Open Networks Project FINAL Report-PUBLISHED.pdfReal time data exchange & Forecasting, Work Streem 1B Open Network Product 3 Energy Networks Association T +44 (0) 20 7706 5100 W E

Real time data exchange & Forecasting, Work Streem 1B Open Network Product 3

Energy Networks Association

T +44 (0) 20 7706 5100 W www.energynetworks.org.uk E [email protected] 1

Open Networks Project

Real time data exchange & Forecasting

Deliverable: P3 Report

10th January 2020

WS & Product Ref: WS1B P3 Restriction: Public

mailto:[email protected]




Document Control Version Control

Version Issue Date Author Comments

1 14/11/2019 WS1B P3 ENA Stakeholders Draft report for WS1B review.

2 13/12/2019 Ali Ahmadi, John West Final Draft report for approval.

3 10/01/2020 Ali Ahmadi Final Draft for publication.





Table of Contents

1 Introduction and P3 Sub-Deliverables Summary ....................................................... 6

1.1 Context.......................................................................................................................... 6

1.2 Future Worlds................................................................................................................. 6

1.3 Product 3 2019 Sub-deliverables Summary ........................................................................ 6

1.4 Further Product 3 work for 2020 ....................................................................................... 8

2 Identify Learnings from other trials/projects ............................................................ 9

2.1 Regional Development Programs ...................................................................................... 9 2.1.1 Transmission Issues....................................................................................................................... 11 2.1.2 South Coast Distribution Issues ...................................................................................................... 11 2.1.3 Service Conflicts ............................................................................................................................ 12 2.1.4 NGESO Managed Approach ............................................................................................................ 13 2.1.5 DSO Managed Approach ................................................................................................................ 13

2.2 Power Potential ............................................................................................................. 15 2.2.1 Active Power Services .................................................................................................................... 18 2.2.2 Reactive Power Services................................................................................................................. 18 2.2.3 DERMS Interactions with ESO and DSO ........................................................................................... 20 2.2.4 DERMS Simulation Results.............................................................................................................. 21

2.3 T.E.F. Projects ............................................................................................................... 25

2.4 Electricity Flexibility and Forecasting Systems (EFFS) ......................................................... 25

2.5 CLASS Project................................................................................................................ 26

2.6 ENTSO TSO-DSO Report ................................................................................................. 26

2.7 Project TERRE ............................................................................................................... 26

3 Manage Service Conflicts and Increase T/D Visibility .............................................. 29

4 Short Term Forecasting – 48 Hours Ahead ............................................................... 34

5 N-3 Operational Tripping Scheme (OTS) .................................................................. 35

5.1 Introduction to OTS and Capped Generation Arrangements ................................................ 35

5.2 Post fault curtailment of DER via DNO ANM to manage transmission constraints ................... 36

5.3 Options and Analysis and Recommendation ...................................................................... 36

5.4 System Architecture and operation .................................................................................. 36

5.5 Data exchange requirements ........................................................................................... 37

5.6 Implementation and system integration ............................................................................ 37

6 Service Conflicts and Data Exchange for National Emergency Conditions ............... 38

7 Appendix A – RACI Matrix ......................................................................................... 39

8 Appendix B – ENTSO-E TSO-DSO Report .................................................................. 40

9 Appendix C – Electricity Flexibility and Forecasting Systems (EFFS) ....................... 42

9.1 Forecasting ................................................................................................................... 44 9.1.1 Data Sources ................................................................................................................................. 45 9.1.2 Accuracy monitoring ...................................................................................................................... 45 9.1.3 Forecast locations and aggregation/ disaggregation ......................................................................... 46 9.1.4 Forecasting adjustments ................................................................................................................ 46

9.2 Conflict avoidance and synergy identification .................................................................... 47 9.2.1 Conflict definition ........................................................................................................................... 47 9.2.2 Conflict identification ..................................................................................................................... 47 9.2.3 Conflict resolution .......................................................................................................................... 48





List of Figures Figure 1 – Current RDP areas ........................................................................................................................... 9

Figure 2 – South Coast Transmission Constraints. ........................................................................................... 11

Figure 3 – UK Power Networks South Coast network for DSO Trial. .............................................................. 11

Figure 4 – Embedded Incremental Service Conflict Example........................................................................... 12

Figure 5 – Embedded Decremental Service Conflict Example. ........................................................................ 12

Figure 6 – Main communication paths between National Grid, UK Power Networks and DER ........................ 13

Figure 7 – Power Potential real and reactive power procurement approach from DSO...................................... 16

Figure 8 – DERMS network integration overview. .......................................................................................... 17 Figure 9 – DERs coordinated & controlled as Virtual Power Plant at a GSP. ................................................... 17

Figure 10 – Active power service commercial process. .................................................................................... 18

Figure 11 – Reactive power services commercial process. ............................................................................... 19

Figure 12 – DERMS interactions with ESO and DSO...................................................................................... 20

Figure 13 – High Level setup for DERMS Desktop Functionality Test in simulation environment. .................. 21

Figure 14 – Reduced schematic of the distribution network where DERMS was tested in simulation

environment. .................................................................................................................................................. 22

Figure 15 – Illustration of TERRE. ................................................................................................................. 27

Figure 16 – NGESO PAS-UKPN DERMS interface. ....................................................................................... 30

Figure 17 – PAS dispatch screen. .................................................................................................................... 31

Figure 18 – the Azure APIM Gateway and IIB architecture ............................................................................. 32 Figure 19 – Transmission & Distribution integration via ICCP ........................................................................ 32

Figure 20 – Capped Generation Approach. ...................................................................................................... 35

Figure 21 – EFFS Workstreams. ..................................................................................................................... 42

Figure 22 – EFFS core functions ..................................................................................................................... 43

Figure 23 – PV output data ............................................................................................................................. 44

Figure 24 – Forecasting accuracy monitoring .................................................................................................. 46

List of Tables Table 1 – Data exchange required for NGESO and DSO managed approach. ........................................... 10 Table 2 – NGESO services embedded at the distribution level.............................................................. 14 Table 3 – Stage 1 Optimisation, Priority Stack Mode. ........................................................................ 14 Table 4 – Stage 2 Optimisation Mode (Reshuffle). ............................................................................ 15 Table 5 – Stage 3 Optimisation Mode (Cost and Volume Optimisation).................................................. 15 Table 6 – Simulation initial condition............................................................................................. 22 Table 7 – Active Power Services Instruction Results per DER.............................................................. 23 Table 8 – GSP initial voltage and reactive power settings. ................................................................... 23 Table 9 – Active Power Services Instruction Results per DER.............................................................. 24 Table 10 – Real-time data exchange signals and methods of communications. ......................................... 29 Table 11 – Forecasting data exchange signals and methods of communications. ....................................... 34

References

[1] UK Power Networks Providing Power Services From Distributed Energy Resources to Transmission

System Operator via a Centralised DERMS Platform, 25th International Conference on Electricity

Distribution, CIRED 2019.

[2] WPD EFFS, Forecasting Evaluation Report, June 2019,

https://www.westernpower.co.uk/projects/effs

[3] ENTSO, TSO-DSO Report, An Integrated Approach to Active System Management, April 2019,

https://docstore.entsoe.eu/Documents/Publications/Position%20papers%20and%20reports/TSO-

DSO_ASM_2019_190416.pdf

[4] Power Potential Project, https://www.nationalgrideso.com/innovation/projects/power-potential.

[5] KASM Project, https://innovation.ukpowernetworks.co.uk/projects/kasm/.


https://www.westernpower.co.uk/projects/effs

https://docstore.entsoe.eu/Documents/Publications/Position%20papers%20and%20reports/TSO-DSO_ASM_2019_190416.pdf

https://docstore.entsoe.eu/Documents/Publications/Position%20papers%20and%20reports/TSO-DSO_ASM_2019_190416.pdf

https://www.nationalgrideso.com/innovation/projects/power-potential

https://innovation.ukpowernetworks.co.uk/projects/kasm/




Acknowledgments We would like to thank Ali R. Ahmadi, Kanan Ganakesavan, Steve Gough, Sarah Hadfield, John West and WS1B P3 Stakeholders, for their expertise, contribution and assistance throughout all aspects of this work. We would like to thank Farina Farrier for her help and support in scoping the P3 deliverables and resourcing. We would also like to thank Matthew White, Ben Godfrey, Ian Pashley, Steve Shaw, Ian Povey, Jason Brogden and Sotiris Georgiopoulos for their valuable contribution towards enabling whole system approach in U.K.





1 Introduction and P3 Sub-Deliverables Summary

1.1 Context

As part of the Open Networks Project’s Workstream 1B (Whole Electricity System Planning and Transmission–Distribution Data Exchange), Product 3 seeks to determine the requirements for real time data exchange and short term forecasting between the National Grid Electricity System Operator (NGESO) and emerging Distribution System Operators (DSOs) to better utilise distribution connected resources. Product 3 is looking across a number of ongoing trials to assess how real time data exchange and forecasting work best in different situations. To enable the effective utilisation of distribution connected resources for balancing, and for economic and efficient network operation in operational timescales, NGESO and DSOs will need an ongoing exchange of information to share service requirements and to identify any ongoing network limitations. Product 3 is taking the outputs of ongoing trials to better define the information transfers to enhance transmission/distribution system coordination and control. It is evaluating how these information transfers would change depending on the complexity of distribution networks and for the future worlds that are being assessed under Workstream 3. Much of this work on information and data transfer is being informed by Regional Development Programmes (RDPs) that are running between network companies to trial different approaches for managing distribution connected resources. For example, the NGESO-Western Power Distribution (WPD) RDP is exploring the deployment of distribution level services in a relatively straightforward network where there are relatively few service providers and few network constraints. The NGESO- UK

Power Network (UKPN) RDP is exploring the deployment of distribution level services in a more complex network environment where there are multiple service providers and multiple network constraints.

1.2 Future Worlds

Following detailed assessment and stakeholder consultation, Workstream 3 of Open Networks has determined that Future World B (Coordinated DSO-ESO Procurement and Dispatch) is the least regrets implementation path for the short to medium term. Both the NGESO-WPD and NGESO-WPD RPDs are examples of Future World B where the ESO and DSO can jointly co-ordinate the procurement and use of distribution resources. The work carried out in this product will inform the wider implementation of World B arrangements. Over the longer term, other future worlds including World A, known as Distribution System Operator (DSO) Coordinates, and World D, known as Electricity System Operator (ESO) coordinates, may also be taken forward. The work carried out under this product will also inform the data exchange requirements for these outcomes.

1.3 Product 3 2019 Sub-deliverables Summary

All of the outputs from Product 3 in 2019 are included in this written report. A P3 sub-deliverable summary in 2019 is tabulated below where the default colour marks the completed deliverables and *Red marks ongoing deliverables which are dependent on RDP projects between DNOs and NGESO. Through the first half of 2019, the responsibilities and requirements for information transfer to support the effective deployment of services under DSO World B were further assessed in sub-deliverables A and B1. As comparisons between the Future Worlds were being developed by Workstream 3, the

requirements for Worlds A and D were also considered by extrapolating the work for World B. For World B, and for Worlds A and D, accountabilities have been identified, data requirements have been defined as have the preferred methods of operational data transfer.





Sub-Deliverable & Output Baseline

Date End of 2019 Position Status

Output A – Publish RACI for “simple” and “complex” networks in Worlds A, B & D.

Feb-19 RACI produced for Worlds A, B & D. (See Appendix A)

Complete

Output B1 – Describe information exchange between DSO and ESO for Worlds A, B & D (“what” & “how”).

Mar-19

RACI matrix converted into a simple table which defines ‘what’ & ‘how’ data/information needs to be exchanged between DSO and ESO for Worlds A, B and D.

(See Section 4)

Complete

*Output B2 – Inform and update this view of information exchange through the RPD results.

Oct-19

Detailed RPD simulation results from the NGESO-UKPN RPD will inform information & data exchange. The algorithm design to obtain simulation results to start once National Grid have

signed-off the design requirements.

Ongoing, 2020 subject to NGESO and DNOs commercial agreement.

*Output C – Identify information & control system architecture for Worlds A, B & D. Support roll-out.

Dec-19 Further RPD results & review of other trials & projects will inform architecture requirements.

Ongoing, 2020 subject to NGESO and DNOs commercial agreement.

Output D – Identify data exchange requirements for short and medium term

forecasting.

Apr-19

Templates were issued and completed to establish Network Company capabilities. A review of the EFFS project was also undertaken.

(See Section 5)

Complete

Output E – Consider service conflicts and identify data exchange for national emergency conditions.

Jun-19 (See Section 8) Complete

Output F – Identify learnings from other trials/projects and update Product 13 report.

Jul-19

Relevant trials & projects were allocated and learnings were summarised through August & September. (See Sections 2.3 to 2.6 and Section 3)

Complete

Output G – Ensure operational interfaces

for TERRE and Wider BM access remain aligned to Product 3.

Dec 2019

TERRE has been delayed to July 2020 due to RTE (France) delays. Wider BM Access is scheduled to go live on 11th December 2019.

Some prequalification of secondary BMUs is beginning. Operational interfaces are likely to be required in some areas to manage ANM impacts.

Complete

With respect to sub-deliverable C, all the Distribution Network Owners (DNOs) in the UK recognise the need to coordinate data across the GSP boundary with the ESO and consider that ICCP is the most appropriate protocol for this data exchange. DNOs are at different stages in establishing ICCP links with the NGESO. For example, as part of the 2019 NGESO-UKPN RPD, UKPN have established a fully functional Inter-Control Centre Protocol (ICCP) link with NGESO exchanging over 100 real-time data

points. The State Estimation running on UKPN’s Distribution Management System (DMS) platform is converging at Ninfield Grid Supply Point (GSP) and the distribution network below the GSP. Further data exchange is needed to ensure the accuracy of the DSO State Estimation is satisfactory for system operation decision making. The short-term forecasting requirements to identify and manage service requirements in operational timescales have been further assessed during 2019 as part of sub-deliverable D. This builds on previous

Open Networks work. Again, the information that will need to be exchanged between network companies has been identified as well as the preferred methods of data transfer. As well as drawing from RDPs, as part of sub-deliverable F, Product 3 has reviewed the outputs of other projects and trials that have been assessing information exchange for distribution level services. These include ongoing UK based innovation projects such as Power Potential and Electricity Flexibility and Forecasting System (EFFS), and related international work.

Through 2019 Product 3 has continued the real-time data exchange discussions from 2018 and described the approach for information exchange between DSOs and ESO for Worlds A, B & D. Sections 4 and 5 of this report describe the list of signals and methods of data exchange for real-time operation and operational forecasting. The most desirable approach is World B – Joint Procurement due to its flexibility when applied to distribution networks with different complexities.





This report has drawn heavily on ongoing trials/projects including Power Potential and the RDPs. The Power Potential project has not yet reached field trials but section 3.2 describes the infrastructure designed and built to facilitate Power Potential project. It explains the information flow between ESO and DSO as well as simulation results obtained from the operation of DERMS when real and reactive power instructions are issued from ESO. For the NGESO-UKPN RDP, section 3.1 of the report describes the N-3, headroom/footroom and loadflow/optimisation approaches proposed under the RDP. The UKPN’s RDP N-3 design and development work started in December 2019, the results of the trial is expected in 2020. Alongside this more general work on operational data transfer, Product 3 has supported the development of the Transmission-Distribution interface to support the implementation of Project Trans European Replacement Reserves Exchange (TERRE) and Wider System Access. Whilst this report is not able to provide any conclusions on the outcomes of Power Potential and RDP projects in 2019 in the context of data exchange, the approaches presented for data exchange in this report are insightful and provide a platform to conduct trials. In 2020, results of the trials will be presented with conclusions on benefits and drawbacks.

1.4 Further Product 3 work for 2020

Now that Open Networks has concluded that World B (Coordinated DSO-ESO Procurement and Dispatch) is the least regrets pathway to DSO (which is supported by the analysis in this paper for real time and short-term forecasting data exchange), Product 3 will focus further on World B in 2020. Work in late 2019 and early 2020 will draw on more detailed modelling and trials to be carried out as part of the NGESO-UKPN RDP and that will further inform World B. Potential service conflicts will be considered in greater detail and any additional information transfers to manage these will be identified. Further trials to inform information exchange will be carried out during 2020. NGESO, UKPN and SSE have agreed to progress with N-3 Operational Tripping Schemes (OTS). Whilst work on Connect & Manage and Service Conflicts has been delayed by the need to finalise commercial arrangements, provided agreements are reached between NGESO and DNOs, it is anticipated that by the end of 2020, we can capture the learnings and begin to write the good practice guide for the implementation of data exchange specifications, N-3 OTS and management of service conflicts. Hence the high level deliverables for 2020 are:

Identify data exchange specifications & implementation guidelines for Whole System activities

including Operational Tripping Schemes, Connect & Manage and Service Conflict Management

based on Regional Development Programme (RDP) outcomes.

Develop a good practice guide for the implementation of data exchange specifications for

Operational Tripping Schemes, Connect & Manage and management of service conflicts.

Telemetry and Telecommunications – Develop resilience, redundancy and latency requirements for the systems used to manage distribution resources.

Advise on the ESO/DSO data exchange and the methodology for operational forecasting.

Following the successful trials of UKPN RDP in the South Coast, the Real Time Data Exchange Product 3 must take the learnings and codify the data exchange information. In addition, Development of a method of orchestration for DSO/DNOs must be considered via CIM. The Business as Usual process must be clarified and documented. Data points have been agreed on a bilateral basis between DSO and ESO for the RDP trials and would benefit to be standardised in a common format. ICCP mapping are known to be codified in CIM and this needs to be investigated.





2 Identify Learnings from other trials/projects

This section provides a summary of recent projects where data exchange and system coordination have been considered between DNOs and NGESO. These projects include RDPs involving NGESO and DNOs, the Power Potential project, the Fusion project, the CLASS project and the EFFS project.

2.1 Regional Development Programs

RDPs have been set-up by NGESO with UKPN, ENW, Scottish Power Transmission (SPT), Scottish Power Distribution (SPD), WPD and SSEN to provide detailed analysis of areas of the network (see Figure 1) which have large amounts of Distributed Energy Resources (DER) and known transmission/distribution

network issues in accommodating that DER. The RDPs promote a whole-system approach. The aims are to accommodate additional capacity for DER ahead of, or instead of, major network reinforcements and to reduce national balancing costs when managing thermal constraints on the transmission network, avoiding conflicts of services, and enhancing T/D system coordination and control. The RDP approach can innovate and push the boundaries of current thinking with a “design by doing” approach to resolving known issues, pushing towards DSO type solutions and informing thinking for

the DSO debate. By solving specific case studies and improving outcomes for customers in innovative ways, it is possible to make progress faster than the more conventional method of agreeing changes in approach at industry forums before trying out these changes. While there are risks that working in this way leads to a lack of standardisation across the GB network, by using RDPs as case studies for the Open Networks project, WS1B’s Real Time Data Exchange Product can take the learnings from the RDPs and making these available to all network operators.

Figure 1 – Current RDP areas

The RDPs are investigating two solutions to enhance transmission and distribution system coordination and control; a simple constraint headroom assessed approach (managed by NGESO) and a DSO load flow assessed approach (managed by DSO). Both of these solutions fit with Future World B (Coordinated DSO-ESO Procurement and Dispatch). Initial sensitivity analysis studies on UKPN and WPD networks, indicated the NGESO managed approach can be achieved with few signals/data exchange where:

There is good understanding of the services conflict likelihood and pattern of occurrence.

Low number of constraints and network configurations which can change the sensitivities.





The few signals for the NGESO managed approach correspond to headroom/footroom on each distribution constraint and fixed sensitivities under intact conditions. When the distribution system becomes more complex (as is already happening in some distribution networks) with an Active Network Management (ANM) system managing multiple active constraints; then the simple constraint headroom assessed approach (NGESO managed) is likely to require more signals to achieve acceptable results. In these cases, the use of Distributed Energy Resource Management Systems (DERMS) and ANM systems on DNO networks, where network data, schedules and information are available, provides the infrastructure to effectively manage the conflicts of services and distribution system optimisation under the load flow assessed approach (DSO managed). NGESO-UKPN Regional Development Programme The remainder of section 3.1 provides further detail on the NGESO-UKPN RPD. For this RPD, the SE-coast network was chosen as UKPN and NGESO identified that transmission capacity issues were beginning to impact on customer connection dates. DER developers rely on the ability to be able to connect to the network quickly, so this was a potential barrier to the growth of renewables in the area. UKPN engaged with their customers both at local and regional events, and established the need to move quickly to resolve the network constraints affecting the connection of further DERs. Through the RDP, UKPN and NGESO developed a set of objectives which were consulted upon, agreed through UKPN DER customer forums and ultimately delivered (or are in the process of being delivered). The “simple constraint headroom assessed approach” and the “DSO load flow assessed approach” will facilitate the following real time data exchanges (for the trial initially) via the Inter-Control Centre Protocol (ICCP) link between NGESO and DSO:

Table 1 – Data exchange required for NGESO and DSO managed approach.

Simple constraint headroom assessed

approach (managed by NGESO)

DSO load flow assessed approach

(managed by DSO)

DSO to provide the following signals to NGESO:

Headroom and footroom information at

points of constraint.

Visibility of ANM operations.

Visibility of passive DG/DER operation.

Visibility of flexibility instructed.

Visibility of flexibility contracted.

DER Sensitivity per key constraints under

intact network condition.

Background data for network modelling

(flows. Topology, switch states, impedance,

ratings etc.).

NGESO sends services instructions in the form of priority stack and volume.

NGESO to provide the following signals to DSO:

NGESO boundary constraint information.

Visibility of flexibility instructed within

distribution network.

Visibility of flexibility contracted within

distribution network.

Background data for network modelling

(flows. Topology, switch states, impedance,

ratings etc.).

Order of DER dispatch (stack priority),

Nominal service, Effective Service and Cost

of MW Services.

DSO sends services optimization results to NGESO.

For the simple constraint headroom assessed approach; the DSO calculates distribution constraint headroom/footroom and sends it to NGESO with DER effective sensitivities against each constraint. NGESO will then optimise dispatch of DER resources without service nullification. The operation of ANM systems maximising the output of Flexible DERs can potentially, at times, conflict with embedded NGESO services - negating service output. For the DSO load flow assessed approach in the South East Coast of UK, NGESO issues their desired MW services stack to DSO (received by ANM). The ANM then performs optimisation on cost or volume in order to deliver the maximum possible MW





volume at GSP without violating any distribution constraints. In this approach, NGESO inputs sufficient details into DERMS, such as pricing, volume and availability within procurement timescales. The NGESO-UKPN RPD evaluates the test results (including optimisation) for each of the approaches and captures benefits/drawback of each approach for whole system operation and meeting all DSO/NGESO operational requirements in the most economic and efficient way.

2.1.1 Transmission Issues

Figure 2 shows the South East Coast transmission network and associated large generation in-feeds. Also highlighted is the overload and voltage collapse that results from one of a number of constraining

network conditions most significantly; the loss of the double-circuit between Kemsley and Cleve Hill. Under this scenario, there will be a requirement to constrain generation to manage the system.

Figure 2 – South Coast Transmission Constraints.

2.1.2 South Coast Distribution Issues

The UKPN South Coast network consists of the following Active constraints:

Figure 3 – UK Power Networks South Coast network for DSO Trial.

The distribution constraints are described below:

PX1 and PX2 (i.e. the 132kV lines between Richborough and Canterbury South).

PMA and PGA (i.e. the 132kV lines between Canterbury South and Canterbury North).

SGT1 and SGT2 at Canterbury North.

PF Route: Hastings Main – Appledore 132kV (not shown in the diagram).

PP Route: Appledore 132kV – Ruckinge 132kV (not shown in the diagram).

PR Route: Ruckinge 132kV – Tee Towards Sellindge 132kV Grid (not shown in the diagram).





2.1.3 Service Conflicts

The ENA Open Network Project have identified the potential for ANMs to, at times, conflict with embedded NGESO services - negating service output. NGESO services embedded in the DNO network may be impacted by ANMs either:

For services which increment - if the ANM is active at the time (or doesn’t have sufficient headroom)

then the service effect will be negated seconds later following ANM action to curtail alternative

generation.

For services which decrement - if the ANM is active at the time the controlled DER will “fill in” the space

made by the service with the extent of the fill in being determined by the volume of other DG being

curtailed prior to the decrement service.

An illustrative example of an incremental service conflict is given in Figure 4 where an ANM is actively curtailing DER to 70MW in order to control the flow on a DNO circuit within its rating limit of 50MW. There is an embedded NGESO service, Short Term Operating Reserve (STOR), within the ANM Zone not itself under ANM control. Should the STOR service be called upon by the System Operator to generate 20MW, seconds later the service’s output would be nullified by the ANM pulling back an equal amount of DG to return the circuit to within its rating.

Figure 4 – Embedded Incremental Service Conflict Example.

An illustrative example of a decremental service conflict is given in Figure 5 where an ANM is actively curtailing distributed generation to 70MW to control the flow on a DNO circuit with a rating limit of 50MW. There is an embedded NGESO service, Enhanced Frequency Response (EFR), within the ANM Zone not itself under ANM control. Should the EFR service automatically absorb power in response to

a rise in system frequency as per its service requirement, the ANM would detect the spare capacity and seconds later the service’s output would be nullified by the ANM releasing an equal amount of previously curtailed DG.

Figure 5 – Embedded Decremental Service Conflict Example.





The two approaches proposed to manage and resolve conflict of services are NGESO managed approach and DSO managed approach.

In the NGESO managed approach; the DSO will calculated distribution constraint headroom/footroom and will send it to NGESO with DER sensitivities. The DER effective sensitivities against each constraint can be shared in advance as standing data or as an additional signal. The NGESO will then perform optimisation on their side to determine the most economic dispatch of embedded NGESO services without service nullification.

In the DSO managed approach; NGESO issues their priority order MW services stack to DSO (received by ANM), the ANM then performs optimisation in order to deliver the maximum possible MW volume at GSP without violating any distribution constraints (maximise headroom/footroom use). In the DSO managed approach, the NGESO also inputs DER contract details into DERMS, such as pricing, volume and availability within procurement timescales. This is explained further in section 2.1.5.

2.1.4 NGESO Managed Approach

NGESO would carry out optimisation analysis on the transmission network and looks for the most

efficient solution taking into account transmission and distribution energy resources. NGESO checks the actions are within distribution network capability using distribution constraint data provided in real time from the DSO.

2.1.5 DSO Managed Approach

The proliferation of DERMS and ANM systems on DNO networks where all the network data, schedules and information are available, provides the infrastructure to effectively manage the conflicts of services and distribution system optimisation. Figure 6 shows the main communication paths and components for end-to-end trial of DSO approach with UKPN. DERMS is designed to interface with NGESO dispatch tools. ANM is designed to manage distribution constrains and conflict of services. Both DERMS and ANM are operating via a centralised DMS platform called PowerOn.

Figure 6 – Main communication paths between National Grid, UK Power Networks and DER

The Conflict of Services algorithm with 3 stages are proposed to be designed inside UKPN’s ANM scheme in order to manage conflicts of services between ANM and any type of embedded NGESO Services listed in Table 2 on points 1-8. It should be noted that the name or exact nature of NGESO services will change over time but that the solution will be flexible enough to define the name and key characteristics of any service. The Conflict of Services algorithm will classify each NGESO services uniquely in order to manage nullifying the NGESO instruction as a result of DERs sitting behind a distribution constraint managed by ANM. The conflict of services must be managed between ANM and NGESO Services with any combination of services type (however, the availability results of each NGESO services must be presented to NGESO and DSO simultaneously.





Table 2 – NGESO services embedded at the distribution level.

Service

Name Description

Service Conflict

Characteristics

1 Connect and

Manage

Generation connections whose output is subject to T constraints). Will be paid by

the NGESO if constrained.

Bidirectional

2 N-3

intertripping

Generation connections to be armed and tripped if a certain T fault materialises.

Will not be paid.

Decremental

3 Power

Potential

Firm DERs signed under Power Potential providing Active and Reactive Power

Services for NGESO.

Bidirectional

4 STOR Dispatched electronically from ENCC. Usually instructed following a generation

loss. Available within 20mins & runs for < 4 hrs.

Incremental

5 EFR Permanently enabled for frequency response. Bidirectional

6 Demand Turn

Up

Several hours ahead of time. Currently by email as new service under

development.

Decremental

7 BM Balancing Mechanism Units embedded in the distribution network. Which can be

incremented or decremented, or armed for frequency response by electronic

instruction

Bidirectional

8 TERRE Trans European Replace Reserve Exchange. Incremental

Three modules will be developed by UKPN for conflicts of services management; Stage 1 Priority Stack Mode, Stage 2 Optimisation Mode (Reshuffle) and Stage 3 Optimisation Mode (Cost and Volume Optimisation) as well as a separate module with direct NGESO dispatch.

Stage 1 Priority Stack: NGESO issues their priority MW services stack to DSO (received by ANM), the ANM assesses how much of NGESO’s desired volume can be delivered at the GSP without violating any distribution constraints. The ANM must then declare available capacity per DER (or for Power Potential only aggregated capacity per GSP) to NGESO via DERMS and/or ANM. Desktop analysis results for Stage 1, under intact network condition is presented Table 3. In this case, NGESO is seeking 33MW of service and the analysis indicates that 10.5MW of service can be provided across the 5 DERs against the priorities given.

Table 3 – Stage 1 Optimisation, Priority Stack Mode.

DER Incremental

Priority Nominal service

(MW) Declared service (MW) Stage 1 Optimisation

aDER1 1 15 5 5MW on aDER1, No violation

aDER4 1 10 10 3.5MW on aDER4, due PX1

O/L

aDER5 2 7 6 0 MW, due PX1 O/L

aDER6 3 15 10 0 MW, due PX1 O/L

aDER9 4 20 2 2 MW, No violation

TOTAL 39 33 10.5

Stage 2 Optimisation Mode (Reshuffle): NGESOESO) issues their desired MW services stack to DSO (received by ANM), the ANM then optimises the stack order in order to deliver the maximum possible MW volume at GSP without violating any distribution constraints (maximise headroom/footroom use). Reshuffle could be done across all services or within a group of services. The ANM must then offer available capacity per DER (or aggregated capacity per GSP) to NGESO via DERMS and/or ANM. Desktop analysis results for Stage 2, under intact network condition is presented in Table 4. In this case, by rearranging the priority stack order, all 33MW of the requested service could be provided across the 5 DERs.





Table 4 – Stage 2 Optimisation Mode (Reshuffle).

DER Incremental


(MW) Declared service

(MW) Stage 2 Optimisation

(Reshuffle)

aDER1 N/A 15 5 9.5MW on aDER1, No

violation

aDER4 N/A

10 10 3.5MW on aDER4, due PX1

O/L

aDER5 N/A 7 6 0 MW, due PX1 O/L

aDER6 N/A 15 10 0 MW, due PX1 O/L

aDER9 N/A 20 2 20 MW, No violation

TOTAL 39 33 33

Stage 3 Cost and Volume Optimisation: NGESO issues their desired MW services stack to DSO (received by ANM), the ANM then performs cost optimisation in order to deliver the maximum possible MW volume at GSP without violating any distribution constraints (maximise headroom/footroom use). Under Stage 3, the ESO inputs DER contract details into DERMS, such as pricing, volume and availability within procurement timescales. The services availability optimisation results from stage 3 must be visualised in both table (i.e. Table 5) and cost curve.

Table 5 – Stage 3 Optimisation Mode (Cost and Volume Optimisation).

DER Incremental


(MW) Declared

service (MW) Cost

(£/MW) Stage 3 Optimisation

aDER1 N/A 15 5 5 13 MW

aDER4 N/A 10 10 12 0 MW

aDER5 N/A 7 6 12 0 MW

aDER6 N/A 15 10 12 0 MW

aDER9 N/A 20 2 4 20 MW, No violation

TOTAL 39 33 33

As demonstrated in Table 4, the DSO managed approach at Stage 2 achieves the full 33MW capacity required by NGESO without violating any distribution constraints. In addition, Table 5 demonstrates that DSO managed approach at Stage 3 achieves the total 33MW capacity requested by NGESO at cheapest cost. Hence the DSO managed approach which has full visibility of the distribution network via DERMS/ANM/PowerOn, can effectively optimise the distribution network and offer services to NGESO.

2.2 Power Potential

The Power Potential project is an innovation project between UKPN and NGESO, looking to utilise the

operation of Distributed Energy Resources in order to provide real and reactive power services at GSP

in the South East Coast.

The project is targeting 4 existing GSPs in the South Coast; Bolney, Ninfield, Sellindge and Canterbury

North. The following diagram illustrates the Power Potential approach from Stage 1-Stage 6:





Figure 7 – Power Potential real and reactive power procurement approach from DSO.

The transmission network, and the areas within the distribution network at this location are at the limit of capacity for transferring generation away from the area. This means for particular faults or conditions on the transmission network, voltage levels at certain points reach values which can violate statutory voltage limits. This constraint is preventing additional generation from being able to connect to the South East transmission or distribution networks. To enable more generation to connect, large-scale network investment is traditionally required. The Power Potential project team aim to find an innovative solution to facilate faster and cheaper alternative DER connection. In the meantime, increasing operating costs are being incurred by the System Operator in managing the existing limitations which impact customers and consumers.

The project aims to create a regional reactive power market which will help defer network reinforcement

needs in the transmission system. The project has the following key deliverables:

A commercial framework using market forces to create new services provided from DER to NGESO via UKPN.

A technical and market solution known as Distributed Energy Resources Management System (DERMS) to support technical and commercial optimisation and dispatch. It includes gathering bids from DER and presenting an optimised view of the services to NGESO split by GSP. The DERMS will be installed in UKPN’s control room.

At a high level, the project solution is envisaged to work as follows:

Gather commercial availability, capability and costs from each DER;

Run power flow assessments to calculate possible availability of each service at the GSP. Once the assessment is complete, a range of service availability and costs will be presented to NGESO

as intra-day availability (or 24 hour rolling window) taking into consideration DER bids, their effectiveness and what the distribution network can allow at the time of service due to current running arrangements. With this information, NGESO the system operator, will decide the level of services to be procured; and





On the day of the response, NGESO will instruct the services to UKPN, the DERMS solution will instruct each DER to change their setpoint as required and will monitor their response.

ESO will define the real and reactive power services instruction at the interface 400kV level (GSP) in the form of:

P service: target set-point (ΔP400 kVset).

Q service: target set-point (V400kVset), droop and dead-band.

Figure 8 – DERMS network integration overview.

DERMS receives the instructions from NGESO and optimises the operation of the distribution network to deliver the required services at the GSP, as shown above. It coordinates the operation and control of DERs as a Virtual Power Plant (VPP). This concept is illustrated below.

Figure 9 – DERs coordinated & controlled as Virtual Power Plant at a GSP.

For the reactive power service, DER are expected to operate under a local voltage droop control scheme whose set-point VDER

set can be adjusted by DERMS. This occurs after receiving a voltage set-point instruction V400kV

set from NGESO or after a large transmission voltage change V400kVmeas, which allow DER

to provide dynamic voltage support.





2.2.1 Active Power Services

The Following summary describes the proposal for calculation, presentation and settlement of Active Power Services via Power Potential. Using the real-time demand, generation and running arrangement, the DERMS load flow engine calculates the available dispatchable Active power (P) Volume at the GSP. DERMS does not make dispatching decisions itself for Active Power (P); however, it can receive an instruction from NGESO to dispatch Active Power. The dispatchable Active power (P) is presented to NGESO in the form of Delta Volume at GSP. In other words, granular view of every DER is not presented to NGESO in real time. The real power (P) increase (generation) and curtailment (down) will be represented in Delta at the GSP. Alongside the dispatchable Active power, the Reactive power (Q) Lag (injecting) and Lead (absorbing) will be represented in Delta Volume at the GSP.

Ahead of real-time, the DERs would have input their P bids with a utilization price (£/MWh) for each commercial interval of the delivery period by 14:00 on the day ahead using a web interface with DERMS. This tendering process is presented below:

Figure 10 – Active power service commercial process.

Using the DER commercial information, real-time demand, generation and network running arrangement, the DERMS loadflow engine calculates the available dispatchable active power (ΔP) volume at each GSP. The DERMS does not make dispatching decisions itself for active power; however, it can receive an instruction from NGESO to dispatch active power.

The dispatchable active power (ΔP) is presented to NGESO in the form of delta volume at each GSP. In other words, granular view of every DER is not presented to NGESO in real-time. For each GSP (400kV delivery point), NGESO can provide DERMS a target set-point MW service instruction.

2.2.2 Reactive Power Services

This section describes the proposed calculation, presentation and dispatch of reactive power services via Power Potential during real-time.

Ahead of real-time, during the auction process (before gate close) each DER provides a declared availability; in absolute terms; and an availability price (£/Mvar/h) and utilisation price (£/Mvarh) for





each commercial interval of the delivery period using a web interface with DERMS. At the close of the commercial auction, selected DERS are contracted to be available to provide Mvar support for each commercial interval during the delivery period.

The diagram below illustrates how the auction will look like for participating in reactive power market:

Figure 11 – Reactive power services commercial process.

Using the DER commercial information, real-time demand, generation and network running

arrangement, the DERMS loadflow engine calculates the available dispatchable reactive power (ΔQ) volume at each GSP, both lag (injecting) and lead (absorbing). It would also send an arming instruction for the reactive power service at the start of the instruction period (to change the DER operating model voltage droop control and provide dynamic voltage support). For each GSP (400kV delivery point), NGESO can provide DERMS a target voltage set-point kV service instruction, to be delivered with a droop characteristic within a certain dead-band, to get Mvar support from DER.

The high level operation of the Power Potential services during real-time can be then summarised as follows:

DERMS receives service requests from NGESO per 400KV delivery point.

Based on these service instructions the DERMS builds a production schedule for DER to be delivered in each operating interval over the remainder of the current commercial operating period. (max 24hrs).

For each operating interval, DERMS has a number of DERs that are contracted to provide power services for a specific price.

At the start of each commercial interval the DERMS instructs each DER to enable their set-points

and confirms that this has been done.





2.2.3 DERMS Interactions with ESO and DSO

The DERMS Solution technical architecture will provide the capability to handle the following services:

Dynamic Voltage Service

o for Low Voltage Management

o for High Voltage Management

Active Power Service

o MW Re–Dispatch (Active Power) Service (for Thermal Constraints)

The solution will be implemented as a redundant server-based software product located within the UK PN’s ICT network and interfacing with various UKPN’s systems and external parties. UKPN’s systems include the existing UKPN Distribution Management System (DMS), PowerON, developed by General Electric. External parties interacting with the solution include the National Grid’s Platform for Ancillary Services (PAS) and Distributed Energy Resources (DER) located in the Distribution Network. At the core of the solution is the facility for DER to provide reactive and active power services to National Grid for their use to operate and secure the transmission system. The volume of P and Q services available for procurement is updated and presented to NGESO every 10s. The key components for end-to-end integration between DSO and ESO are as follows:

1. PAS (Platform for Ancillary Services). 2. PAS –DERMS web interface. 3. DERMS Control Platform

a. Service Functionality. b. Network Security Analysis. c. Forecaster. d. Outage Planning Inputs. e. Commercial Implementation. f. Mandatory Trial Interface. g. Future Availability.

4. IIB (IBM Integration Bus). 5. PowerOn screens and RTU Logic integration via DNP3. 6. NGESO to UKPN ICCP link. 7. PowerOn CIM via ICCP.

Figure 12 – DERMS interactions with ESO and DSO.





2.2.4 DERMS Simulation Results

This section presents the simulation results obtained from testing the functionality of the DERMS when providing real and reactive power services to NGESO during real-time operation. The tests were conducted with DERMS real-time algorithm running on a Microsoft Azure pre-production simulation environment. The desktop test arrangements are presented below:

Figure 13 – High Level setup for DERMS Desktop Functionality Test in simulation environment.

The DERMS end-to-end simulation based functional testing included two different instructions:

1. P service: target ΔP (MW) at 400 kV GSP. 2. Q service: target Vset voltage (kV) at 400 kV GSP, droop percentage and dead-band in kV.

The simulation was tested on a 100 bus bar distribution network connected to transmission network at 400kV Canterbury North GSP via two supergrid transformers (SGT1, SGT2). The distribution network contained 5 DERs which can provide Power Potential active and reactive power services:

Maidstone and Canterbury DERs at 132kV. ALL, Barming and Sheepway DERs at 33kV.





A reduced schematic of the distribution network connected to Canterbury North GSP and location of the 5 DERs are shown in figure below:

Figure 14 – Reduced schematic of the distribution network where DERMS was tested in simulation

environment.

2.2.4.1 Active Power Services Results

The initial DER condition presented in Table 6 illustrates the technical real/reactive power capability of each DER and its associated bids.

Table 6 – Simulation initial condition.

DER Commercial/Technical Data

DER No. 1 2 3 4 5

Maidstone Canterbury ALL Barming Sheepway

£/MW1 2 1 3 4 5

£/Mvar2 2 1 3 4 5

Max MW 100 200 30 20 20

+ve Mvar 50 50 10 10 10

-ve Mvar -50 -50 -10 -10 -10

DER Initial Setpoint

DER P set (MW) 100 200 30 20 20

DER Q set (Mvar)

0 -48 -10 -10 -10

1Instantaneous active power utilisation price. 2Instantaneous reactive power utilization price. Reactive power availability price not reflected in table as DER will have been

already procured.





As shown in Table 6, all DERs are operating at their maximum real power export setpoint (i.e. base case study with initial setpoints). The first study in Table 7 with 0MW instruction at GSP level, illustrates the DER’s initial condition.

Table 7 – Active Power Services Instruction Results per DER.

DER No. Transformer flows ΔP at GSP

DER output

1 2 3 4 5 SGT1 SGT2 Total Flow

P Instruction

P (MW) 100 200 30 20 20 -171.28 -169.12 -340.4 1st Study: 0 MWs Q (Mvar) 0 -48 -10 -10 -10 73.75 72.25 146

P (MW) 100 200 30 20 20 -171.28 -169.12 -340.4 2nd Study: 50 MW Q (Mvar) 0 -48 -10 -10 -10 73.75 72.25 146

P (MW) 100 126 30 20 20 -132.99 -131.32 -264.3 3rd Study: -74MW Q (Mvar) 0 -50 -10 -10 -10 74.84 71.83 146.7

P (MW) 100 200 30 20 20 -171.28 -169.12 -340.4 4th Study: 0MW Q (Mvar) 0 -48 -10 -10 -10 73.75 72.25 146

P (MW) 0 0 0 0 0 11.41 11.25 22.66 5th Study: -370MW Q (Mvar) 0 0 0 0 0 -3.97 -4.59 -8.56

The second study in Table 7 demonstrates that the instruction of extra 50MW export at GSP level is not possible because the DERs were already operating at their maximum export capacity, hence the DER outputs and SGT flows do not change. The third study, shows the instruction of 74MW reduction at GSP level. The DERMS algorithm instructs the cheapest DER, which is DER 2 (Canterbury DER), at a cost of £1/MW, to reduce from 200MW to 126MW. The fourth study shows the return to the initial conditions by instruction a delta P of 0MW at GSP level. The fifth study, demonstrates the instruction

of 370MW reduction at GSP level. In order to achieve the instruction, DERMS has instructed all DERs to operate at 0MW in order of cheapest cost first. Hence the net export across both SGTs have reduced from -340.4MW to a net import of 22.66MW, whilst keeping all distribution network constraints within limits.

These studies show effective optimisation of DER real power services by DERMS as a VPP, in order to achieve NGESO’s instructions at the GSP whilst keeping all distribution network constraints within limits.

2.2.4.2 Reactive Power Services Results

The DER commercial and technical data and initial set-points are the same as Table 6. However, the initial GSP voltage setings are as shown below in Table 8:

Table 8 – GSP initial voltage and reactive power settings.

Reactive Power Initail Setpoints

GSP Deadband (kV) 1

GSP Droop (%) 4

GSP Target Voltage (kV) 400

Slack Bus bar Voltage (kV) 404





This initial condition is demonstrated by the sixth study in Table 9, with NGESO request of a target voltage of 400kV, the cheapest and most effective DERs are instructed by DERMS to operate at their leading reactive power, absorbing vars.

Table 9 – Active Power Services Instruction Results per DER.

DER No. (P in MW and Q in Mvar)

Trans. flows

Slack Volt. (kV) GSP Target Voltage (kV)

DER output

1 2 3 4 5 Total Flow

Voltage Achieved

Q Instruction

P 100 200 30 20 20 -340.4 404.00 6th Study: 400kV Q 0 -48 -10 -10 -10 146

P 100 200 30 20 20 -341.22 404.35 7th Study: 404.97kV Q 0 0 0 0 0 59.67

P 100 200 30 20 20 -341.41 405.49 8th Study: 415kV Q 50 50 10 10 10 -114.88

P 100 200 30 20 20 -338.4 403.48 9th Study: 395kV Q -50 -50 -10 -10 -10 206.52

The seventh study shows the change in GSP target voltage set-point to 404.97kV, the slack bus is

operating at 404kV, hence the DERMS instructs DERs to operate at a zero reactive power as the target voltage and slack bus bar voltage is within the deadband.

The eighth study shows a GSP target voltage increrase to 415kV. Hence instructing the cheapest and most effcetive DERs first, DERMS requests all DERs to operate at their maximum lagging reactive power limit, injecting vars.

The nineth study shows the GSP target voltage reduced to 395kV. Hence instructing the cheapest and

most effcetive DERs first, DERMS requests all DERs to operate at their maximum leading reactive power limit, absorbing vars.

The Canterbury North GSP connected to the transmission network is very stiff which means changing the voltage profile must be coordinated across the South Coast GSPs to achieve the best results.

These studies show effective opimisation of DER reactive power services by DERMS as VPP, in order to achieve NGESO’s voltage instructions at GSP whilst mantaining all distribution network consraint within limits. This section has presented the technical framework proposed in the Power Potential project for coordination and control of transmission and distribution networks in Great Britain during real-time operation. This project will improve interaction between NGESO, UKPN and DER, through a technical and commercial solution. Power Potential is a significant step towards the transition of DNO to DSO, and informing the future role of the NGESO.

This section also presented the simulation results obtained from testing the functionality of the DERMS when providing real and reactive power services to NGESO. The simulation results demonstrate that the DERMS algorithm works by coordinating and controlling the operation of DERs as VPP in order to meet NGESO’s requirements at GSP level.

The successful demonstration of DERMS functionality in real-time under simulation environment shows effective interactions and mechanics between ESO and DSO can lead towards procuring the cheapest

resources available in order to manage transmission network constraints. The proliferation of DERMS on distribution network where all the network data, schedules and information are available, provides the infrastructure to effectively manage/optimise the distribution system constraints and offer services to NGESO.





2.3 T.E.F. Projects

The T.E.F. Projects comprise of TRANSITION (SSEN with ENWL), EFFS (WPD) and FUSION (SPEN); three Ofgem Network Innovation Competition (NIC) projects focusing on developing and testing a number of elements associated with the transition to a DSO model and smart gird architecture. TRANSITION and FUSION are progressing through the system development and trial design phases with primary trials taking place in 2021 through to 2023 (albeit smaller tests are proposed in 2020 to support early testing of industry principals). Data exchanges haven’t been finalised, but initial high level design work outlines the base requirements that, where practicable, has been made available on the associated project websites. The intention is to base the solution on a data integration framework that is capable of presenting information in standard formats, including CIM, and test the assumptions made in this document and the subsequent work through 2020. TRANSITION and FUSION have aligned activities to maximise learning and will build on the relevant EFFS outputs expected in 2020 and 2021. Electricity Flexibility and Forecasting System (EFFS) is a shorter project, operating over three years, thus it is expected to lead the T.E.F. engagement on data exchange and forecasting initially, acting as a testbed where viable. The learning generated through the upcoming EFFS trials will inform the later TRANSITION and FUSION trials with both projects mandated to directly adopt outputs where practicable, offering enhanced value to network customers. As the EFFS project is the primary T.E.F. conduit for data exchange and forecasting in the short term, it is discussed separately in Section 2.4 and Appendix C.

2.4 Electricity Flexibility and Forecasting Systems (EFFS)

The EFFS project is being delivered by WPD and project partners including AMT-SYBEX, Smarter Grid Solutions and NGESO. The project is focussed on enabling the DSO role through the development of capabilities for forecasting and flexibility service co-ordination. Further detail is provided in Appendix C. EFFS looks to support DSO transition by trialling a system to enable the planning and dispatch of flexibility services in operational timescales. Software is being developed to forecast network demand,

generation and storage. This will help distribution networks to co-ordinate flexibility services at a local level. The potential to utilise services for energy balancing is recognised as well as distribution network management. Building on work by Open Networks, EFFS will determine arrangements for co-ordination and conflict resolution with other parties that are using flexibility services. One EFFS objective is to provide specifications for data exchange between the DSO and other parties. Design documents for the various functional areas of EFFS have been published indicating data items

and processes for data exchanges. Interface specifications will be finalised in the coming months. Where possible these will be open source such that they can be implemented in different technical solutions. Testing these across the FUSION and TRANSITION NIC projects would form part of the trials. Operational forecasting is a large part of EFFS. The anticipated operational horizons are within day, day

ahead and week ahead. It is envisaged that the forecasting capability would be directly integrated into the EFFS solution with automatically scheduled runs delivering the required profiles. EFFS would support a range of forecasting algorithms (multi linear regression, heuristic/machine learning etc.), together with data sources and interfaces. EFFS has not yet reported on co-ordination and prioritisation. It recognises that without co-ordination, actions by one party could negate the actions of another, and the potential for actions by other parties introduces additional uncertainty into forecasting. Workshops with NGESO, other DNOs and other stakeholders were held to help define the scenarios that are considered to be conflicts, methods for how the conflicts would be identified and a simplified matrix of service conflicts between NGESO and DSOs to be used in the trial to test the approach. Once proven the service conflict matrix could be expanded to include conflicts between other parties such as DSO-DSO or Supplier-DSO.





2.5 CLASS Project

Customer Load Active System Services (CLASS) is now a business as usual activity and is provided by the DNO to the ESO under its normal commercial arrangements for procuring balancing services. The roles of ESO and DSO were not defined in the project. Communications are in accordance with the ESO specifications, in particular NGESO’s specification for the Platform for Ancillary Services (PAS). For frequency response services, performance data is provided on request through the provision of data files via email. For Fast Reserve services, a web interface exists for communication. Data exchange is from DNO to ESO. The information exchanged over the web interface includes nomination, activation and deactivation requests and acknowledgements, real time (every 15s) metering data, performance data (1 minute data files). The purpose of CLASS is for DNOs to provide balancing services to ESO. This project involved ESO/DSO system coordination with DNO control room to ESO control room information/data exchange.

2.6 ENTSO TSO-DSO Report

This report was produced by various European ESO and DSO organisations to consider a broader framework for Active System Management (ASM) [3]. Some further detail is provided in Appendix B. ASM is defined as the actions taken by ESOs and DSOs to monitor and ensure that grid operational parameters are within satisfactory ranges. These actions comprise operational planning processes, grid monitoring and control, data exchange and interactions with market parties delivering services.

The report focusses on active power management and solving congestion through the market based allocation of flexibility services. It recognises that ESOs and DSOs each have roles to play as system operators and neutral market facilitators. National flexibility resource registers are proposed with information on the location of connection points that can provide flexibility services to system operators. The registers would make all potential flexibility resources visible to DSOs and ESOs. They might also be used to register connections and to settle flexibility services between market parties.

Specific flexibility products are not considered in detail but products for portfolio optimisation, balancing and congestion management should be sufficiently aligned to permit the optimal market-based allocation of flexibility between these different purposes. Forecasting is discussed in the report. Grid reinforcement planning (longer term) and grid utilisation (shorter term) are noted. If the capacity of the electricity grid is insufficient to cope with changes,

flexibility services can be used as a complement to grid reinforcement.

2.7 Project TERRE

Under the EU Electricity Balancing Guideline (EBGL), “standard” European wide services for electricity balancing and frequency management are being introduced in participating countries. Three different types of service cover the overall timescales for reserve services – “Frequency Containment Reserves”, “Frequency Restoration Reserves” and “Replacement Reserves”. Project TERRE (Trans European Replacement Reserves Exchange) is establishing a European platform for the exchange of “Replacement Reserves”. (These are reserves scheduled to help manage energy imbalance or to maintain system margins after short term services have been deployed.) This will be the first ”standard” service and it is planned to begin by mid-2020. Cross border exchanges will depend on available interconnection capacity. For example, exchanges between Great Britain and Continental Europe will use capacity on the existing High Voltage Direct Current (HVDC) interconnectors. The ESO has been leading TERRE implementation in Great Britain. The ESO would interface with the European wide platform for TERRE and it would dispatch services through the GB balancing market





arrangements. A significant GB code change has been the introduction of “Secondary Balancing Mechanism Units” (Secondary BMUs) to enable more straightforward participation for smaller service providers in the GB balancing market.

Figure 15 – Illustration of TERRE.

As well as supporting TERRE, the Secondary BMU arrangements will provide opportunities for DERs to provide other services to the GB balancing market. DNOs will need to interface more closely with the ESO to facilitate effective DER participation. European Network Codes and GB code frameworks recognise that DNOs/DSOs need to be involved in the management of reserve services to ensure the effective use of DER given the increasing potential for distribution network constraints. In GB for example, if a participating DER is connected within an area covered by an ANM arrangement, the DER’s contribution to the TERRE service might be counteracted by the ANM scheme operation. Project TERRE was identified in early-2018 as an area where ESO-DNO interfaces need to be developed. The main interactions are at i) Prequalification and ii) Utilisation. i) Prequalification – For DER looking to participate in TERRE, the ESO and DNOs check to ensure that the DER can participate effectively. The ESO collects data from prospective DER providers and for providers connected to distribution networks, data would be forwarded to DNOs. This data would include any generation or demand facilities that are being aggregated to form a BMU (or Secondary BMU), MW capacities, locations, technology types and whether there are operational restrictions in their Connection Agreements. The respective DNO’s would review this data and any potential limitations to service provision would be flagged to the ESO and to the prospective service provider. ii) Utilisation – Just ahead of the service being utilised, the ESO and DNOs may need to liaise to ensure that DER are able to participate effectively given prevailing network conditions. For near real-time interactions, it is anticipated that direct DNO involvement won’t be required when TERRE initially goes live as the limitations through ANM systems and other DNO network restrictions are likely to be small. As levels of DER service providers and ANM use increase, DNO information for near real-time DER participation is likely to be required. Options to manage near real-time ESO-DNO interactions could include ESO notification to DNOs of prospective service providers and DNO’s highlighting any limitations to the ESO or to the service providers. The focus of TERRE is to provide the European wide standard “Replacement Reserve” service. The GB changes will also enable the wider use of DER for other GB balancing services. Specific operational forecasting is not being developed as part of the TERRE/Wider System Access work. Additional forecasting may be required for DNOs to manage near real time interactions.





The project has not considered co-ordination and prioritisation in detail. Any processes agreed through the wider Open Networks work could be incorporated into the processes to manage TERRE and Wider System Access. The project is intended to bring down energy balancing costs by making a wider range of providers available to the ESO. Through TERRE, a pan-European market will be established for “Replacement Reserves”. Through Wider System Access, it is anticipated that additional distribution connected resources will be made available to provide GB balancing services.





3 Manage Service Conflicts and Increase T/D Visibility

The following table demonstrates the list of real-time signals and method of data exchange for managing conflict of services between DSO and ESO in addition to providing increased transmission/distribution system visibility:

Table 10 – Real-time data exchange signals and methods of communications.

World A – DSO World B – Joint Procurement World D – ESO DSO to ESO ESO to DSO DSO to ESO ESO to DSO DSO to ESO ESO to DSO

Wh

at

Da

ta t

o E

xch

an

ge

●Available volume and

cost per GSP (MWs and

MVArs). ●Optimised Output of

DERs.

●Cost information

per DER.

●Volume per GSP.

●Time Duration of

when the service is required.

For “simple” case (from

section 2.1.4):

●Headroom/footroom.

●Fixed sensitivities.

For “complex” case:

●Optimised results

(stage 1-3 from DSO managed approach from section 2.1.5).

●Active Curtailment of

DERs by ANM (MWs).

●Service instructions (stack

and volume).

●Stack Order for DERs.

●Costs.

●Arming the Stack and

confirmation of DER volume.

●Visibility of FDGs.

●Visibility of DERs Connected

Customers. ●Visibility of ANM.

●All CBs status and Analogues.

●FDG Setpoints.

●Active Curtailment of DERs by ANM

(MWs).

●Maintenance Schedule.

●Fault Outages.

●Operational liaison re network

changes.

●Control and Dispatch.

●Response to Operational Liaison

(i.e. moving open points and running arrangement changes).

Data Types Method of

Communications

Data Types Method of

Communications

Data Types Method of Communications

Ho

w t

o E

xch

an

ge

th

e

Da

ta

Analogues + CB status. ICCP. Analogues + CB status. ICCP. Analogues + CB. ICCP.

N-3 instructions + alarms.

ICCP. N-3 instructions + alarms.

ICCP. Control and Dispatch. ICCP.

Availability of volumes. Web service. MW dispatch decision. Web service. Maintenance and ongoing operational liaison.

Web services.

Volume of instructions. Web service. Service conflicts. Web service.

Delivered Outputs. Web service.





In summary, Table 10 illustrates in World A “DSO Co-ordinates”; the DSO can do both simple and complex distribution network optimisation to manage service conflicts and maximise the ESO’s access to participating DER real and reactive power services whist providing new revenue streams for DER customers. In World A, relatively limited ESO data is required to be sent to the DSO in the form of volumes of services, orders and duration of time. In World D “ESO Co-ordinates”, the ESO responsibilities would include design and operation of the ANM system, whilst the DNO would retain responsibility for delivery and maintenance of the wider distribution network. The ESO would be responsible for the implementation of ANM on the distribution network. The ESO can do both simple and complex optimisation on the distribution network. However, larger volumes of data exchange are required from the distribution network to ESO compared to World A. World B (Coordinated DSO-ESO Procurement and Dispatch) is the most flexible approach where the ESO and DSO are working together to utilise distribution network assets and facilitate DER access to the network and new revenue streams. In World B, there are more options on where to draw the line to minimise the volume of data exchange depending on the complexity of the distribution network flows and actors. Both parties can work together to minimise data exchange and manage conflict of services. There is more flexibility and options to manage the service conflicts and increase T/D network visibility. The Method of Communications for Data exchange are explained further below: PAS is a web-services link to National Grid’s Platform for Ancillary Services (PAS). The NGESO main interface with DERMS/ANM is the Platform for Ancillary Services (or PAS). PAS is a new NGESO control and monitoring solution to support and enhance existing and future reserve and frequency services. It sits in NGESO control room and has been designed for a range of ancillary services. For example, for Power Potential, PAS will receive from DERMS/ANM volume availability and cost for each Power Potential service at the GSP level as demonstrated in Figure 16 [4].

NMS (Power On)

DERMS

RTU

Co

ntr

ol &

Co

nfi

g si

gna

ls

AN

M d

atap

oin

ts

NM

S d

ata

po

ints

Fails

afe

dat

apo

ints

Me

asu

red

va

lue

s

Sta

tus

sign

als

DER CONTROLLER

Control Signals

Customer issued datapoints & measured values

Status signals

Measured valuesANM Datapoints

Loca

l/R

emo

te

Loca

l Dat

apo

ints

Control Engineer

Local Operative

For a full list of signals exchanged between the NMS, RTU, and DER, refer to the points list EDS 05-9610a

PAS Bid Response

Bid Award

Bid Request

NG UKPN DER

Remote select

NMS Datapoint

Test scopeWeb Interface

Figure 16 – NGESO PAS-UKPN DERMS interface.





It will also allow NGESO control room engineers to provide instructions for the different services, in real-time. An example of PAS instructions in the form of voltage setpoint dispatch screen is shown in Figure 17 where voltage setpoints at GSP are sent to the DSO:

Figure 17 – PAS dispatch screen. The parameters that affect the instruction are:

400kVTarget Voltage: Target voltage at the GSP. VActual (DERMS): Actual GSP voltage as recorded by DERMS. Deadband %: band to keep the voltage within certain predetermined limits. Droop %: voltage droop is the intentional loss in output voltage. The GSP Voltage Droop

displays the droop value in Mvar/kV units. The IBM Integration Bus (IIB) and Azure aspects of the Power Potential solution are supporting elements – tested with the associated PAS-DERMS and DERMS web interface functions detailed later in this document, rather than tested as individual components. UKPN Enterprise Service Bus (ESB) tool – IBM Integration Bus (IIB) is used as the messaging platform to provide the integration services in Power Potential project. IIB is a proprietary software tool from IBM and is a leading middleware product in the UK. The capabilities of IIB have been used to develop the scalable and robust integration solution required in Power Potential. Azure API Management (APIM) Gateway is a cloud based API Gateway service. It can be configured with security policies like Restricted Caller IPs (IP Filtering), Rate Limit, Validating Authorisation tokens etc., which will be executed to validate the incoming requests before forwarding them to DERMS through IIB. These IS hosting functionalities are relevant to two parts of the DERMS architecture.

1. The NGESO PAS – DERMS web services communication goes through the Azure API Management Gateway and IIB.

2. Both the DERMS Pre-production and Production UKPN Gridview (Dashboard) and DER Dashboard applications are hosted within the App Service environments on Azure cloud. Dashboard applications fetch the data using Azure Gateway service end points which are configured to retrieve the data from DERMS backend systems.

All the DERMS Test Environment services hosting the Dashboards (UKPN/ DER) Applications, CIM Core, Service Modules, Simulator and Future Availability Modules are deployed on Azure Virtual Machines within the UKPN Azure cloud infrastructure.





Figure 18 – the Azure APIM Gateway and IIB architecture

In order to assist UKPN and NGESO in their roles of DSO and ESO the KASM, Power Potential and RDP projects have installed an ICCP link between their respective Control Rooms. The ICCP link provided the medium so that metering from points of interest on the respective distribution/transmission system can be sent to the other party’s network management system [5].

UK Elect r icit y Syst em

UKPN (Dist r ibut ion N et wor k Oper at or )

DM S

Cont r ol Room

Engineer

N at ional Gr id (Tr ansmission Syst em Oper at or )

IEM S

Cont r ol Room

Engineer

ICCP

Figure 19 – Transmission & Distribution integration via ICCP

ICCP is the global standard for providing data exchange over wide area networks (WANs) between utility control centres. The protocol was specifically design for the purpose that it is being used for in this implementation: the exchange of real time information between utility control centres. It enables the exchange of real-time and historical power system monitoring and control data, including measured values, scheduling data, energy accounting data, and operator messages. NGESO already use ICCP to communicate with other DNOs and the Scottish System Operator. The ICCP components are integrated into the SCADA systems. The data exchange provides the end user of a real-time view of both networks as the ICCP data are integrated with the existing SCADA data

API Mgmt Gateway

IIB Service Bus

DERMS

ServiceSimulator and FutureAvailability

CIM Core

NG PAS

Web Server

Grid View - UKPN

Grid View - DER CimphonyCore





already in their control system. The conformance blocks within the ICCP data defines and limits the scope of the data that can be exchanged over the link. The ICCP will provide a facility (if the data are consumed by PowerOn Fusion) where control room engineers will have visibility of the current operational conditions of the transmission system. This will help them to better understand the operation of the distribution systems and hence assist when undertaking important operational actions. The ICCP is an IEC point-point protocol, so there are limitations with using ICCP as a bus to pass signals from NGESO-PowerOn over ICCP, then over PowerOn-DERMS ICCP. This arrangement has been part of Power Potential’s agreed design since 2018, and re-uses the existing ICCP link set up for the KASM project. However the protocol limitation was only just identified as a barrier at final integration stage. Originally identified this was for passing to DERMS National Grid’s selected voltage references at the four GSP voltages, then identified this would also be required for Q flows at SGTs since these are not monitored on UKPN network. NGESO has added P and Q to the V points being shared over ICCP. Identifying this limitation is a significant learning to the wider industry and the project will be expected to provide detail engineering rationale and may lead to engagement/feedback to the IEC standardisation body responsible for ICCP standard. A manual workaround was developed by the project to be able to pass the signals, and successfully verified on 17th October 2019 to pass the Ninfield 400kV voltage from NGIEMS to PowerOn to DERMS in around 1ms (below the timescale resolution of the timestamp). This was then extended to other voltage, active and reactive power analogue data which needed to be sent across the link.





4 Short Term Forecasting – 48 Hours Ahead

The following table demonstrates the list of signals and method of data exchange between DSO and ESO to facilitate short term operational forecasting on transmission/distribution sides:

Table 11 – Forecasting data exchange signals and methods of communications.

World A – DSO World B – Joint Procurement World D – ESO DSO to ESO ESO to DSO DSO to ESO ESO to DSO DSO to ESO ESO to DSO

Wh

at

Da

ta t

o E

xch

an

ge

●DSO every half hour Forecast (Demand

and Generation at 11kV and 33kV feeder level). Presented to ESO at GSP level.

●ESO National

Forecast (Demand and Generation).

●Interconnector

Forecast Flows.

●DSO every half hour Forecast

(Demand and Generation at 11kV, 33kV and 132kV feeder level). Presented to ESO at GSP level.

●DSO) Outage.

●E&W forecasted 132kV active

flows.

●ESO National Forecast

(Demand and Generation).

●Interconnector Forecast

Flows.

●N/A ●N/A




Ho

w t

o E

xch

an

ge

th

e

Da

ta

●Web with CIM data exchange format ●Web with CIM data

exchange format

●Web with CIM data exchange

format

●Web with CIM data exchange

format

●Web with CIM

data exchange format

●Web with CIM data

exchange format





5 N-3 Operational Tripping Scheme (OTS)

5.1 Introduction to OTS and Capped Generation Arrangements

Historically, Operational Intertripping Systems (OTS) are used to allow transmission connected generation to operate freely pre-fault. In the event of defined network faults occurring, an intertripping signal would be sent to certain generators to remove a predefined volume of MW generation from the system immediately post-fault to manage thermal, voltage or stability issues.

As part of the RDPs it has been proposed to extend OTS arrangements to DNO ANMs to curtail DER output under certain N-3 conditions3 to manage thermal overloads. This will enable maximum connection capacity to be available to DER in an economic fashion. UKPN, NGESO, WPD and SSEN have jointly agreed to implement N-3 OTS, hard-tripping DERs during abnormal network conditions (i.e. faults) on the transmission network. The three DNOs and NGESO have agreed to implement a Capped Generation Approach as part of their RDPs. (The design of N-3

OTS arrangements between ESO and DSO may not immediately concern SP, NPG and ENW.)

Figure 20 – Capped Generation Approach.

In broad terms under the Capped Generation Approach, NGESO would send DSOs a CAP per GSP (see Figure) to ensure that a post-fault boundary capability would not be exceeded in the event of a fault. The DSO’s ANM system will subtract the CAP from the total DER generation behind that GSP to determine the curtailment required for N-3 event under that GSP. For example, GSP1 has 95MW total DER capacity, NGESO sends a CAP of 70MW, therefore to determine the curtailment required for N-3 under that GSP; 95MW-70MW=25MW. To achieve this curtailment, the ANM system would send a 10MW curtailment signal to Solar1_1 and a 15MW curtailment signal to Solar 1_2. The design and development of the N-3 OTS algorithm with Capped Generation Approach began on June 2019 between UKPN and NGESO.

3 An N-3 condition in the transmission system is defined as a planned single circuit outage followed by a fault

outage on double circuit.





5.2 Post fault curtailment of DER via DNO ANM to manage transmission constraints

To manage transmission constraints post fault using DER, it is necessary to install a system to curtail DER quickly by an automatic action in the event of a rare but significant N-3 condition in the transmission network that severely reduces the capacity available in real time. It is proposed to use NGET’s standard OTS system to detect faults on the transmission system and send a tripping/de-load signal on a per GSP basis to a DNO wide system ANM, which will curtail generation in the required GSPs. It is proposed that NGESO will instruct the DNO on the requirements for generation curtailment via an ICCP link. DNO’s have indicated a wish to fully automate their response to such instructions.

5.3 Options and Analysis and Recommendation

There are various logics/ways that can be used to curtail DER. Some examples are, predefined MW blocks, delta MW reduction, and cap on DER. While there are pros and cons of each option, “Cap on DER” appears to be the most positive. This option will fit well with the existing ESO processes and allows full ANM potential pre-fault. Hence the

recommendation is to request DNO’s to provide the capped generation option. This is a cap on the generation in the GSP that is controlled by the ANM and not the GSP in total. (The latter would be very difficult to manage owing to the volume of generation and demand the ANM cannot control). The selected option allows NGESO to send the required MW values to the DNO as an instruction from an iEMS screen via an ICCP link this will enable the DNO to act on the instruction automatically if they chose to do so ensuring the scheme is designed such that any real-time algorithms the ANM is running

do not impede the scheme’s ability to achieve the overall response time.

5.4 System Architecture and operation

The diagram below shows the simple overall coms architecture of the total N-3 intertripping system.

There are primarily 4 key stages to the intertrip process – prior to arming, arming, tripping/triggering

and disarm/restore. In the prior to arming process (i.e. steady state), the DNO will provide visibility of all DER controllable generation available to be curtailed under each Grid Supply Points (GSP) to NGESO. NGESO will continue to monitor the outputs and system conditions.





The next phase is arming. This is where NGESO identifies a possible N-3 event and makes a request to the DNO for the MW volume to be available to be reduced per GSP. At this stage, it is an indicative request only to be ready. The ANM will not actually curtail any generation at this stage. The ESO will also instruct the relevant TO OTS to start monitoring circuits for fault outages. The third phase is the tripping/triggering phase where, if there is a double circuit fault in the relevant system, a trip signal will be sent from the TO OTS to the DNO ANM. Upon receipt of the TO trip signal the ANM will have an agreed duration to reduce the generation to the cap level set at the arming stage. Currently the agreed timescale to achieve this is 30 seconds. The fourth and the final phase is disarming/restoring when the system conditions are back to normal. NGESO will coordinate the whole process of resuming the system operation. It is worth noting that the third phase (generation intertripping) does not happen very often. The national electricity system’s historical data shows that a particular double circuit fault in the transmission system is typically a 1 in 100-year event. The double circuit fault happening during another single circuit outage (which lead to an N-3 scenario) makes it an even smaller probability event.

5.5 Data exchange requirements

To make the whole intertripping arrangement work, a number of data points are expected to be exchanged between ESO, TO and DNOs. At a high level it is summarised as follows:

DNOs will share real time controllable DER behind each GSP with the ESO. This will be via

ICCP link.

The TO will share outage and circuit status information. This is currently a business as usual

process.

The ESO will send arming, restore instruction to the DNO and TO.

The TO will send trip signal to the DNO via a dedicated link.

5.6 Implementation and system integration

Currently this arrangement is being implemented in the south coast where transmission constraints for certain N-3 conditions are either caused or exaggerated by increasing DER volumes. UKPN, WPD and SSE are working with NGESO and NGTO to implement N-3 intertripping arrangements. UKPN currently have an ICCP link with NGESO. WPD and SSE are in the process of implementing similar ICCP links. At the same time the NGTO is working with all three DNOs to establish dedicated trip signal links from their OTS systems (Sellindge OTS in the South East and Melksham OTS in the South West). The whole N-3 system is being built to be full dual redundant in such a way that a single software or hardware failure will not result in the whole system being unavailable. The only exception would be the control and comms links to individual DER from the DNO ANM. This is owing to the cost benefit consideration. The NGESO and DNOs have agreed adequate failsafe actions to manage such single comms failures with individual DERs. This level of redundancy will make the operational intertripping of DER equivalent to traditional transmission protection arrangements and will allow NGESO to fully utilise it in managing the whole system. It is expected that the first such system will be operation from Q2 2020 in the South East followed by South West in early 2021.





6 Service Conflicts and Data Exchange for National Emergency Conditions

Where the services are aligned - Service price marginally rises with increase in requirement, so national services are procured elsewhere where they can be satisfied. Distribution services take priority through pricing. T and D to work on better information sharing to enable co-optimisation, though this might naturally occur through markets. Where there is conflict - if this is just out to the markets, it will be resolved by price, but at a cost to consumers. The conflict to be identified ahead of procurement by sharing information and formal co-optimised processes. This should be built bottom up, i.e. distribution needs satisfied, balance notified to ESO to them procure national requirements from the residual. For real time issues, a system control hierarchy needs to be agreed. Again, preferred proposal would be bottom up, acknowledging for severe system issues, ESO has veto capability. The following key points must be considered during emergency conditions:

Maintaining system security and stability nationwide.

The order where ESO performs the emergency actions.

Under National Emergency, DSO/DNOs are under ESO instructions. However, DSO/DNO also

advises the ESO on settings which keeps the network secure.

During major events such as loss of transmission connected generation (e.g. 9th of August 2019

event), Low Frequency Relays (LFR) operating and disconnection customers on the distribution

network, ANM curtailing DG customers and contributing to frequency collapse.

As the number of DERs and ANM systems increase on the DSO system, the system emergency

instructions could be bottom-up with provisional information from ESO with ESO coordination

capability.

DSO providing available DER MWs and MVArs volume increase/decrease at GSP for system

security.

Emergency codes need to be amended to consider ANM and DER operation during national

emergency actions.





7 Appendix A – RACI Matrix

P3 stakeholders have developed a RACI where simple and complex networks have been mapped to World A (DSO Coordinates), World B (Coordinated DSO-ESO procurement and dispatch) and World D (ESO Coordinates) and defined responsibilities for DSO/ESO, methodology of Service Conflicts management and dispatch. The Matrix is used for mapping key high level interactions for each world.

DSO Joint Procurement ESO

Word A World B World D

TSO DSO TSO DSO TSO DSO

Data supporting operational planning/forecasting

Generation dispatchable I R,A I R,A R A, I

DER dispatch priority N/A R,A I R,A I R,A

Generation passive N/A R I R,A R A, I

Demand fixed N/A R A R R,A N/A

Demand flexible N/A R I R,A R A, I

Head room / foot room N/A R C R,A R A, I

System data (system topology) N/A R I R R A, I

forecasted Flows N/A R I R R N/A

Constraints DNO N/A R I R A R

Constraints TSO R I, A R I R N/A

MW Dispatch decision I R, A C R,A R, A I

Dispatch duration N/A R, A I R,C R, A N/A

Data exchanged in real time

network configuration (switch status) N/A R I R R N/A

Generation (including ANM control) N/A R I R,A I R,A

Flows (analogues) N/A R I R R I

MW Dispatch deployment N/A R I R,A R N/A

manage conflict N/A R, A I R,A R, A N/A

Emergency action R A R,A I R,A N/A

Exception / Post event data

Conflict occurred N/A R I R R I

Error (between request and deployment) N/A R I R R N/A

cost settlement N/A R C R,A R N/A

RACI represents: R - Responsibility, A - Accountable, C - Consulted, and I – Informed





8 Appendix B – ENTSO-E TSO-DSO Report

This report was produced by various European ESO and DSO organisations to consider a broader

framework for Active System Management (ASM) [3]. ASM is defined as the actions taken by ESOs and DSOs to monitor and ensure that grid operational parameters are within satisfactory ranges. These actions comprise operational planning processes, grid monitoring and control, data exchange and interactions with market parties delivering services. The report focusses on active power management and solving congestion through the market based allocation of flexibility services. The report recognises that ESOs and DSOs each have roles to play as system operators and neutral market facilitators. The report proposes national flexibility resource registers with information on the location of connection points that can provide flexibility services to system operators. It may also be possible to use the register to register connections and to settle flexibility services between market

parties. The objective of the register is to gather and share information on potential sources of flexibility. In the report, the focus is on the provision of flexibility services to system operators. Qualifying connections would be registered by the connecting system operator. All potential flexibility resources would then be visible to DSOs and ESOs.

The report is largely focussed on congestion management and the use of market based flexibility services for active power management. Specific products are not considered in detail but there is agreement that products should comply with the needs of system operators to perform economically efficient congestion management. These requirements should be clearly specified to enable successful product design and development. This requires sufficient transparency to enhance the mutual understanding of system operator requirements and market party capabilities.

Flexibility products for portfolio optimisation, balancing and congestion management should be sufficiently aligned to permit the market-based allocation of flexibility between these different purposes with the objective of an optimal allocation. This requires interoperability between products to enable the exchange between markets. Products should be either an option (available capacity) for the purchasing system operator or a direct activation. This option can always be forfeited if there is no need to activate the product for congestion

management. Availability products have to be designed properly to avoid a decrease in market liquidity due to non-activation of contracted products. Furthermore, different situations in Member States might require more short or more long-term products or a combination of both. The availability of short-term products like day ahead or intraday could be guaranteed through forward markets, which trade short-term products for specific periods in advance. To ensure the right balance between availability and market liquidity, DSO and ESO co-ordination is needed.

Forecasting is discussed in the report. In the forecasting phase, the planning of grid reinforcement (year and months ahead) and grid utilisation forecast (months ahead, weeks ahead, day-ahead and intraday) are taken into account. If the capacity of the electricity grid is insufficient to cope with the expected rise in consumption or production of electricity, or new usage patterns start impacting normal grid operation, flexibility services can be used as a complement to grid reinforcement. Forecasting is undertaken in different timeframes. The accuracy of the predicted flow of electricity in a certain area typically improves closer to real-time. Some forecasts consist of long-term planning analysis made years in advance (before the preparatory phase) and some forecasts are updated and performed up until real time (for example using real-time weather data and remote monitoring devices on the grids).





System operators should have access to good schedules with relevant locational information, to perform proper forecast for congestion management and make efficient and secure decisions. Different market models are discussed for congestion management and balancing. These include i) separate ESO and DSO congestion management, ii) combined ESO and DSO congestion management and iii) combined congestion management and balancing. It is recognised that different market models will be appropriate for different national situations such that a single model is not preferred. The project should make a wider range of providers available to the ESO.





9 Appendix C – Electricity Flexibility and Forecasting Systems (EFFS)

The EFFS Project is being delivered by WPD and a number of project partners including AMT-SYBEX, Smarter Grid Solutions and the National Grid ESO. The project looks to support DSO transition by developing and trialling a system to enable the planning and dispatch of flexibility services in operational timescales. Software is being developed to forecast network demand, generation and storage using project developed algorithms. This capability will help distribution networks to manage the co-ordination of flexibility services at a local level. Building on work by Open Networks, EFFS will determine optimal arrangements for co-ordination and conflict resolution with other parties that are using flexibility services. One of the objectives of EFFS is to provide specifications for data exchange between the DSO and other parties. EFFS is structured around 4 workstreams as shown below in Figure 21:

Figure 21 – EFFS Workstreams.

The project is primarily focussed on enabling the DSO role through development of capabilities for forecasting and flexibility service co-ordination. The ESO is a participant in EFFS and it is intended to liaise on flexibility service co-ordination. However,

it is recognised that this work should be aligned with Open Networks work in this area. The system design is split into the functional areas for EFFS, which are:

Forecasting;

Capacity engine;

Service management;

Optimisation;

Scheduling;

Conflict avoidance and synergy identification;

Market interface; and

Reporting.





The functional areas were defined in the requirements phase of the EFFS project’s Workstream 1 and are represented conceptually in Figure 22 below.

Figure 22 – EFFS core functions

The areas of forecasting and conflict avoidance are most directly relevant to this Open Networks deliverable and so sections from the System Design Summary document (https://www.westernpower.co.uk/downloads/64093) are given below. Further details are given in the design documents for the functional areas of Forecasting (https://www.westernpower.co.uk/downloads/64090) and Conflict avoidance. (https://www.westernpower.co.uk/downloads/64102) Other data exchanges relevant to each functional area are outlined in the design documents relevant to each area, e.g. market interfaces.


https://www.westernpower.co.uk/downloads/64093






9.1 Forecasting

The project investigated several forecasting methodologies during Workstream 1 Smarter Grid Solutions (SGS) was tasked with assessing whether machine learning techniques could perform better than traditional statistical models and Capita Data Science team were tasked with providing independent validation of the proposed models and findings. This work covered a range of assets such as transformers at Grid Supply Points (GSPs), Bulk Supply Points (BSPs), Primary substations and large load customers, but also wind and solar generation sites. Forecasts for time horizons from day-ahead to six-months ahead were considered and the impact of including or excluding certain model features, such as different types of weather data, were tested. Three forecasting approaches were evaluated:

Extreme Gradient Boosting (XGBoost) – a machine learning approach;

Long Short-Term Memory (LSTM) – a machine learning approach; and

Auto-Regressive Integrated Moving Average (ARIMA) – a statistical approach. This work concluded that while all the techniques were capable of being tailored to provide reasonable forecasts, the machine learning technique XGBoost gave the best overall balance between accuracy of the results and the effort required to set up and maintain the forecasts. It also found that, as expected, input data quality was an important factor in the quality of the forecast. Surprisingly, while the inclusion of historic weather data improved the quality of the forecasts, they were able to perform reasonably well without this data. This was thought to reflect the model determining seasonal variations from the week-of-year and month-of-year features as a proxy for weather data. However, it was anticipated that day-ahead and week-ahead forecasts would benefit from inclusion of weather forecast data, especially when the predicted weather would be different from the seasonal averages. Similarly, while time-series forecasting methods can be used for forecasting wind or solar generation, they are not the recommended method. While a time-series model may be able to determine a general relationship between the weather data and the generation output, engineering models can better represent the non-linear impacts introduced by inverters, protective control systems etc. and so are the recommended method to model wind and solar generation. Figure 23 is an example of PV generation output data that can be used4.

Figure 23 – PV output data

4 https://www.renewables.ninja/


https://www.renewables.ninja/




During the Workstream 2 system design phase, the focus shifted from which model to implement to the practical questions around implementing the forecasting algorithm. The questions addressed and the conclusions are summarised in the following sections.

9.1.1 Data Sources

The forecasting tool will be provided with historic time-series data from existing SCADA monitoring. This is held within PowerOn5 and is routinely exported in the form of a HISTAN file. This file can be used to provide both the initial bulk data required for setting up the models and the ongoing requirements for continuing data so that the models are working with up-to-date inputs. Existing forecast weather data for solar and wind values will be used of to provide inputs for modelling. This is provided at BSP group granularity and therefore EFFS will need to be able to relate forecasting sites to the various weather forecast locations. The model requires additional forecast data for temperature and for the actual weather experienced. This data, along with a bulk set of historical recorded data for training the model, could be provided by Meteogroup6 at the same level of disaggregation as the current forecasts.

The design has identified the need for a process step for data correction of the time-series data from SCADA. The system will include repositories for both raw and cleansed data with processes to identify data issues being run against the raw data. Further work is planned for the build phase to determine how this can be automated and the degree to which data correction or substitution can be automated. Options for data substitution have been explored and will be examined further in the next phase. Similarly, the work to assess the potential impact of data quality has begun with a review of the monitoring points at a representative location in the trial area.

9.1.2 Accuracy monitoring

The accuracy metrics suggested by the SGS / Capita Data Science work have been adopted for continuing use within EFFS. There are two purposes to calculate accuracy metrics for the forecasts during the EFFS project. Firstly, it will increase the opportunities for learning what factors affect forecasting accuracy by providing a much larger set of results to analyse. The second reason is to see how forecasting accuracy changes over time. If the rate at which forecast accuracy deteriorates is similar across all sites, or all types of site then the process to retrain forecasts can rely on timed schedules. If, however the deterioration of forecasting accuracy varies widely between sites, then the process to identify sites that require retraining and schedule that retraining will necessarily be more complex.

5 https://www.gegridsolutions.com/products/brochures/uos/PowerOn_Control.pdf 6 https://www.meteogroup.com/sites/default/files/180807_weather_data_api_-_corp_factsheet_1.pdf


https://www.gegridsolutions.com/products/brochures/uos/PowerOn_Control.pdf

https://www.meteogroup.com/sites/default/files/180807_weather_data_api_-_corp_factsheet_1.pdf




With forecasts covering a wide range of time-horizons and being frequently refreshed there is the potential to calculate a large number of accuracy metrics, but to support their calculation forecasts need to be stored. For example, a forecast for three months in advance, which would be otherwise overwritten within a week, needs to be stored for three months until the actual data from the forecast date is available to calculate the accuracy metrics with. The requirements for sample sizes and retention periods have been specified for the project. Figure 24 contains an accuracy comparison for day ahead forecasts between forecast and actual HH values.

Figure 24 – Forecasting accuracy monitoring

9.1.3 Forecast locations and aggregation/ disaggregation

While the forecasting algorithm was tested at a variety of locations, the PSS®E7 power flow analysis software has the capability of aggregating power flows at one voltage level to determine the impact at a different voltage. Therefore, there is no requirement to forecast the load at a BSP transformer if the loads at all the downstream primary transformers and any customers directly connected to the 33kV networks can be forecast. Similarly, PSS®E can manage the aggregation to create a load profile for GSP transformers given the profiles of the relevant BSP transformers and any 132kV connected customers. The same process would apply to 66kV networks. While this reduces the total number of sites that require direct forecasts, it may still be useful to create a small number of forecasts at BSP or GSP transformers for validation purposes.

9.1.4 Forecasting adjustments

The time-series data that is used by the forecasting algorithms is expected to reflect network loadings for standard running arrangements. These are expected to occur for the majority of the time, and non-standard loading values due to maintenance or unplanned outages would be highlighted as outliers by the data cleansing process. However, the most onerous conditions for the network are more likely to be experienced when the network is abnormally arranged and therefore there is a need to adjust the forecasts accordingly. Forecasting the load for these non-standard arrangements using the same forecasting algorithms that are used for normal running is not practical due to the difficulties of identifying when the required running arrangements happened in the past to select the appropriate data, but also because the number of data points would be small, and in discontinuous blocks. Adjusting the load values that include the impact of embedded generation to reflect the total demand

7https://new.siemens.com/global/en/products/energy/services/transmission-distribution-smart-grid/consulting-

and-planning/pss-software/pss-e.html


https://new.siemens.com/global/en/products/energy/services/transmission-distribution-smart-grid/consulting-and-planning/pss-software/pss-e.html

https://new.siemens.com/global/en/products/energy/services/transmission-distribution-smart-grid/consulting-and-planning/pss-software/pss-e.html




when generation is disconnected requires modelling of the embedded generation downstream of a primary substation. As generation can be added to nodes within the PSS®E model, the approach of creating virtual generators, aggregating embedded generation of different types at the primary busbars has been adopted. A similar correction factor is required to reflect that the load on each transformer at a multi-transformer site feeding different busbar sections that are joined by a bus-section circuit breaker will be different according to whether the bus-section is open or closed. A method for estimating the proportional split has been devised based on the expected aggregated load of the outgoing feeders for each busbar section.

9.2 Conflict avoidance and synergy identification

The ESO has utilised flexibility for many years and involves sending short notice instructions to participants connected to the distribution system, requiring them to alter their demand or generation in return for payment. The options for customers to sell this type of service are growing with opportunities extending beyond the ESO to include DNOs and traders within energy markets. As the number of flexibility service users increases, so does the potential for conflicts arising between services. There is a very clear requirement to consider how the actions of different actors within system can operate without reducing efficiency, increasing costs or presenting unnecessary risks to the system resilience, both locally and as a whole. The functionality explored surrounding conflict avoidance reflects that EFFS is designed to operate in Open Networks future world B, where DNO and ESO are both involved in the co-ordination of flexibility services and exchange data to facilitate this. Synergies between services have also been considered in the context of how the system might identify them. This is purely to support information gathering to assist policy development. There are no activities in the scope of EFFS to reduce the services procured or scheduled on the basis that there may be a beneficial effect from a third-party service. Key design considerations are summarised below.

9.2.1 Conflict definition

There are many types of conflict between users of the network, but they have sufficient predictability or low impacts that they can feed into the forecasting but do not necessarily cause network issues. This could for example be two embedded generators shutting down to carry out annual maintenance at the same time. The impact would result in increased demand through the upstream network to supply more electricity from elsewhere, but it wouldn’t necessarily result in any real risk to the network. Our definition of conflicts between flexibility services are events that result in flexibility services being unavailable due to scheduling errors between multiple parties, services being counteracted by third party actions and combined actions that result in network issues for either the DNO or ESO. For example, we could have a situation where the ESO requires a Flexibility provider to start a generator to support national system balancing, but this would then trigger a nearby windfarm equipped Active Network Management (ANM) to reduce output by the same capacity and cancelling out the initial request.

9.2.2 Conflict identification

The different types of conflict require different data and approaches to identify them. For a scheduling conflict, a simple comparison can be made for the asset ID, the date and time of the service to be delivered along with the type of service being delivered (demand turn up that benefits both DNO and ESO would not necessarily constitute a conflict). Determining whether one service will negate or partially negate another will require some consideration of where the services are impacting the





network and the locations of the desired change. Where possible this will use network hierarchy information, however where a simplified process is not sufficient then power flow analysis will be used to determine whether one service is reducing the impact of another.

Conflict identification and quantification would be a beneficial activity whether or not conflict resolution activities were in place as it would provide much needed information to the industry. It is quite possible that where certain types of conflict present a low risk the most practical solution is not to require one party to alter their planned use of services, but rather to factor the risk of impact of that type of service into the safety margin applied within the capacity engine.

9.2.3 Conflict resolution

The potential principles for conflict resolution have been investigated. One potential option is to compare the marginal cost of using an alternative flexibility service for both parties. For the purposes of the EFFS trial it is sufficient to simply have values that can be compared by the resolution algorithm. The methodology to calculate marginal costs is likely to be best addressed within an industry wide forum such as Open Networks. Similarly, as the objective of the EFFS trials in relation to conflict avoidance are to prove that the data exchanges and processes are sufficient this can be achieved without engineering real conflicts in services and complete system-to-system interfaces but by using representative data.
