time of the outage and do not trip following the system fault. In this case the capabilities evaluated in Section 6.3 become available immediately and the demand can be supported without interruption45.

(b) Circuit available immediately, generation delayed

(c) Circuit delayed, generation available immediately This situation would occur if the second circuit is connected to the system via a

normally open point and the infeed requires various switching events to take place before it can pick up any demand. The generation however is assumed to be running at the time of the outage and does not trip because of the system fault. As this would necessitate islanded operation, it is normal practice to trip the generation and not reconnect it until the network is re-established. In this case, the demand cannot be supported immediately, but only after the time it takes to restore supply through the alternative network connection. This can generally be accomplished within three hours using manual switching, which is sufficient for Class B although not for Class C. In the latter case, the switching needs to be achieved remotely or automatically.

45 It should be noted that P2/5 (and therefore presumably P2/6) defines “immediate” restoration as one in which the interruption lasts less than one minute.

Page 126 of 150


(d) All sources delayed This situation is similar to that described in (c) above when the generation is

deliberately tripped to avoid islanded operation.

6.8. Concluding Comments The examples included in this section serve to illustrate the procedure for assessing

whether the system is P2/5 or P2/6 compliant. There are an infinite number of possible alternative scenarios, but it is felt that those included are sufficient for the purposes of the two objectives defined in Section 6.1. In reality, this procedure does not differ from that already established in ACE Report 51 and P2/5 and used for nearly 30 years.

Page 127 of 150


7. Application Guide

7.1. Introduction This Guide summarises the procedure for assessing whether the supply to a

particular group demand complies with the revised security standard, P2/6. The procedure is based on that underpinning the existing standard P2/5. The only modifications of note are those that relate to the inclusion of modern generation and the methodology used in this project for determining the information that should be included in the “new” security standard, P2/6. As this is a summary only, the reader may need to refer to the main text of this report for further and more comprehensive details.

7.2. Approaches Three possible approaches have been identified:- (i) look-up table(s) approach - this has the merits of simplicity and should be

free of erroneous applications. However, it is restricted to specific generation types and typical or average availability data. In the case of non-identical units, the capabilities are approximate since generation contributions are assessed separately for each type and unit size and then added arithmetically. This gives a pessimistic result, but generally of acceptable precision. This approach can also incorporate the concept of capping of the security contribution of generation so that it does not violate the materiality conditions which require generation to be less significant than the circuits

(ii) graphical approach - this is an extension of the previous approach and is

still reasonably simple. It accommodates a wider range of data and choices, but still requires approximate summations for non-identical units

(iii) computer program approach - this permits an unrestricted choice of data and mixed generation types. The prototype analysis package is implemented in MS Excel using VBA environment and a User Guide is provided (Section 3) for its easy application. The package calculates the security contributions only and these need to be assessed for compliance in the same way as performed with either of the two previous approaches.

7.3. Systems and Data Whichever of the three approaches is used to test for compliance with the P2/6

security standard, the system characteristics needs to be assessed. This is identical in concept with P2/5. This assessment is used to determine whether the system is sufficiently normal to allow the application of either the look-up table approach or the graphical approach. If any of the conditions or constraints used to produce the look-up tables or graphical tables are

Page 128 of 150


insufficiently relevant then, as in P2/5, special studies would need to be performed. This would require using the computer program approach.

The following discussion summarises the important aspects of the key parameters

and criteria. Greater detail is given in the main body of the report.

7.3.1. Generation Capacities For the purposes of determining the security contribution of generation, the

following definitions of capacity are used:- (a) non-intermittent generation plant. The actual DNC or sent out capacity is

used, i.e. the actual sustained output of the plant after deduction of the auxiliary loads of the plant but not the separate site loads which are, ideally, separately identified.

(b) intermittent generation plant. The nameplate rating is taken as the

capacity.

7.3.2. Generation Availabilities The analyses carried out in this project used the values of availability set by the Data

Collection and Processing Report under contract K/EL/00303/05: these values were approved by Workstream 3. The Data Collection and Processing Report gives availability guidelines for various forms of plant. It also shows that the availability can vary significantly for many types of plant.

Although, in order to increase the concept of relevance, it is theoretically best to use data specific to a particular plant, or similar plant operated in a similar manner, this may not be possible in practical terms because of paucity of data. In such cases, the need to increase the concept of confidence means that use of pooled data becomes essential.

The average is a sound statistical concept that recognises variability or uncertainty. Use of an average means that sometimes the output result is greater and sometimes less than the one eventually observed for a particular plant. An average can be obtained for a specific plant, or a set of similar plant, or for a set of plant that may have some divergence in characteristics. However, if the sample size of sets is restricted, it could require many years of cyclic up and down data to ascertain actual characteristics of a specific plant or type of plant. In reality exactly the same problem exists with circuits, the variability of individual circuits can vary widely but the average performance is the basis of P2/5 and hence should be of P2/6. Similarly, P2/5 treats all generation as having the same average availability (86%). Finally, if a plant is operating poorly then remedial action is likely; when it will possible revert to something near average. For these and related reasons, it is preferable to use the average availability from sets of similar type as determined in the Data Collection and Processing Project and used in the preparation of the tables in this report.

Page 129 of 150


Therefore, it is suggested that the average availability value for each type of plant should be used as the first indicator of the security contribution from generating plant connected to a specific demand group.

For complex generating plant with a number of auxiliaries, there is a possibility that

the plant may exhibit of a number of derated or partial output states. There are a number of ways these can be treated; one of which is the equivalent forced outage rate (EFOR) as defined in IEEE Standard 762. This approach attempts to modify the unavailability by dividing the partial output states between the fully up state and the fully down state to give a two-state model of the unit. It should be noted that it is an approximation but one that gives pessimistic results46, which can be argued as being the more appropriate situation to be in.

7.3.3. Generation De-minimus Criteria These criteria have been agreed by WS3, and are not elaborated here. However

these will be particularly relevant for circumstances where generation is of a significant aggregate capacity relative to the capability of the network into which it feeds.

7.3.4. Generation Winter-Time Operating Regime The winter operating period must be ascertained, e.g. whether it operates for 8 hr or

12 hr or whether it is continuously available. In the case of restricted operating times:- •

•

•

the approaches used assume that the increasing demand at the start-up time is the same as the decreasing demand at shut-down time if this is not so, then the contribution may be less than the approaches suggest in the extreme, if the operating period does not span the peak demand at all, the contribution from such generation is zero.

7.3.5. Intermittent Generation

The value of Tm for the situation being assessed must be deduced. P2/5 requires that some or all demand (depending on Class of Supply) should be restored within 15 min or 3 hr or after the repair time. If generation is to be used for such purposes, it must have a reasonable expectation of persisting for the subsequent periods to allow compliance with the standard.

For intermittent generation, such as wind and hydro, the choice of Tm is a material issue related to the nature of the networks into which the intermittent generation is connected. Systems, in which the demand can be restored by transfer capacity but only after a switching period, may allow intermittent generation to make a larger security contribution during this

46 R.Billinton and R.N.Allan. “Reliability evaluation of power systems”. Plenum Publishing, New York, Second Edition, 1996

Page 130 of 150


switching period after which it is no longer required. This effectively indicates a short Tm value.

(a)

Recommended47 values for Tm are shown in Table 31.

7.3.6. System Demand The winter peak demand of the demand group, adjusted for any disaggregated site

loads or non-de minimis plant, must be determined.

7.3.7. Materiality Criteria In order to use Table 2 of P2/5, generation should not be dominant and consequently

materiality criteria were introduced to prevent this happening. These criteria have been revised to be compatible with the methodology and the revisions of P2/6. These revised criteria are:

The contribution of generation specified in the “new Table 2” of P2/6 is based on the

assumptions that:-

the cyclic rating of the largest distribution circuit is greater than F% of the total sent-out capacity of the N1 largest generating units

(b) the cyclic rating of the two largest distribution circuits is greater than F% of the total sent-out capacity of the (N1+1) largest generating units

If the materiality condition is not satisfied, then the maximum security capacity Cg of the generating units can be capped at the maximum value that satisfies the above assumptions, i.e. for identical units:-

from first materiality condition Cg < Cc/F.N1

from second materiality condition Cg < 2.Cc/F.(N1+1)

7.3.8. First and Second Circuit Outages

The FCO and SCO are as defined presently in P2/5. Under these outage conditions, the generation must have a persistence that can match either the time for switching the demand to other alternative sources (associated with a Tm value of between ½ to 3hr), or the time for repair of the outage (associated with a Tm value of one or two days).


Page 131 of 150


7.3.9. Time Periods for Compliance Testing It is critically important to note that the capability assessment needs to be done for

each of the time periods specified in “Table 1 of P2/5”. For instance, in the case of Class C, the two time periods of concern are the demand that must be recovered in 15 min and the demand that must be recovered in 3 hr. Both periods must be assessed separately since the required demand and the number of circuits and the amount of generation could be different in each of these. Intermittent generation, such as wind, is expected to contribute significantly to the short periods.

Compliance with P2/6, as in P2/5, is required for each time period.

The overall procedure used for Approaches 1 and 2 is shown in the process diagram of Figure 51.

7.4. Approach to Use The following guidelines indicate which of the three approaches to use. • Look-up Table Approach (Approach 1). This approach is preferred because

of its practical simplicity provided the underpinning conditions are satisfied, i.e. technology is known, and the associated average availability of non-intermittent generation or the data used for intermittent generation is considered acceptable. If the materiality condition is satisfied or the unit capacity is capped, then the simple “new Table 2” can be used. If the materiality condition is not satisfied and the unit capacity is not capped, then the more extensive table “equivalent to Table A.1 of ACE Report 51” can be used. If mixed generation systems exist, then the capability contributions of each of the individual groups is summated arithmetically.

• Graphical Approach (Approach 2). This approach is an extension of the Approach 1 and is used if one or more of the underpinning conditions of Approach 1 are not satisfied, i.e. the technology is not one of those included, the associated average availability of non-intermittent generation or the data used for intermittent generation is not considered acceptable. Materiality and mixed generation systems are treated the same as in Approach 1.

• Computer Package Approach (Approach 3). This approach is a

computerised model of the methodology used to create the tables used in Approaches 1 and 2. It offers the ability to accommodate a wide range of data and assumptions, and permits the underpinning conditions of the other approaches to be relaxed and modified. It is therefore appropriate for special studies and bespoke analyses.

7.5. Procedure Used For Approaches 1 And 2

The process starts with checking the size and numbers of generators in the group

being examined. If the generation capacity is less than de minimis level then it is neglected.

Page 132 of 150


If the materiality conditions are met, then the FCO is the largest circuit and the

process continues with the calculation of the system capability under this outage condition. If the materiality conditions are not met, then the options are to cap the generation

capacity (process shown in Figure 51) and to use the largest circuit as the FCO, or not to cap and use one or more generators as the equivalent FCO. In both cases the process continues with the calculation of the system capability under this outage condition (process shown in Figure 52).

The system capability is determined as described in the following sections.

7.5.1. Approach 1 • if the materiality condition is satisfied or the generating units are capped, the

system capability is the summation of the following contributions:- remaining circuit cyclic capacity plus any transfer capacity plus the appropriate generation contribution from “new Table A.2” or

“new Table 2” (Table 33), using values of F from sub-tables “Table A.1a” (Table 34) or “Table A.1b” (Table 35) for non-intermittent and intermittent generation respectively. If more than one generation group exists, the generation capability is the arithmetic summation of the individual contributions

plus any additional contribution from generation having an operational period less than 24hr using Table 33

• if the materiality condition is not satisfied and the unit capacity is not capped, the system capability is the summation of the following contributions:- circuit cyclic capacity

7.5.2. Approach 2

plus any transfer capacity plus the appropriate generation contribution from “new Table A.1”

(Table 37) using values of F from sub-tables “Table A.1a” (Table 34) or “Table A.1b” (Table 35) for non-intermittent and intermittent generation respectively, and N1 from “Table A.1c” (Table 36). If more than one generation group exists, the generation capability is the arithmetic summation of the individual contributions

plus any additional contribution from generation having an operational period less than 24hr using Table 37.

• if the materiality condition is, or is not, satisfied, the system capability is the

summation of the following contributions:- remaining circuit capacity

plus any transfer capacity plus the appropriate generation contribution from “new Table A.1”

(Table 37), using values of F and N1 from Tables 38 and 39

Page 133 of 150


respectively for non-intermittent generation, and values of F from Figures 54 and 55 for wind farms and small hydro generation respectively. If more than one generation group exists, the generation capability is the arithmetic summation of the individual contributions

plus any additional contribution from generation having an operational period less than 24hr using Table 37.

It is critically important to note that this capability assessment needs to be done for

each of the time periods specified in “Table 1 of P2/5”. For instance, in the case of Class C, the two time periods of concern are the demand that must be recovered in 15 min and the demand that must be recovered in 3 hr. Both periods must be assessed separately since the required demand and the number of circuits and the amount of generation could be different in each of these. Compliance with P2/6, as in P2/5, is required for each time period.

If the demand exceeds the system capacity under first circuit conditions in any one

time period, the system is declared as not complying with the security standard. If the system is compliant under the first circuit outage condition, then the process

moves to checking if a second circuit outage condition needs to be considered. P2/5 requires this security requirement for group demands in excess of 100MW.

The assessment of system capability under a SCO is similar to that under a FCO

(process shown for a circuit as SCO in Figure 53). The only difference of note is that either the next largest circuit is outaged, or one more largest generating unit is outaged: the one removed being dependent on whether the materiality condition is or is not satisfied respectively.

If a SCO security assessment is required and the system demands can be satisfied,

then the system is deemed to comply with the security standard.

7.6. Procedure Used For Approach 3 This approach expects a set of input data to be supplied by the user. The detailed

procedure for this and for using the package is described in detail in Section 3 of this report.

Table 32 - Recommended48 values for Tm

group demand switching repair maintenance

A n/a n/a n/a B 24 hours 24 hours 2 hours C 3 hours 5 days 18 hours D 3 hours 15 days 24 hours E 3 hours 90 days 24 hours


Page 134 of 150


Table 33 - Format of new “Table A.2”

contribution after first circuit

contribution after second circuit

notes type of generation outage outage classes of supply A-E

classes of supply D and E only

landfill gas F % of DNC F % of DNC

CCGT F % of DNC F % of DNC

sewerage: spark ignition F % of DNC F % of DNC

sewerage: GT F % of DNC F % of DNC

CHP

other CHP F % of DNC F % of DNC

waste to energy F % of DNC F % of DNC


F % of nameplate rating wind the values in these two rows

apply to the complete site irrespective of number of units F % of nameplate

rating F % of nameplate rating small hydro

SMALLER OF SMALLER OF


above value OR 11 % of group demand (12 % of firm capacity)

above value OR 11 % of group demand (12 % of firm capacity)

the values in these two rows assume that the operating period is such that operation spans the peak demand, and the demand at start-up is the same as the demand at shut-down, i.e. operation is symmetrically placed on the daily load curve. If these conditions do not apply, the contribution could be optimistic, in which case the generation ought to be treated as a special case. At one extreme, the contribution would be zero if the operating period did not span the peak demand at all.

SMALLER OF SMALLER OF above value above value OR OR plant operating for 12 hr 12 % of group demand (14 % of firm capacity)

12 % of group demand (14 % of firm capacity49)

NOTES all the values in this table apply from the point of time when the generation is connected or reconnected to the demand group. This may be immediately if the generation does not trip at the time of the outage or is not tripped for other technical reasons. Otherwise it will be from the point of time when the generation is reconnected


49 this being the total effective capability of the remaining circuit capacity plus the effective capacity of any generation operating for more than 24hr per day.

Page 135 of 150


Table 34 – “Table A.1a”, F factors in % for non-intermittent generation

number of units


landfill gas 63 69 73 75 77 78 79 79 80 80

CCGT 63 69 73 75 77 78 79 79 80 80


sewerage: GT 53 61 65 67 69 70 71 71 72 73 CHP

other CHP 53 61 65 67 69 70 71 71 72 73

waste to energy 58 64 69 71 73 74 75 75 76 77

Table 35 - “Table A.1b”, F factors in % for intermittent generation

Tm, hr type of generation

½ 2 3 18 24 120 360 >360

single site 28 25 24 14 11 0 0 0 wind multiple

sites 28 25 24 15 12 0

small hydro 37 36 36 34 34 25 13 0

Table 36 – “Table A.1c”, Number of generators (N tributing to FCO 1) con

number of units


landfill gas 1 2 2 2 2 2 3 3 3 3

CCGT 1 2 2 2 2 2 3 3 3 3


sewerage: GT 1 2 2 3 3 3 4 4 4 4 CHP

other CHP 1 2 2 3 3 3 4 4 4 4

waste to energy 1 2 2 2 3 3 3 3 4 4

wind all, i.e. complete wind farm

small hydro all, i.e. complete site

Page 136 of 150


Table 37 - Format of new “Table A.1” 1 2 3 4

MAXIMUM ADDITIONAL CONTRIBUTION IN % BY

GENERATION NOT CONTINUOUSLY AVAILABLE

OUTAGE CONDITION

SYSTEM CAPACITY INCLUDING GENERATION AVAILABLE FOR 24H

for 8hr for 12hr

(a) total transmission/distribution cyclic capacity plus 11 12 no outage (12) (14) (b) maximum sent-out capacity of all generation units x F = SC0

SC0 minus the larger of

maximum sent-out capacity of any N1 generation units x F = SC1

SC0 minus the largest of

maximum sent-out capacity of any N2 generation units x F = SC2

N1 = appropriate value from Table A.1c N2 = N1 + 1

(c) cyclic capacity of any transmission/distribution circuit first circuit outage

11 12 or (12) (14) (d)

(e) the sum of (c) and (d) above or 11 12 second circuit

outage (f) cyclic capacity of any two transmission/distribution circuits (12) (14) or (g)


the percentages in columns 3 and 4 represent the % of group demand and, in parentheses, the % of firm capacity; the latter being the total effective capability of the remaining circuit capacity plus the effective capacity of any generation operating for 24hr per day. These two values are equivalent.

the percentages in columns 3 and 4 are the maximum contributions because it is assumed that the operating period is such that operation spans the peak demand, and the demand at start-up is the same as the demand at shut-down, i.e. operation is symmetrically placed on the daily load curve. If these conditions do not apply, the contribution will be less

Page 137 of 150


Table 38 – F factors in % as function of availability and number of generators

number of units availability (%) 1 2 3 4 5 6 7 8 9 10

5 3 5 5 5 5 5 5 5 5 5

10 7 10 10 10 10 10 10 10 10 10

15 10 14 15 15 15 15 15 15 15 15

20 13 19 19 20 20 20 20 20 20 20

25 16 23 24 24 25 25 25 25 25 25

30 20 27 28 29 29 29 30 30 30 30

35 23 31 32 34 33 34 34 34 35 35

40 26 34 36 37 38 38 39 39 39 39

45 30 38 40 41 42 43 43 43 43 44

50 33 41 44 45 46 47 47 47 48 48

55 36 45 47 49 50 50 51 51 52 52

60 40 48 51 52 53 54 55 55 56 56

65 43 51 54 56 57 58 59 59 60 60

70 46 54 58 60 61 62 63 63 64 64

75 50 57 61 63 65 66 67 68 68 69

80 53 61 65 67 69 70 71 71 72 73

85 58 64 69 71 73 74 75 75 76 77

90 63 69 73 75 77 78 79 79 80 80

95 69 74 78 80 82 83 84 85 87 88

98 75 79 82 85 89 92 92 93 94 94

Page 138 of 150


Table 39 – Number of generators (N1) contributing to FCO

number of units availabil

ity (%) 1 2 3 4 5 6 7 8 9 10

30

35 9

40 7 8 9

45 6 7 8 8

50 5 6 7 7 8 all units

55 5 6 6 7 7

60 4 5 5 6 6 7

65 4 4 5 5 6 6

70 3 4 4 4 5 5 6

75 3 3 4 4 4 5 5

80 2 3 3 3 4 4 4 4

85 2 2 3 3 3 3 4 4

90 2 2 2 2 3 3 3 3

95 2 1 2 2 2 2 2 2 2

98 2 1 1 1 1 2 2 2 2

Page 139 of 150


Page 140 of 150

Start Calculation of the system capability to support

demand in the case of materiality rules satisfied

Does the system configuration satisfy

the materiality criteria?

Is the required demand less than system capacity SC1?

Non compliance with security of supply standard

Stop

Is peak demand more than 100 MW?

Calculate system capacity with first circuit outage SC1

Is the required demand less than system capacity SC2?

Compliance with security of supply standard

YES

NO

NO

YES

Stop

NO

YES

Non compliance with security of supply standard

Stop

NO

Compliance with security of supply standard

Stop

YES

Calculate system capacity with second circuit outages SC2

Cap the capacities of generation units

Materiality criteria to be

imposed?

YES

NO

Figure 50 – Process diagram for

application of security standard


Start Calculation of capped unit capacity

Calculate capped capacity CC1

Capped capacity CC1 is cyclic rating of the largest distribution circuit divided by the factor F (Table A.2) and number of generators contributing to the first circuit outage N1 (Table A.1c)

Calculate capped capacity CC2

Capped capacity CC2 is cyclic rating of the two largest distribution circuits divided by the factor F (Table A.2) and number of generators contributing to the first circuit outage N2 = N1 + 1 (Table A.1c)

Calculate capped capacity CC = min(CC1, CC2) CC is the maximum capacity or the generation units when the

materiality conditions are satisfied

Stop

Figure 51 – Process diagram for capping capacity of generating units

Page 141 of 150


Start Calculation of the system capability to support demand after the first circuit outage

Calculate capacity SC1a

Calculate capacity SC1b = min(SC<24, SCc)

Capacity SC1b is an additional capacity calculated as a minimum of maximum sent-out capacity of all generation units operating for less than 24 hours per day times factor F (Table A.2) and percentage (given in parentheses in Table A.2) of SC1

Calculate capacity SC1 = SC1a + SC1b SC1 is the capacity the system could usually be expected to support

demand after the first circuit outage

Stop

Is materiality criteria satisfied or imposed?

YES

NO


Capacity SC1a is a total remaining transmission/distribution cyclic capacity after first circuit outage plus maximum sent-out capacity of all generation units operating 24 hours per day times factor F (Table A.2)

Capacity SC1a is a total transmission/distribution cyclic capacity plus maximum sent-out capacity of remaining generation units (after removing generators according to Table A.1c) after equivalent first circuit outage operating 24 hours per day times factor F (Table A.2)

Figure 52 – Process diagram for assessing capability after first circuit outage

Page 142 of 150


Start Calculation of the system capability to support demand after the second circuit outage


Capacity SC2a is a total remaining transmission/distribution cyclic capacity after second circuit outage plus maximum sent-out capacity of all generation units operating 24 hours per day times factor F (Table A.2)

Calculate capacity SC2b = min(SC<24, SCc)

Capacity SC2b is an additional capacity calculated as a minimum of maximum sent-out capacity of all generation units operating for less than 24 hours per day multiplied by factor F (Table A.2)

and percentage (given in parentheses in Table A.2) of SC2a

Calculate capacity SC2 = SC2a + SC2b SC2 is the capacity the system could usually be expected to support

demand after the second circuit outage

Stop

Figure 53 – Process diagram for assessing capability after second circuit outage

Page 143 of 150


0%

5%

10%

15%

20%

25%

30%

0 12 24 36 48 60 72 84 96 108 120

Persistence Tm (h)

Con

trib

utio

n (%

)

Single site Multiple sites

Figure 54 – F factors in % as function of persistence Tm for wind farms

10%

15%

20%

25%

30%

35%

40%

0 24 48 72 96 120 144 168 192 216 240 264 288 312 336 360 384

Persistence Tm (h)

Con

trib

utio

n (%

)

Figure 55 – F factors in % as function of persistence Tm for small hydro

Page 144 of 150


8. Conclusions

8.1. Objectives

In addition a number of related technical issues had to be resolved. Consequently this final report contains the definitive outcome of the discussion and resolution of these issues, together with the proposals for developing a new security standard P2/6, a proposed new “Table 2” and supporting text, an application guide, examples illustrating the approach and application, and a description of the associated analysis package.

8.2. Technical Issues

(i) Treatment of single unit generation systems

As part of its wide brief on the impact of distributed and modern forms of generation, the DTI50 has mandated Workstream 3 to focus on several short-term network solutions, one of which relates to the immediate problem of how best to assess the contribution of distributed generation to network security. This project formed part of these developments and followed other related projects conducted by UMIST.

The main objective of this project was to use the methodology developed in the previous FES project [“Security Contribution from Distributed Generation”, ETSU/FES Project K/EL/00287 Extension. Final Report by R.N.Allan, G.Strbac and K.Jarrett, UMIST, 11 December 2002] to assess the security capability of modern distributed generation in order to review Table 2 and related text of Engineering Recommendation P2/5, and to propose information and results that could be used to create a new P2/6 that takes into account: modern type of generating units; unit numbers; unit availabilities; and capacities.

The specification was to apply this methodology in a way that would reflect the

attributes of present-day generation but constrained in two very specific respects:- (i) the application had to be simple, easy to implement and achievable in the

short term (ii) the application had to be consistent with that used to develop the generation

contributions specified in the present P2/5.

During the previous Methodology project, a number of issues were raised but, as

they were not considered to be material at that time, were postponed until this project. These are summarised in the following paragraphs.

During the previous project, one issue raised was that a single unit should not be

relied on to provide security. A significant part of this project, and of the report itself, was spent on addressing this issue. From the analytical assessment, it was agreed that single units do have a security capability and should be given credit for doing so. It was also agreed that

50 In order to implement the recommendations of the DTI/Ofgem, a Distributed Generation Co-ordinating Group together with a supporting Technical Steering Group (TSG) has been established. A number of workstreams are being pursued; Workstream 3 being focussed on short-term network solutions.

Page 145 of 150


the methodology quantifies this capability in a way that is compatible with the consideration of all other units and with the philosophy of the existing P2.6. Consequently, single units are included as entities in the development of the proposals for a new “Table 2”.

For generation to provide security, the output must remain at or above a certain

required level for a minimum period of time, defined as Twith intermittent generation such as wind due to its significant variability. This persistence time has a considerable impact on the capability that can be associated with intermittent generation and is related to the duration of the system conditions for which such generation may be able to avoid or reduce customer disconnections. There are three distinct system conditions, each of which can be associated with different minimum persistence times. These are:-

(iii) maintenance activities.

Proposals for required values of T conditions, are given in the report.

(ii) Effect of shape of load duration curves (LDC)

During the Methodology project, it was observed that the shape of the LDC affected the capability associated with the generation. It was decided to leave any further study until the present project in order to perform the sensitivity studies using real data for the generating plant and for the LDC. Nine typical LDCs were obtained by the Data Collection and Processing project51. These individual LDCs have been used separately in these sensitivity studies. It was observed that the shape of the LDC affects the contribution delivered by generation; the effect being greater with increasing unit availability up to about 90% and with decreasing number of units. Although this may seem to be an effect that should be included as a parameter in the development of P2/6, it should be noted that the variation between minimum and maximum is, in most cases, of the same order or less than the variation in contribution caused by practical ranges of unit availability. It can therefore be recommended that an average LDC should be used as well as an average value of unit availability in the development of a “Table 2” similar to that of Table 2 in P2/5. If this assumption is considered unacceptable in specific situations, then such situations should be considered special cases and assessed individually using the analysis-package approach. The average LDC obtained from these nine typical LDCs was used to obtain the proposed values for P2/6.

(iii) Persistence of intermittent generation, Tm

m. This is generally only a problem

(i) switching activities (ii) repair activities

m, depending on system

(iv) Time resolution of intermittent generation output profiles

All intermittent generation output profiles are obtained by averaging over a specific period, frequently ½ hr periods. One issue was whether this was acceptable or whether resolution times of 5 min or even 1 min would be necessary. Sensitivity studies showed that the security capability of intermittent generation decreased as the resolution time decreased,


Page 146 of 150


as expected. However, since the thermal time constant of overhead lines tends to be the most critical and these are generally of the order of several minutes, it is recommended that a 5 min resolution for quantifying the contribution of wind generation should be used. This value is used in the derivation of the “new Table 2”. In other cases, for instance with circuits composed of cables and/or transformers, it may be sufficient to use a 30 min resolution. Appropriate correction factors are proposed and included in the report.

(v) Ride-through capability Ride-through capability is of great importance in knowing how, or even whether,

distributed generation can contribute to system security. This capability does not affect the effective generation contribution of a particular generation site to be determined by the methodology and to be specified in a table similar to Table 2 of P2/5. What it does affect is whether this contribution is available immediately, after a certain time period, or not at all. This depends on whether the generation plant is, or is not, tripped following a circuit outage; and if tripped, how long it takes to restore its network connection. These scenarios in no way affect the methodology, nor the effective generation contributions that the methodology determines. It is only the way that these contributions are used that is dependent on the scenario. In some cases, the contribution becomes available immediately (similar to base loaded stations in the current P2/5) and sometimes the contribution only becomes available after a period.

(vi) Risk to loss of supply There is a view that the risk to supply (presumably meaning the impact on customer

reliability) could be degraded by using generation instead of circuits: this belief may exist because of the view that the reliability of generating units is less than that of overhead lines and underground cables. However the benefit of reinforcing a system using generation is clearly an alternative to constructing a new circuit or upgrading an existing one. Reinforcing with generation would undoubtedly increase reliability although perhaps by less than that due to an additional circuit. The discussion in the report identified that:-

• a revised P2/5 and Table 2, i.e. a new P2/6, should not cause customer

reliability to be less than that envisaged by P2/5, and therefore a DNO’s licence

•

•

if generation exists but is ignored in determining customer reliability levels, then the actual customer reliability may be in excess, even considerably in excess, of the minimum set by P2/5. if generation exists in a network, either its contribution to security should be included to maintain consistent treatment of all customers throughout the system, or the principles and criteria used in P2/5 should be completely changed. The latter is not the objective of the present review of P2/5, which is solely intended to be an update of Table 2 of P2/5, all principles and criteria remaining otherwise unchanged. A more fundamental review of P2/5 is possible and suitable approaches have been published previously52 by two of the authors of this report.


Page 147 of 150


(vii) Other issues Four other issues have also been considered in parallel with this project. These

were:-

disaggregation of demand and generation output • • • •

de minimus levels of capacity effect of remote generation effect of widespread anti-cyclonic weather

Although part of the original specification of this project, these were transferred

from this project and became the direct responsibility of WS3.

8.3. Implementation of Methodology There are three main ways of implementing the methodology. This report

recommends all three approaches; the one to be used depending upon circumstances, the data available and the need for precision in the output. The three approaches are:-

(i) look-up table(s) approach - this has the merits of simplicity and should be

free of erroneous applications. However, it is restricted to specific generation types and typical or average availability data. In the case of non-identical units, the capabilities are approximate since generation contributions are assessed separately for each type and unit size and then added arithmetically. This gives a pessimistic result, but generally of acceptable precision. This approach can also incorporate the concept of capping of the security contribution of generation so that it does not violate the materiality conditions which require generation to be less significant than the circuits

(ii) graphical approach - this is an extension of the previous approach and is

still reasonably simple. It accommodates a wider range of data and choices, but still requires approximate summations for non-identical units

(iii) computer program approach - this permits an unrestricted choice of data and mixed generation types. The prototype analysis package is implemented in MS Excel using VBA environment and a User Guide is provided for its easy application. The package calculates the security contributions only and these need to be assessed for compliance in the same way as performed with either of the two previous approaches.

An Application Guide has been written and included in this final report.

Page 148 of 150


8.4. Computer Program Package

One requirement of the project was to develop a software package that could perform the required assessments using the agreed methodology. This was an essential step in order to analyse many generation scenarios and to perform a wide range of sensitivity studies. It was agreed that the package need only be of a prototype form, not of a commercial grade and not commercially supported by UMIST. Instead, the specification was to produce a simple spreadsheet application package that an engineer with a reasonably knowledge of the approach could use. This prototype package has been developed in conjunction with Workstream 3. Its objective is to calculate the capability contribution to security of supply from distributed generation connected to a particular demand group. The application has been developed using Microsoft Excel® and Visual Basic for Applications®.

8.5. New Table(s) for Inclusion in P2/6

8.5.1. Base Table

The data used to create the “new” tables for inclusion in P2/6 were established from the Data Collection and Processing project53, also funded by FES. This separate project proposed availabilities to be used and these were agreed by WS3.

Table 2 of P2/5 is a derived table. The base table is Table A.1 of ACE Report 51,

although this may not be evident from P2/5 itself. Therefore the first requirement is to establish the base table. The following types of generation were identified by the Data Collection and Processing project and these are included specifically in the tables:-

• landfill gas • CCGT • CHP

sewerage using spark ignition sewerage using gas turbines others

• waste to energy • wind • small hydro

These tables are easy to use since they only need knowledge of generation type and number of units in the case of non-intermittent generation, and generation type and the degree of required persistence (Tm) in the case of intermittent generation. If the data underpinning the security capability used for these types of generation plant or other types of generation are being considered, then a more general set of tables and figures are also included.


Page 149 of 150


8.5.2. Materiality Test According to P2/5, generation should not be dominant and consequently materiality

criteria were introduced to prevent this happening. A revised set of criteria are proposed that are compatible with the methodology and P2/5. These are:

The contribution of generation specified in Table 2 (of P2/6) is based on the

assumptions that:- (a)

the cyclic rating of the largest distribution circuit is greater than F% of the total sent-out capacity of the N1 largest generating units

(b) the cyclic rating of the two largest distribution circuits is greater than F% of the total sent-out capacity of the N2 largest generating units

8.5.3. New “Table 2”

In the application of this revised set of criteria, the F factors are those specified in “Tables A.1, A.1a and A.1b” and N1 and N2 largest generating units correspond to the number of generating units to remove for a FCO (N1) and a SCO (N2) given in “Table A.1c”. On this basis, the base table of “Table A.1” simplifies to that “Table A.2”, the equivalent of “Table 2” in P2/5.

If the materiality test is not satisfied, then the generation would become the most

significant FCO. If this is unacceptable for technical and/or commercial reasons, then P2/5 states that separate risk and economic studies may need to be undertaken. However, a more practical alternative would be to use a process of “capping” the generation, as proposed in the report. This can be achieved by evaluating the maximum capacity that each generating unit could have such that the materiality criterion is satisfied.

8.5.4. Aggregating Units

If more than one type of generation exists in a demand group then the capabilities must be combined. The exact way is to aggregate them using the methodology. Alternatively the individual capabilities could be arithmetically summated. The summation results are always less than those given by the aggregation approach and therefore its use would always give results that err on the side of caution. In addition, the difference between the two values is generally small particularly as the number of units is increased. Since the aggregation approach increases the degree of precision and not necessarily the degree of accuracy, these results indicate that, provided the number of separate groups is relatively small, the simplicity of the summation approach justifies its use for evaluating the security capability of groups having non-identical units.

It is therefore suggested that the summation approach be used for most case studies

and that the aggregation approach be used only for those cases in which greater precision is deemed necessary, or the number and diversity of the groups give cause for concern about whether the summation approach is sufficiently precise.

Page 150 of 150

DG Contribution to network security_R3.pdf

Documents

widespread

de minimus

engineering

examine sco

daily load

synthesising

month winter

future energy