Page 1
MURDOCH UNIVERSITY
School of Engineering and Information Technology
Bachelor of Engineering Honours
2016
Effective Demand
Response Management in a
Low Voltage Grid
By: Mohammad Essarras
Thesis Supervisor: Dr. GM Shafiullah
Unit Coordinator: Prof. Parisa Bahri
Page 2
Declaration
This thesis is presented for the award of a Bachelors of Engineering Honours degree
in Electrical Power and Renewable Energy at Murdoch University. I declare that the
work presented in this report is of my effort and that all sources used in the
preparation of this report have been cited and acknowledged to the best of my ability.
Name:
Signature:
Date:
Page 3
Abstract
This project investigates the application of demand response on a low voltage grid
with a comparison between traditional and effective demand response strategies to
minimise the voltage violations in a network. For the project, a low voltage grid was
constructed using the DIgSILENT PowerFactory 15.2 power systems analysis
software. The low voltage grid built represents a common low voltage Australian
distribution grid. This was achieved by using standards, technical manuals, and
literature related to Australian distribution networks. Six different scenarios were
considered to examine the performance of demand response on the constructed grid.
Only two scenarios required the use of demand response; the grid under peak loading
condition and the grid under average loading condition with future PV penetration
values. For the peak loading condition, the voltage limits were breached with a
minimum bus voltage of 0.93 Vp.u. The traditional demand response strategy equally
reduced the 22 loads in the network from a scaling factor of 1.00 to 0.839. This
resulted in a total load reduction of 35.4 kW which equated to a 16.1% load
reduction. The effective demand response strategy reduced only 4 loads by a total of
23.8kW which equated to a 10.8% reduction of the total load.
For the average load with future PV generation values condition, voltage limits were
breached with a maximum bus voltage of 1.072 Vp.u. The traditional demand response
strategy reduced the PV generation uniformly from a scaling factor of 1.00 to 0.874.
This resulted in a total PV generation reduction of 27.72 kW which equated to 12.6%
and was distributed between 22 PV systems. The effective demand response strategy
reduced the PV production of 4 PV systems by a total reduction of 26.07kW which
equated to 11.85 %. The results showed that using an effective demand response
management with a direct load control strategy is a more efficient solution than using
a traditional demand response strategy.
Page 4
i
Table of Contents
Declaration.............................................................................................................................. i
Abstract .................................................................................................................................. ii
Acknowledgement ................................................................................................................. v
Acronyms and Abbreviations ............................................................................................... vi
Chapter 1- Introduction ..................................................................................................... 1
1.1 Overview ................................................................................................................. 1
1.2 Background ............................................................................................................. 1
1.3 Objective ................................................................................................................. 4
1.4 Project Outline ........................................................................................................ 4
1.5 Significance of Project ............................................................................................ 5
Chapter 2- Literature Review ............................................................................................ 6
2.1 Overview ................................................................................................................. 6
2.2 The current and future Australian utility grid design.............................................. 6
2.3 Demand response (DR) ........................................................................................... 7
2.4 LV Australian network ........................................................................................... 8
2.4 DIgSILENT PowerFactory ................................................................................... 11
2.5 Power flow simulation .......................................................................................... 12
2.6 Sensitivity analysis ............................................................................................... 12
2.7 The current state of the Australian Grid ............................................................... 13
2.8 The predicted future of the Australian grid........................................................... 14
Chapter 3- Methodology .................................................................................................. 15
3.1 Overview ............................................................................................................... 15
3.2 Project structure .................................................................................................... 15
3.3 Grid construction using DIgSILENT PowerFactory 15.2 .................................... 16
3.4 Scenario characterization ...................................................................................... 19
3.4.1 Scenario 1: Average load condition .............................................................. 19
3.4.2 Scenario 2: Peak load condition .................................................................... 20
3.4.3 Scenario 3: Average load condition with current PV penetration values ...... 20
3.4.4 Scenario 4: Average load condition with future PV penetration values........ 21
3.4.5 Scenario 5: Peak load condition with current PV penetration values............ 21
Page 5
ii
3.4.6 Scenario 6: Peak load condition with future PV penetration values ............. 22
3.5 Scenario simulation ............................................................................................... 22
3.6 Traditional DR ...................................................................................................... 22
3.7 Effective DR management .................................................................................... 23
3.8 Traditional and Effective DR flow charts ............................................................. 25
Chapter 4- Results and Analysis ...................................................................................... 26
4.1 Overview ............................................................................................................... 26
4.2 Results ................................................................................................................... 26
4.2.1 Scenario 1: Average load condition .............................................................. 26
4.2.2 Scenario 2: Peak load condition .................................................................... 27
4.2.3 Scenario 3: Average load condition with current PV penetration values ...... 31
4.2.4 Scenario 4: Average load condition with future PV penetration values........ 32
4.2.5 Scenario 5: Peak load condition with current PV penetration values............ 36
4.2.6 Scenario 6: Peak load condition with future PV penetration values ............. 37
4.3 Result Analysis ..................................................................................................... 38
Chapter 5- Conclusion ..................................................................................................... 41
5.1 Overview ............................................................................................................... 41
5.2 Conclusion ............................................................................................................ 41
5.3 Obstacles and difficulties ...................................................................................... 42
5.4 Future recommendations ....................................................................................... 43
References ........................................................................................................................... 44
Appendix A: Project grid parameters .............................................................................. 47
Appendix B: Simulation parameters and results ............................................................. 51
Scenario 1 ........................................................................................................................ 51
Scenario 2 ........................................................................................................................ 52
Scenario 3 ........................................................................................................................ 55
Scenario 4 ........................................................................................................................ 56
Scenario 5 ........................................................................................................................ 59
Scenario 6 ........................................................................................................................ 60
Appendix C .......................................................................................................................... 61
Newton-Raphson load flow equations ............................................................................. 63
Jacobian Matrix load flow sensitivity equations ............................................................. 64
Page 6
iii
List of Figures and Tables
Figure 1: Project method structure flow chart ..................................................................... 15
Figure 2: DIgSILENT PowerFactory diagram .................................................................... 18
Figure 3: Demand response flow chart ................................................................................ 25
Figure 4: Voltage level against load scaling factor plot for feeder A buses ........................ 29
Figure 5: Voltage level against generation scaling factor plot for feeder A buses .............. 34
Figure 6: Scenario 2 Voltage VS Scaling factor plot .......................................................... 53
Figure 7: Scenario 2 Voltage VS Scaling factor plot .......................................................... 57
Table 1: Average load condition simulation results 26
Table 2: Peak load condition simulation results .................................................................. 28
Table 3: Peak load condition simulation results after traditional DR application ............... 29
Table 4: Violated buses and voltage values......................................................................... 30
Table 5: Load flow sensitivity values .................................................................................. 30
Table 6: Load reduction values ........................................................................................... 30
Table 7: Peak load condition simulation results after effective DR application ................. 30
Table 8: Average load condition with current PV penetration simulation results............... 31
Table 9: Average load condition with future PV penetration simulation results ................ 33
Table 10: Average load condition with future PV generation simulation results after
traditional DR application ................................................................................................... 34
Table 11: Violated buses and voltage values....................................................................... 35
Page 7
iv
Table 12: Load flow sensitivity values ................................................................................ 35
Table 13: Required change in active power ........................................................................ 35
Table 14: Average load condition with future PV penetration value simulation results after
effective DR application ...................................................................................................... 35
Table 15: Peak load condition with future PV penetration simulation results .................... 36
Table 16: Peak load condition with future PV penetration simulation results .................... 37
Table 17: Project transformer parameters ........................................................................... 47
Table 18: Project transformer parameters ........................................................................... 47
Table 19: Project cable lengths ............................................................................................ 48
Table 20: Project bus parameters......................................................................................... 49
Table 21: Project loads and PV generators, locations and values ....................................... 50
Table 22: Scenario 1 parameters and simulation results ..................................................... 51
Table 23: Scenario 2 parameters and simulation results ..................................................... 52
Table 24: Scenario 2 Change in bus voltage due to change in scaling factor ..................... 53
Table 25: Scenario 2 load flow sensitivities ........................................................................ 54
Table 26: Scenario 3 parameters and simulation results ..................................................... 55
Table 27: Scenario 4 parameters and simulation results ..................................................... 56
Table 28: Scenario 4 Change in bus voltage due to change in scaling factor ..................... 57
Table 29: Scenario 4 load flow sensitivities ........................................................................ 58
Table 30: Scenario 5 parameters and simulation results ..................................................... 59
Table 31: Scenario 6 parameters and simulation results ..................................................... 60
Page 8
v
Acknowledgement
I would like to foremost acknowledge both my parents for their loving support and
patience throughout the course of my studies. Furthermore, I would like to
acknowledge the amazing academic staff at Murdoch University, all of which
contributed in helping me throughout the course of my study, especially my
supervisor Dr. GM Shafiullah who has supported and stood by me since the
beginning of my project and to my mentor and friend Mr. Md Moktadir Rahman
who was a source of inspiration and guidance throughout my thesis. I would like to
acknowledge my classmates who I shared ideas and knowledge with and Murdoch
University for the professionalism and dedication towards its students and staff. I
truly thank them all form my heart and wish them the very best.
Page 9
vi
Acronyms and Abbreviations
AC Alternating Current
AMI Advanced Metering Infrastructure
DER Distributed Energy Resource
DC Direct Current
DG Distributed Generation
DR Demand Response
HV High Voltage
kW Kilowatt
LV Low Voltage
MEPS Minimum Energy Performance Standard
MW Megawatt
MV Medium Voltage
P Active Power
P.f Power factor
PV Photovoltaic
Q Reactive power
S Apparent power
V Volts
Vp.u Voltage Per Unit
W Watt
Page 10
1
Chapter 1- Introduction
1.1 Overview
This chapter addresses background information, objectives, the outline and the
significance of this project. The background information will introduce concepts,
statistics, issues faced, and strategies currently practiced. The project objectives
will outline the main aims that the project intends to achieve. The significance of
the project will highlight the motives behind this project.
1.2 Background
Electricity power systems are an essential part of the infrastructure of any modern
society. Over the past century, electricity power systems have grown in size and
capacity to meet the increasing power demand of these growing societies (Hughes,
1983). With this growth, challenges arise and one of the most confronting
challenges is meeting residential power demands. Residential power demands are
challenging as they fluctuate with great contrast. In Australia, residential power
demands can fluctuate between 25% and 45% of the total energy demand in a
given network (AEMC, 2012). In addition, residential areas are powered by low
voltage (LV) networks with high currents (AS/ACIF, 2006). These high currents
contribute to high power losses that result in a drop in the voltage levels of these
networks. In response, power producers tend to overproduce electric power to
accommodate for the power fluctuations and to prevent voltage levels from
dropping below required limits. This approach is not only inefficient due to the
increase power losses associated with increased power flowing through
transmission lines, but as 63% of this energy is generated from coal, it has a high
Page 11
2
contribution towards Australia’s carbon emissions (Australian Energy Update,
2016).
A viable alternative to this is to distribute the load profile to reduce the contrast
between base load demand and peak load demand through schemes such as
demand response (DR). DR helps ease the balance between supply and demand by
involving the consumer to participate in reducing electricity power consumption at
peak load periods, and thus acting as a peak shaver to the load profile. The
traditional DR schemes being used today involve price based or incentive based
programs which only encourage the consumers, and thus could have a reliability
factor left unaddressed in the event of the consumers not being persuaded into
responding to the demand (U.S Department of Energy, 2006). In addition, these
programs targeted peak shaving of peak loads in a conventional electric system
built on the foundation of a unidirectional power flow. However, with the
introduction and recent increase in distributed generation (DG) in electricity grids,
the foundation on which these conventional electricity distribution systems were
built has started to change (Cappers, 2016).
This change has raised questions over the integration of DG into electricity
distribution systems, especially renewable distributed energy resources (DER),
which are growing rapidly due to economic, political and environmental factors
worldwide (Purchala, 2010). This concern is due to the intermittent nature of
renewable DER’s which not only change the power flow from unidirectional to
bidirectional, but do this at an unpredictable and rapid rate (ARENA, 2015). This is
a growing concern in Australia as it currently has the highest residential PV
penetration levels of 16.2% and is predicted to continue growing, with some
suburbs in Brisbane and Adelaide currently reaching 50% PV penetration
Page 12
3
(Australian Energy Council, 2016). The concern is that these residential DER’s
collectively can impact the electricity network grid significantly with no
administration or oversight from grid operators (Mauch, 2006). New schemes such
as the Advanced metering infrastructure (AMI) have been introduced to combat
this obstacle, it is now shifting the conventional grid and transforming it into a
modern “Smart Grid”, this has the ability to facilitate the required communication
between the grid operators and consumers (Wolfs, 2009). Similarly, this
advancement can be used in conjunction with the traditional DR and transform
traditional DR into an effective modern DR that can be managed to work in
conjunction with the modern smart grid. This will allow DR to achieve better
results when reducing the load in peak load periods, as it gives operators the ability
to target the most effective nodes in any given network. In addition, this opens a
new door for operators to monitor and manage residential PV power flow in today's
bi-directional power flow grid.
Page 13
4
1.3 Objective
The objective of this project is to investigate the difference between using a
traditional DR in comparison to an effectively managed DR on an Australian LV
residential network. In addition, this study will also investigate the use of an
effectively managed DR on residential PV to illustrate its viability and potential in
a modern bidirectional power flow grid.
1.4 Project Outline
- This project will begin by introducing the literature review and the studies
the project was based on.
- The methodology of the project will be explained in detail, including the
methodologies for grid construction, scenario characterization, and the
scenario simulation process.
- The results of the simulated scenarios will be presented with the discussion
of the outcomes in relation to the objectives of this project.
- The conclusion, obstacles faced and future recommendations will conclude
this report.
Page 14
5
1.5 Significance of Project
Utility grids are one of the largest and most complex networks known to mankind.
As demand for energy grows and the fundamental design on which utility grids
were built on changes, utility grids are under threat of having to be redesigned and
reconstructed. To upgrade or change the entire infrastructure of these large-scale
networks would mean a devastating blow to any economy. This would not only
affect utility providers and shareholders, but would also burden the consumers as
electricity prices would have to rise to compensate for the colossal changes. This
has pushed for a large investment into innovation for means to avoid such an event.
Demand response has been a cost effective and successful scheme for many
decades practiced by utility grids worldwide. However, as the centralized grid
changes towards becoming a more DER intensive grid, DR needs to adapt to the
changes and prove that it can remain to be a useful tool in utility grids. Therefore,
this project aims to improve the use of DR and provide a foresight to the potential
it holds in the near future by investigating current and future scenarios.
Page 15
6
Chapter 2- Literature Review
2.1 Overview
This chapter addresses the literature on which this project was built on. This
chapter will also address the Australian standards and technical guides for the
requirements and parameters of an LV Australian grid. The software manual for
DIgSILENT PowerFactory 15.2 will be examined for the clarification of the
simulation process.
2.2 The current and future Australian utility grid design
The current conventional Australian utility grid is made up of 3 main elements:
generation, transmission, and distribution. This type of grid has been dominant
over the past century. The fundamentals of this design consisted of a large
centralized generation unit that supplied the demand required by the consumers
connected to its grid. This design accommodates a reserve margin for the
continuously fluctuating power demand. The central generation units are required
to maintain a reserve margin (or reserve capacity) to readily meet any sudden rise
in power demand. The power flow in this type of grid is described as a
unidirectional power flow, as the power flows from production to consumption
(Lasster, 2003).
In the near future, this conventional design will start to shift due to two main
factors. The first is due to the depleting reserve capacity in transmission lines as the
utility grid and power demands grow (Lasster, 2003). The second is the increase in
DER’s which is causing the utility grid to shift from the conventional
unidirectional power flow to a bidirectional power flow grid which allows the
Page 16
7
consumption end of the network to produce power and inject it back into the larger
network body (Lo, 2013).
2.3 Demand response (DR)
DR is a strategy that reduces energy consumption at peak demand periods in a
network. This is done by encouraging consumers in reducing or shifting their
electricity demand during peak load periods in response to time-based rates or
financial incentives. The traditional DR strategy consisted of time-based rates that
offered consumers time-varying rates that vary with the value of electricity in
different time periods. This often encouraged the consumer to shift electricity
usage to less costly periods outside peak demand (U.S Department of Energy,
2006).
A more effective strategy is the incentives strategy which encourages consumers to
participate in programs that reduce their loads upon request by a program sponsor
often referred to as an "Aggregator." Participating consumers agree to respond with
load reduction upon request from the aggregator at high electricity demand periods
in return for a financial incentive. This strategy often involves a direct load control
approach in which the aggregator has direct control over certain consumer
equipment (U.S Department of Energy, 2006). This form of DR is referred to in
this project as an effectively managed DR strategy as it involves an aggregator or a
grid operator to manage the DR efficiently and thus has a more effective approach
to restoring voltage limits.
For the direct load strategy the current traditional grid needs to upgrade into a
smart grid. The smart grid has an integrated two-way communication system
between the power producer and the power consumer. These grids rely on the
Page 17
8
integration of an AMI (Advanced Metering Infrastructure) where smart meters are
installed to provide grid operators with better visibility into lower voltage networks
at the distribution end of the grid. This enables power producers to monitor power
consumption in real time which in turn allows better energy production and
delivery management (Hossain, 2013)
AMI is a system that allows two-way communication between consumers and
utilities. The network requires the installation of "Smart meter" at the user end. The
smart meter communicates by sending packages of information at short time
intervals to the utilities. This is a vital part of the smart grid and provides vision for
grid operators into the consumer ends of the network (U.S Department of Energy,
2015).
2.4 LV Australian network
Electricity grids around the world share the concept of distributing their power to
consumers by using LV grids. Interchangeable with LV network, an LV grid is a
grid that operates on voltages of up to 1000V AC. The LV grid is the last level in
the electricity grid structure which begins with the power generation at the power
plant. LV grids supply power to residential, commercial and industrial entities
(Siemens, 2016).
Low voltage networks distribute the power received from HV or MV transmission
lines by stepping down the voltage using transformers which consumers can then
use to run their equipment and loads. In Australia, any voltage up to 1000V AC is
considered a low voltage. However, most Australian low voltage grids deliver
power using 415V in a three-phase configuration or 240V single-phase.
Regulations and standards have been put in place to ensure the safety and stability
Page 18
9
of running these grids. These standards and regulations differ between utility
operators. In this project, the standards and regulations used in Western Australia,
regulated by Horizon Power and Western Power were used as a reference.
Horizon Power and Western Power own and operate the grids in the North-West
interconnected system (NWIS), the South-West interconnected system (SWIS) and
the non-interconnected system (NIS) and has placed regulations and standards for
these grids (Horizon Power, 2015). The LV networks within those grids consist of
overhead and underground networks. Horizon Power has specified that all future
LV networks will be built as underground LV networks unless within an area
where overhead cables are still in use. Standards have been put in place to regulate
each one of these LV network types. In recent years as PV systems have become
increasingly popular, standards and manuals have been revised to address the
changes and their effects on Australian LV networks. For this project, these
manuals and standards have been studied to gather the necessary parameters for
constructing an LV grid that resembles a common LV network in Australia. From
the Western Power technical rules (Western Power, 2016), Western Power
distribution connections manual (Western Power, 2015), Australian standard for
power transformers (AS 2374.1.2:2003, 2016), Horizon Power distribution design
manual (Horizon Power, 2014), Horizon Power technical rules (Horizon Power,
2013), and Australia and New Zealand standard for electrical installations
(AS/NZS 3008.1.1:2009, 2009), it was found that in Western Australia for a SWIS
LV network the operation conditions are as follows:
-The system supply is to function within 2.5% of the system nominated frequency
of 50 Hz.
Page 19
10
-The permitted voltage levels at any given node in the network must be within the
limits of 6% of the standard supply voltage.
-Distribution systems must be designed to supply the maximum load for the area
served while considering changes in the foreseeable future.
-Liquid cooled distribution transformers of a 200kA size require a minimum of
98.73% efficiency rating.
-Distribution transformers may operate to up to 1.4 of their maximum rating.
-Consumer loads are measured in their active power values.
-Power factor requirement for consumer loading in Western Australia is no less
than 0.8. However, it is a requirement by most distributors that the P.f remains
between 0.9 and 0.95.
-A load power factor correction device is to be installed with loads below 0.8 P.f.
-Most PV system outputs have a power factor output of 1.
-Cables supplying an LV network will have a three-phase configuration of ABC-N.
-Cable loadings allow a tolerance of 10% minimum.
-In an LV distribution network all residential, commercial and industrial
installations supplied are to be within 500m from the distribution transformer.
-Cable de-rating factor of overhead cables with more than 600mm spacing is 1.
Page 20
11
2.4 DIgSILENT PowerFactory
DIgSILENT PowerFactory is a power systems analysis software used to simulate
networks for the applications of generation, transmission and distribution of power.
It is the preferred program in grid operations by Western Power. The DIgSILENT
PowerFactory manual was a vital tool in understanding the calculation methods
used by the program for analysis purposes. The manual was used as a reference in
carrying out the grid construction and scenario simulations. It showed the technical
background including the equations used. DIgSILENT PowerFactory uses the
Newton-Raphson power equations for load flow calculation and the Jacobian
Matrix for load flow sensitivity calculation. From the user manual, the following
was extracted:
-Loads under normal conditions were modeled as general loads where the P and Q
are constant without voltage dependency.
-The load scaling factor can be adjusted manually at each load or collectively by
using the load scaling factor entry when calculating load flow.
-The calculation method to be used in a balanced grid is a positive sequence AC
load flow.
-The generation scaling factor can be adjusted manually at each PV generator or
collectively by using the generation scaling factor entry when calculating load
flow.
-The active and reactive powers of loads and generators will be scaled equally.
-Load flow analysis can be calculated for diagonal elements or to a single bus bar.
Page 21
12
2.5 Power flow simulation
Power grids are complicated networks with expensive equipment connected to
serve the purpose of delivering high amounts of electric power. It would be costly
and dangerous to make alteration to a grid based on a trial and error approach.
Therefore power flow simulation tools are used to predict the performance of a
power system in the real world. This is performed by using numerical and complex
calculation methods. The Newton-Raphson algorithm (see Appendix C) is a
common numerical technique used for solving non-linear equations in power flow
simulations. This method uses sequential linearization and iterations of multi-
variables to reach solutions.
2.6 Sensitivity analysis
The power flow sensitivity analysis determines the effect of an independent
variable on a dependent variable under certain conditions. The sensitivity analysis
of a model investigates the potential changes and the results of these changes to
optimize the outputs of the model. In any system, the optimal version of that
system is always preferred as it means the best potential use of that system is in
effect (Pannell, 1997). Electricity distribution grids are constantly striving to
achieve maximum efficiency in its operation; therefore, a sensitivity analysis is
vital to the optimization of a distribution grid as a large portion of the losses occur
in the transmission and distribution. Sensitivity analysis of a grid contributes
positively towards the overall performance of the grid. It does so by taking full
advantage of the maximum available resources and selecting the most viable
technical parameters. The sensitive analysis of a system also helps illustrate the
constraints and limits of the system components. The Jacobian matrix (see
Page 22
13
Appendix C) is a mathematical approach used to calculate the sensitivities of the
bus voltages in power distribution systems. It uses iterations to the change in active
and reactive power to calculate the change in voltage for a specific bus. This
enables an accurate prediction of bus voltage performance of the voltage of a bus
relative to changes in power demand.
2.7 The current state of the Australian Grid
The current state of the Australian grid is important as it currently delivers power
to more than 20 million people across Australia. Australia currently produces 3
times the amount of energy that it currently consumes (Office of the Chief
Economist, 2016). The surplus reserve capacity found in the SWIS network
between the years 2016 and 2017 was 23% (Department of Finance Public Utilities
Office, 2016). This is a troubling statistic as the yearly energy consumption rise in
Australia does not exceed 2%. Only 14% of the electricity generated in Australia
comes from renewable energy resources in comparison to almost 63% coming from
coal. Australia is a solar rich country with almost 65% yearly sunny days. Also,
with the dropping prices in PV system installations, many Australian consumers
are installing PV systems with an average size being 5 kW. This has made the
average residential PV penetration levels in states such as Queensland reach 30%
with some networks reaching as high as 50%.
The current peak to average demand ratios present in Australian LV networks
range between 1.61 and 2.7 (Utility Magazine, 2015). Supplying electricity with a
large difference in demand ratios requires ramping down or ramping up electricity
generator outputs. This can be difficult with coal electricity generators as they are
not quick in changing their generation speeds. This could be concerning as it leads
to power quality issues. Also with the expansion of the grid and increase in loading
Page 23
14
the reserve capacity is depleting and thus this strategy of over producing electricity
cannot last. Other more effective strategies are being implemented or developed.
The demand response strategy of time-based rates has been used for decades and
has proven to be effective. Other more effective strategies are in their development
stages. The use of aggregators and direct load control are slowly expanding with
the first step of installing smart meters being implemented. All new connections are
fitted with the smart meters. The overall current state of Australian LV grids is
stable with power outages only occurring in sever conditions.
2.8 The predicted future of the Australian grid
For the future of the Australian grid renewable DER’s are being studied closely as
a viable alternative to the extensive coal energy production option currently being
used. The expected forecast for coal energy consumption is a 2% drop in the next
10 years. Renewables on the other hand are expected to increase by 2.1% (Office of
the Chief Economist, 2016). The annual growth of all sectors in energy
consumption is expected to rise by only 1% (Office of the Chief Economist, 2016).
The cost of electricity is expected to continue increasing with distribution network
costs being the major contributor. The peak demand to average demand ratio is also
expected to grow and plays a major role in increasing electricity prices. Residential
PV generation systems are expected to grow with a steeper incline as PV prices
continue to drop. Studies are currently being conducted on battery installations and
micro grid development especially in rural areas.
Page 24
15
Chapter 3- Methodology
3.1 Overview
This chapter explains the methodology of the project. The structure of the project is
addressed and defines the steps taken in accomplishing this project. The grid and
the construction process are then illustrated. The methods for scenario building and
execution are defined. This is followed by an explanation of the process of
applying demand response and effective demand response management.
3.2 Project structure
The project was built by initially investigating methods addressed in the literature
review. A model was then constructed and scenarios were designed to represent the
current and future states of LV networks in Australia. Simulations were run and
adjustments were made. Results were then recorded and used for the discussion and
recommendations for future projects.
Figure 1: Project method structure flow chart
Literature review Scenario
characterisation
Project topic
Model
construction
Scenario
simulation
Result analysis Result discussion Conclusion
Recommendation
Page 25
16
3.3 Grid construction using DIgSILENT PowerFactory 15.2
The structure of the LV network grid (see figure 3) was constructed based on the
guidelines specifications and standards of the following:
1- Western Power distribution connections manual (WADCM) 2015.
2- Western Power technical rules 2016.
3- Australian standard for power transformers (AS 2374.1.2:2003) Part1.2: MEPS
requirements for distribution transformers.
4- Australia and New Zealand standard for electrical installations (AS/NZS
3008.1.1:2009) Part 1.1: Cables for AC voltages up to and including 0.6/1 kV-
typical Australian installation conditions.
5- Aerial – Nexans Olex cable catalogue.
Using DIgSILENT PowerFactory a new project was selected on which to build the
LV network. The nominal frequency of supply was selected as 50 Hz. The power
supplied to the LV grid needed to come from the greater HV or MV network of the
grid and therefore the power supply was modeled as an external grid. The present
standard voltage for an MV distribution network used in the SWIS is 22kV and this
was used as the MV end bus voltage. The external grid was connected to the MV
bus. A single 200kVA transformer was selected to step down the voltage from
22kV to 415V for the LV network. The transformer MV side was connected to the
22kV MV bus and the LV side was connected to a 415V main feeder bus. The
main feeder bus then supplied power to 4 feeders labeled A, B, C and D specifying
their designated supply areas. Feeders A, B, C and D delivered power to 7, 6, 7 and
2 LV buses respectively. Some nodes were connected to a single bus, for example
LV bus A5 was connected to a single node, others nodes were connected to 2
buses, for example LV bus A6 and A7 share the same node (see figure 2). A single
Page 26
17
load and a single PV unit were connected to each LV bus. Each load was modeled
as a general 3 phase load and represented three single phase consumers. The PV
unit was modeled as a static generator. The parameters of the transformer, buses,
lines and loads were then entered for each respective element (see Appendix A).
The values of the grid components such as feeder lengths, bus loading and bus PV
generation values were kept uniform throughout the design. This ensured that the
outcomes of the application of DR were clear and that no other elements were
involved in influencing the results. Also, no other voltage manipulating means
were present as it was assumed that these means had been depleted before resorting
to DR.
The project grid built using DIgSILENT PowerFactory is presented in figure 2. The
parameters are presented in Appendix A.
Page 27
18
Figure 2: DIgSILENT PowerFactory diagram
Page 28
19
3.4 Scenario characterization
The scenario characterization objective was to represent present and future
conditions that could be faced by a typical Australian LV grid. These scenarios
were built on the literature investigated while complying with the standards.
The scenarios considered were:
1- Average load condition
2- Peak load condition
3- Average load condition with current PV penetration values
4- Average load condition with future PV penetration values
5- Peak load condition with current PV penetration values
6- Peak load condition with future PV penetration values
3.4.1 Scenario 1: Average load condition
In this scenario, the grid was to operate under normal conditions with an average
demand for power by the loads. The purpose of this scenario was to examine the
grid performance and the parameters chosen in the grid construction. The average
loading assumed was 40% of peak demand. This value of loading was chosen as it
was extracted from the literature that the average peak to average demand ratio was
2.5 and thus the average loading was assumed as 40% of the peak demand. Also,
the transformer loading in response at this consumer loading value was close to
50%. This value also allowed some tolerance for change if the peak loading in later
conditions appeared to exceed the grid component limitations. To apply the 40%
loading the loading scale of all the buses on the grid was adjusted to 0.4 of the
maximum demand. The simulation scenario was saved and the component
Page 29
20
performance values were recorded for ease of adjustments if necessary as the
scenario simulations progressed.
3.4.2 Scenario 2: Peak load condition
For the peak load scenario, the grid was required to respond to a peak demand in
power by the consumers connected to the grid. The purpose of this scenario was to
test the grid performance under maximum loading conditions. From the literature,
it was found that the loads can exceed the transformer rating under peak load
conditions. This scenario was built to overload the transformer and examine the
effects of applying DR to reduce the overload in the transformer. Due to the peak
demand to average demand ratio assumed, the peak load was simulated as 100% of
the available bus loadings. This was applied to the grid by adjusting the load scale
factor of the loads to 1. The collective power demand of the loads was checked to
be 220 kW. The simulation scenario was saved and was then run and the voltage
per-unit values were recorded.
3.4.3 Scenario 3: Average load condition with current PV penetration values
From this scenario, the requirement was to represent the current state of the
Australian LV network. A PV generation value was added to resemble the effects
of residential PV systems on the network as this was the case in the real world. The
PV penetration values were chosen to be 30%. This was assumed to be an
appropriate penetration value as values ranged between 0 and 50% in extreme
cases. This scenario presents an insight to current performance of LV networks in
Australia under the real-world conditions where PV penetration levels are taken
into consideration. This scenario was simulated by adjusting the load scaling factor
Page 30
21
to 0.4 to represent the 40% average loading and the PV generation values were
adjusted to 0.3 to express the 30% PV penetration value.
3.4.4 Scenario 4: Average load condition with future PV penetration values
The objective of this scenario was to examine the effects of future PV penetration
values. As the price of PV systems drops and the price of electricity increases it
was assumed that more PV systems will be installed in the near future (Office of
the Chief Economist, 2016). The value chosen for examination was 100% PV
penetration value. This value was chosen as it represented a grid in the future
where all the consumers have a PV system installed. Also, it was assumed that the
outputs of the PV systems were equal to the maximum peak demand. This scenario
would illustrate the effects of PV systems at low power consumption times if the
PV systems were designed to meet the consumers highest load needs. This scenario
was implemented in the simulation by increasing the PV generation scale factor to
1 to equate to the 100% PV penetration value and the load was adjusted to 0.4 to
represent the average loading of 40% on the grid.
3.4.5 Scenario 5: Peak load condition with current PV penetration values
From this scenario, the requirement was to examine an Australian LV network
under full loading with the consideration of the current real world PV penetration
values. This scenario illustrates the effects of current PV generation on an
overloaded grid. This would show if the current PV generation values were
beneficial to the grid. This scenario was simulated by adjusting the load scaling
factor to 1 to represent the full loading of consumer buses and the PV generation
values were adjusted to 0.3 to express the current 30% PV penetration value.
Page 31
22
3.4.6 Scenario 6: Peak load condition with future PV penetration values
For this scenario, the goal was to investigate the effects of future PV penetration
values on the current grid in peak load periods. This will allow the grid and its
components to be investigated under the maximum range of loading and
generation. The scenario was simulated by increasing both the load and generation
factors to 1. In addition to the bus voltages, the line and transformer loadings were
observed in this scenario to record any load limitation breach.
3.5 Scenario simulation
For the scenario simulation stage, the scenarios characterized for this project were
run individually on the model constructed. The scenario parameters were entered
by accessing the load and PV generation scaling factors. The simulation was then
run using the load flow calculation method. The method was selected as a Newton-
Raphson load flow classic equation. The calculation method was chosen as
balanced, positive sequence, AC load flow. The tap adjustments were removed as
the transformer was assumed to have already exhausted the available tapings. All
the other load flow calculation parameters in DIgSILENT PowerFactory were kept
standard as they are not required to be altered. The voltage per-unit values of each
bus were recorded. For the scenarios that encountered bus voltage limit violations,
the buses violating the voltage limits were specified for the implementation of
traditional DR and effective DR management.
3.6 Traditional DR
The traditional DR approach involved applying demand response in a uniform
fashion (see figure 3). This represented a time-based rate being applied to all
consumers. After the voltage violating buses were specified from the scenario
Page 32
23
simulations the loads were reduced with iterations to the load scaling factor. The
load flow calculation process was repeated for each iteration. The values of the
power reduced and the voltage limits were recorded for each scaling factor. The
values of loads reduced were plotted against the voltage levels of the bus with the
largest voltage violation. A linear trend line was inserted and the equation of the
trend line identified the ratio of power reduction to bus voltage reduction. The
equation was then used to reduce the loads by a scaling factor that corrected the
voltage of the most violated bus. The values of reduced power, number of
consumers involved and the bus voltages of the network were recorded.
3.7 Effective DR management
The effective DR management approach involved applying demand response to
specific buses (see figure 3). These buses were chosen based on their effectiveness
in the rectification of bus voltage limit violations in the network. This represented
an incentive based, direct load control DR being applied to the most effective
consumers. After the voltage violating buses were specified from the scenario
simulations a load flow sensitivity analysis was performed for each of the voltage
violating buses. The sensitivity of each bus was recorded. The voltage violation
value of each bus was calculated by finding the absolute value of the difference
between the current voltage limit and the acceptable voltage limit. This value was
then multiplied by the sensitivity of that bus to attain the required change in load or
change in PV generation. As each bus had a varying effect on other bus voltages in
the network, the process of effective DR management was conducted by changing
the load or PV generation of the most effective buses. If there remained a voltage
violation in the network the second most effective bus was considered and the
process was repeated until all the buses in the network were within the desired
Page 33
24
voltage limits. The values of reduced power or PV generation, number of
consumers involved and the bus voltages of the network were recorded.
Page 34
25
3.8 Traditional and Effective DR flow charts
Figure 3: Demand response flow chart
Yes
No No
Yes
Start
Grid
construction
Characterize
scenario
Run load flow simulation of
scenario
Bus voltage
limits violated?
Implement traditional DR
by uniform load reduction
Start
Grid
construction
Characterize
scenario
Run load flow simulation of
scenario
Bus voltage
limits violated?
Run load flow sensitivity
Implement effective DR by
reducing loads or PV generation of
the most violated buses
End
End
Bus voltage
limits violated? Yes
No
Traditional DR Effective DR
Page 35
26
Chapter 4- Results and Analysis
4.1 Overview
This chapter will present summarized simulation results from each scenario with
the interpretation and discussion. (for detailed results refer to Appendix B)
4.2 Results
4.2.1 Scenario 1: Average load condition
For the average loading condition, the assumption was made that the loadings were
at a scaling factor of 0.4. This represented the average loading of 40% on a low
voltage grid. The power factor of the loads was assumed to be 0.95 as some utility
providers require a P.f of at least 0.95. From the results, it was found that the
minimum and the maximum bus voltages were within the required limits and thus
the application of DR was unnecessary.
Table 1: Average load condition simulation results
Loads active 22
Load kW 10
Load scaling factor 0.4
Load P.f 0.95
PV active 0
Generation kW 0
Gen. scaling factor 0
Gen. P.f 0
Minimum bus voltage (LV Bus A6, A7) 0.972
Maximum bus voltage (LV Bus D1) 0.985
Page 36
27
4.2.2 Scenario 2: Peak load condition
The peak load condition was assumed to be 2.5 times the average load as this was a
common average to peak load ratio. This meant a loading scale factor of 1 meaning
the loads at each bus were at their maximum of 10kW. The minimum bus voltage
measured at LV bus A6 and A7 was 0.93 Vp.u.
A traditional and effective DR was applied to the condition. For the traditional DR,
the values of voltage and load scaling factor were plotted against each other and a
trend line was inserted. A required load scaling factor of 0.839 was calculated
which meant the load had to be reduced by 16.1% from the current value. A
uniform load reduction was applied reducing all loads from 10 kW to 8.39 kW. The
simulation was run again and the minimum bus voltage was 0.94 at LV bus A6 and
A7. The same scenario was repeated but this time with the application of effective
DR. The load flow sensitivities for the buses with the most violated voltage levels
were targeted first. The buses of LV bus A6, A7, B6 and C7 had the most influence
on the grid. This meant that any change in these buses would have a more
significant effect on the grid than the same change being applied elsewhere. The
required load reductions calculated from the load flow sensitivities were found to
be 10.323, 10.323, 5.563 and 7.913 kW for buses A6, A7, B6 and C7 respectively.
Since LV bus A6 and LV bus A7 were on the same node, half the load reduction
was applied to each. The results after the effective DR application show that the
bus voltages were now within acceptable limits but the % reduction of each bus
was up to 79%. This would be impractical in a real world situation. However, less
overall load reduction was applied than the traditional DR with a better result of
0.943 for the minimum voltage. The total reduction in percentage was 10.8% for
Page 37
28
the effective DR and 16.1% for the traditional DR. The effective DR also involved
fewer consumers (3 per bus) of only 12 compared to 66.
The results show that effective DR management reduces the load reduction
quantity and the number of consumers involved. However, from the effective DR
management it was found that 10.363 kW were required to be reduced from a
single bus (LV bus A6), this bus was connected to another (LV bus A7) and thus
half the reduction was required. Also 7.913 kW was required to be reduced from
LV bus C7 which was almost 80% of the load. From this result, we can see that
with the effective DR management, not all the load can be reduced from one single
bus and thus more consumers need to be involved to split the load reduction
quantity. Also, with the application of effective DR the targeted Vp.u was 0.94 and
yet the simulated Vp.u was 0.943, this was above the required amount and thus it is
evident that the effect of load reduction was amplified as more than one bus was
reduced and each bus had a secondary effect on all other bus voltages.
Table 2: Peak load condition simulation results
Loads active 22
Load kW 10
Load scaling factor 1
Load P.f 0.95
PV active 0
Generation kW 0
Gen. scaling factor 0
Gen. P.f 0
Minimum bus voltage (LV Bus A6, A7) 0.930
Maximum bus voltage (LV Bus D1) 0.964
Page 38
29
Applying traditional DR:
Lowest voltage bus: LV bus A7
Figure 4: Voltage level against load scaling factor plot for feeder A buses
For LV bus A7:
| |
Target Vp.u for bus = 0.94; load scaling factor = 0.839
After the generation scaling factor of 0.839 was applied, the results were as
follows:
Table 3: Peak load condition simulation results after traditional DR application
Consumers involved 22
Total load reduction kW 35.4
Total load reduction % 16.1%
Minimum bus voltage (LV Bus A6, A7) 0.941
Maximum bus voltage (LV Bus D1) 0.970
0.915
0.92
0.925
0.93
0.935
0.94
0.945
0.95
0.955
0.96
1 0.99 0.98 0.97 0.96 0.95
Vo
lta
ge
p.u
Load scaling factor
LV bus (A1)
LV bus (A2)
LV bus (A3)
LV bus (A4)
LV bus (A5)
LV bus (A6)
LV bus (A7)
Page 39
30
Applying effective DR management:
Table 4: Violated buses and voltage values
Location V p.u
LV bus (A5) 0.934551
LV bus (A6) 0.930016
LV bus (A7) 0.930016
LV bus (C5) 0.934551
LV bus (C6) 0.934551
LV bus (C7) 0.932283
LV bus (B4) 0.936819
LV bus (B5) 0.936819
LV bus (B6) 0.934551
Load flow sensitivities of the most effective buses:
Table 5: Load flow sensitivity values
Location
dv/dP
(Vp.u/MW)
dv/dQ
(Vp.u/MVar)
LV bus (A6) 1.034035596 0.965557
LV bus (A7) 1.034035596 0.965557
LV bus (B6) 1.02092259 0.958599
LV bus (C7) 1.02539365 0.961473
| |
Load reduced in kW:
Table 6: Load reduction values
Location ΔP kW Reduction kW % Reduction
LV bus (A7) 10.323 5.162 51.62 %
LV bus (A6) 10.323 5.162 51.62 %
LV bus (B6) 5.563 5.563 55.63 %
LV bus (C7) 7.913 7.913 79.13 %
(As LV bus A6 and LV bus A7 are on the same node half the load reduction was applied to each)
Table 7: Peak load condition simulation results after effective DR application
Consumers involved 4
Total load reduction kW 23.8
Total load reduction % 10.8%
Minimum bus voltage (LV Bus A6, A7) 0.943
Maximum bus voltage (LV Bus D1) 0.973
Page 40
31
4.2.3 Scenario 3: Average load condition with current PV penetration
values
For the average loading condition with current PV penetration values scenario, the
scenario loading condition was the same as scenario 1 but with the difference being
that PV generation was included. The PV generation scaling factor was raised to
0.3. This represented a grid PV penetration of 30%. The power factor of the PV
generation units was assumed to be 0.8 for ease of simulation purposes only, in the
real world this value is 1 or close to 1. From the results, it was observed that the
minimum and the maximum bus voltages were within the required limits and thus
the application of DR was unnecessary. Also, unlike scenario 1, the minimum and
maximum voltage limits were now raised to 1.002 and 1.006 respectively. This
indicates that with the current PV penetration levels in Australian LV networks, the
grid is being assisted in its power delivery.
Table 8: Average load condition with current PV penetration simulation results
Loads active 22
Load kW 10
Load scaling factor 0.4
Load P.f 0.95
PV inputs active 22
Generation kW 10
Gen. scaling factor 0.3
Gen. P.f 0.8
Minimum bus voltage (LV Bus A6, A7) 1.002
Maximum bus voltage (LV Bus D1) 1.006
Page 41
32
4.2.4 Scenario 4: Average load condition with future PV penetration
values
In this scenario, the load scaling factor was 0.4 and the PV penetration was raised
to 100%. This meant a generation factor of 1, which is an assumed value for future
PV penetration values. In this scenario, an overvoltage was present in the end buses
of feeders A, B and C. From this it was concluded that future PV penetration values
may occur complications to the grid if left unmanaged. The maximum bus voltage
measured at LV bus A6 and A7 was 1.072 Vp.u. The application of demand
response was necessary to reduce these voltages to the acceptable limit of 1.06 Vp.u.
A traditional and effective DR was then applied. For the traditional DR, the values
of voltage and load scaling factor were plotted against each other and a trend line
was inserted. A required generation scaling factor of 0.874 was calculated which
meant the PV generation output needed a reduction of 12.6% from the current
value. A uniform PV generation reduction was applied reducing all PV outputs
from 10 kW to 8.74 kW with a power factor of 0.8. The simulation was run again
with the changes to PV generation and the maximum bus voltage was now 1.060 at
LV bus A6 and A7. The same scenario was then repeated but with the application
of effective DR. The load flow sensitivities for the buses with the most violated
voltage levels were targeted first. The buses of LV bus A6, A7, B6 and C7 had the
most influence on the grid. The required PV generation reduction values calculated
from the load flow sensitivities were found to be 10.295, 10.295, 7.074 and 8.699
kW for buses A6, A7, B6 and C7 respectively. As LV bus A6 and LV bus A7 were
on the same node, half the generation reduction was applied to each bus. The
results after the application of effective DR show that the bus voltages were now
within acceptable limits. Also, less generation reduction was applied to the PV
Page 42
33
generators than the traditional DR with a better result of 1.052 for the minimum
voltage. The total reduction in percentage was 11.85% for the effective DR and
12.6% for the traditional DR. The effective DR also involved fewer consumers (3
per bus) of only 12 compared to 66.
The results show that effective DR management reduces the necessary PV
generation reduction quantity and the number of consumers involved. However,
from the effective DR management it was found that the highest reduction required
from the buses was around 8.7 kW. This would be almost the entire PV generation
quantity. As consumers in Australia mostly receive a feed in tariff, this may reduce
the gains they receive from the installation of PV generation and in turn the
payback period of the PV system installed. This will need to be addressed and
agreed upon by the consumer and the aggregator managing the PV generation
system. Also with the application of effective DR the targeted Vp.u was 1.06 and
yet the simulated Vp.u was 1.052, this was considerably below the required amount
and thus it was evident that the effect of generation reduction was amplified as
more than one bus was reduced with each bus influencing all other bus voltages.
Table 9: Average load condition with future PV penetration simulation results
Loads active 22
Load kW 10
Load scaling factor 0.4
Load P.f 0.95
PV inputs active 22
Generation kW 10
Gen. scaling factor 1
Gen. P.f 0.8
Minimum bus voltage (LV Bus A6, A7) 1.041
Maximum bus voltage (LV Bus D1) 1.074
Page 43
34
Applying Traditional DR:
Highest voltage bus: LV bus A7
Figure 5: Voltage level against generation scaling factor plot for feeder A buses
For LV bus A7:
| |
Target Vp.u for bus = 1.06; PV generation scaling factor = 0.874
After the generation scaling factor of 0.874 was applied, the results were as
follows:
Table 10: Average load condition with future PV generation simulation results after traditional DR
application
Consumers involved 22
Total Gen. reduction kW 27.72
Total Gen. reduction % 12.6%
Minimum bus voltage (LV Bus A6, A7) 1.033
Maximum bus voltage (LV Bus D1) 1.060
1.03
1.035
1.04
1.045
1.05
1.055
1.06
1.065
1.07
1.075
1.08
1 0.99 0.98 0.97 0.96 0.95
Vo
lta
ge
p.u
Generation scaling factor
LV bus (A1)
LV bus (A2)
LV bus (A3)
LV bus (A4)
LV bus (A5)
LV bus (A6)
LV bus (A7)
Page 44
35
Applying effective DR management:
Table 11: Violated buses and voltage values
Location V p.u
LV bus (A6) 1.068182
LV bus (A7) 1.068182
LV bus (C7) 1.066254
LV bus (C5) 1.064327
LV bus (C6) 1.064327
LV bus (A5) 1.064327
LV bus (B6) 1.064326
LV bus (B4) 1.062399
LV bus (B5) 1.062399
Load flow sensitivities of the most effective buses:
Table 12: Load flow sensitivity values
Location
dv/dP
(Vp.u/MW)
dv/dQ
(Vp.u/MVar)
LV bus (A6) 0.808264959 0.789635871
LV bus (A7) 0.808264959 0.789635871
LV bus (B6) 0.814306683 0.79576899
LV bus (C7) 0.812111712 0.79358049
| |
Generation reduced in kW:
Table 13: Required change in active power
Location ΔP kW
LV bus (A6) 10.295
LV bus (A7) 10.295
LV bus (B6) 7.0742
LV bus (C7) 8.6994
(As LV bus A6 and A7 are on the same node half the generation reduction was applied to each)
Table 14: Average load condition with future PV penetration value simulation results after effective DR
application
Consumers involved 4
Total Gen. reduction kW 26.07
Total Gen. reduction % 11.85%
Minimum bus voltage (LV Bus A6, A7) 1.034
Maximum bus voltage (LV Bus D1) 1.052
Page 45
36
4.2.5 Scenario 5: Peak load condition with current PV penetration values
For the peak load with current PV penetration values scenario, the scenario loading
condition was the same as scenario 2 but the difference being that the current PV
generation value of 30% was considered. The PV generation scaling factor was
selected to be 0.3. From the results, it was found that the minimum and the
maximum bus voltages were within the required limits and thus the application of
DR was unnecessary. Also, it was found that the PV penetration values assisted in
maintaining the acceptable voltage limits. This indicates that with the current PV
penetration levels in Australian LV networks, PV generation is beneficial to the
grid when coinciding with peak loading conditions.
Table 15: Peak load condition with future PV penetration simulation results
Loads active 22
Load kW 10
Load scaling factor 1
Load P.f 0.95
PV inputs active 22
Generation kW 10
Gen. scaling factor 0.3
Gen. P.f 0.8
Minimum bus voltage (LV Bus A6, A7) 0.958
Maximum bus voltage (LV Bus D1) 0.980
Page 46
37
4.2.6 Scenario 6: Peak load condition with future PV penetration values
For this scenario, the loading and the PV generation was assumed to be at their
maximum. The scaling factors for both were increased to 1 representing 100%
loading and generation. From the results, it was found that the minimum and the
maximum bus voltages were within the required limits and thus the application of
DR was unnecessary. Also, future PV penetration values can assist in maintaining
the acceptable voltage limits under peak loading conditions.
Table 16: Peak load condition with future PV penetration simulation results
Loads active 22
Load kW 10
Load scaling factor 1
Load P.f 0.95
PV inputs active 22
Generation kW 10
Gen. scaling factor 1
Gen. P.f 0.8
Minimum bus voltage (LV Bus A6, A7) 1.019
Maximum bus voltage (LV Bus D1) 1.039
Page 47
38
4.3 Result Analysis
From the results of the scenarios, the following observations were made:
-From scenario 2 and scenario 4, it was found that the effective DR involved less
consumers and required less power reduction than traditional DR. This could be
beneficial to aggregators that are required to pay an incentive to consumers as they
can now pay incentives to a less number of people. This also allows aggregators to
install less direct load control devices, or if this was to become the norm for all
consumers in the future, it would inform aggregators which consumers to start with
in the installation process.
-From the scenarios above, only a change in active power was necessary to achieve
the required voltage level change when applying effective DR. Also, the reactive
sensitivities would not be used as it would not be conventional in the real world. It
would require a change in the power factor or the use of a power factor correction
tool which is not as practical as simply changing the active power of the loads by
turning them off.
-From scenario 2 and scenario 4, it was found that the effective DR management
was applied the reduction in PV generation yielded a larger change in the voltage
level than that of reducing the loads. This could be due to the lower power factor in
the PV generation units. The change in the apparent power by reducing the PV
generation’s active power is larger than the change in apparent power of reducing
the loads.
Page 48
39
-From the simulations, it was observed that the end buses were the most sensitive
to changes and in turn the most effective for the application of DR. This is evident
as the calculation of the Jacobian matrix process attracts a greater value change as
the bus increases in series.
-From scenarios 4 and 6, the predicted future PV penetration levels in Australian
LV networks were found to assist the grid under peak loading conditions but the
PV generation output needed management as they raise voltages at end buses
beyond the acceptable limits.
-From the results in scenarios 3 and 5, it was found that the current PV penetration
values are not of a concern to the operation of the grid in fact assist the grid at high
peak demand periods. However, their effect to change voltage levels cannot be
overlooked as they have a large influence when they operate collectively.
-In scenario 4, it was assumed that traditional DR can be used to reduce the PV
generation output. However, in the real world this is not that case as PV generation
can only be switched on or off at the consumer end by the consumers. Even if the
inverters were equipped with PV generation level controllers it would not be
practical to expect consumers to use them or to know when to reduce their PV
generation outputs. It is also a time when businesses are operating, requesting a PV
generation drop from a business during working hours would be counterproductive
as the business would generally consume most of the PV generation output.
-In scenarios 2 and 4, the effective DR load and generation were over-reduced from
the sensitivities calculated. This is due to the fact that any change to any bus has an
extended effect on the rest of the buses in the grid. Also, this could be due to the
reactive powers present in the grid.
Page 50
41
Chapter 5- Conclusion
5.1 Overview
This chapter will draw the conclusions from the project, the difficulties
encountered and future recommendations.
5.2 Conclusion
The project objectives were achieved with satisfactory results. The use of
traditional and effective DR management was investigated and the difference in
performance was illustrated through several scenarios. The results were consistent
throughout the simulation process thus supporting the findings.
From the results, it was found that DR was a beneficial strategy to maintaining bus
voltage limits. The use of an effective DR strategy proved to be a more efficient
strategy than traditional DR in maintaining bus voltages. However, the amount of
reduction from each bus needs to be limited to a maximum value as was seen
necessary from the results.
PV generation was found to be beneficial to bus voltage regulation in peak loading
scenarios. As future PV generation increases and PV penetration values rise,
regulations and strategies will be required to maintain bus voltages within limits at
peak PV generation periods.
It can be seen that the effective DR management method is a less resource-
demanding solution to the correction of bus voltage violation. This is beneficial in
reducing the need to overproduce electricity and therefore reduces the cost of
electricity bills of consumers. The environmental benefit is also an important
aspect, as less energy production equates to less carbon emissions.
Page 51
42
5.3 Obstacles and difficulties
The first obstacle was faced at the beginning of the project when searching for an
Australian LV grid to apply DR scenarios on. It was extremely difficult to find a
current operational grid to use for the project as most grids found lacked technical
information and parameters. This forced a new grid to be constructed by relying on
the manuals and standards used in Australia. The difficulty encountered when
constructing the grid was that the current grids operating in Australia varied in
some aspects from the standards highlighted by Horizon Power and Western
Power. For example, the voltage of a three phase LV network defined by Horizon
Power and Western Power is 415V, yet the grid details found for the Perth solar
city project expressed the voltage of the grid to be 400V. Also, the required voltage
limits of the buses in a grid differed as sources claimed a tolerance of +10% and -
6% volts while others required ±6 %.
Another challenge was component sizing. In technical manuals and standards, it
was required that components such as transformers and cables be sized to meet the
grid’s expected loadings. This was difficult as the loadings and lengths of cables
themselves had to be assumed for the project. This resulted in a design with
multiple variables which required many simulations using a lengthy trial and error
approach.
Another major obstacle was the use of the DIgSILENT PowerFactory software.
This software considers many parameters when conducting simulations. Most of
these parameters can be left as standard, however an extensive study of the user
manual is required to be able to understand the effects each parameter has on the
calculation output. Also, the software itself is not the most user-friendly and
requires some experience to use.
Page 52
43
5.4 Future recommendations
As a large portion of the project time was spent on grid construction, finding a
ready-built grid would be beneficial for future projects as this allows more focus
towards DR simulations. In this project, different scenarios were examined using
the same grid. Future studies could investigate the outcomes of using the same
scenario on different grids to expand the scenario spectrum of DR. The addition of
battery banks and large scale renewable DER’s could be added in future projects to
examine the use of DR in conjunction with these elements added to the grid.
From the effective DR management, it was noticed that there tended to be an over-
reduction of loads or PV generation. For future projects the effect of reactive power
on the DR process could be investigated. Also, the sensitivity calculation of the
effective DR process could include the sensitivities of each bus and its effect on
other buses in the grid for increased accuracy of load reduction values.
The system used for this project was a three-phase balanced system. Future projects
could investigate a three-phase unbalanced system which resembles a closer system
to the real world. Also, special conditions could be investigated in future projects,
such as DR under fault or line maintenance conditions where DR could be used to
assist the performance of the grid.
Page 53
44
References
Ausgrid. 2015. "Network Standard: NS125 Construction of Low Voltage Overhead Mains."
Ausgrid. December 01. Accessed October 25, 2016.
http://www.ausgrid.com.au/~/media/Files/Network/Documents/NS%20and%20NUS/N
S125.pdf.
Australian Energy Council. n.d. "Renewable Energy in Australia: How do we really compare?"
AEC. Accessed October 15, 2016.
https://www.energycouncil.com.au/media/1318/2016-06-23_aec-renewables-fact-
sheet.pdf.
Australian Energy Market Commission. 2012. "AEMC." FINAL REPORT: Power of choice
review - giving consumers options in the way they use electricity. November 30.
Accessed August 09, 2016. http://www.aemc.gov.au/Media/docs/Final-report-
1b158644-c634-48bf-bb3a-e3f204beda30-0.pdf.
Australian Standard. 2016. AS 2374.1.2-2003: Power Transformers, Part 1.2: Minimum
Energy Performance Standard (MEPS) requirements for distribution transformers.
Standards Australia.
Australian Standard. 2006. AS/ACIF S009:2006: Installation Requirements for Customer
Cabling (Wiring Rules). Standards Australia.
Australian Standard. 2009. AS/NZS 3008.1.1:2009: Electrical installations-Selection of cables.
Part1.1: Cable for AC voltages up to and including 0.6/1 kV - Typical Australian
installation conditions. Sydney: Standards Australia.
Cappers, Peter, Jason MacDonald, Jane Page, Jennifer Potter, and Emma Stewart. 2016. Future
Opportunities and Challenges with Using Demand Response as a Resource in
Distribution System Operation and Planning Activities. The Department of Energy
Offices of Energy Efficiency and Renewable Energy, Ernest Orlando Lawrence:
Berkeley National Laboratory.
CAT Projects & ARENA. 2015. "Investigating the impact of Solar Variability on Grid
Stability." Australian Renewable Energy Agency (ARENA). March. Accessed October
15, 2016. http://arena.gov.au/files/2015/03/150302-Impact-of-Variability-Report-for-
public-release.pdf.
Daintith, John. 2008. Dictionary of Computing. Oxford: Oxford University Press.
Department of Finance Utilities Office. 2016. "Final report: Reforms to the reserve capacity."
Government of Western Australia: Department of Finance Utilities Office. April.
Accessed October 25, 2016.
http://www.finance.wa.gov.au/cms/uploadedFiles/Public_Utilities_Office/Electricity_
Market_Review/Reforms-to-the-Reserve-Capacity-Mechanism-Final-Report.pdf.
Energy Safety WA. 2008. "WA Electrical Requirements." Government of Western Australia:
Department of Commerce, EnergySafety. July. Accessed October 25, 2016.
https://www.commerce.wa.gov.au/sites/default/files/atoms/files/waer_july_2008.pdf.
Hoke, Anderson, Rebecca Butler, Joshua Hambrick, and Benjamin Kroposki. 2012.
"Maximum Photovoltaic Penetration Levels on Typical Distribution Feeders." IEEE
Page 54
45
Transactions on Sustainable Energy. Accessed September 19, 2016.
http://www.nrel.gov/docs/fy12osti/55094.pdf.
Horizon Power. 2014. "Standard: Distribution Design Manual Vol 4 - Underground Cable
Distribution." Horizonpower. June. Accessed October 25, 2016.
https://horizonpower.com.au/media/1236/hpc_5dc_07_0004_2014___distribution_desi
gn_manual_vol_4_.pdf.
Horizon Power. 2013. "Technical Rules." Horizonpower. October. Accessed October 25, 2016.
https://horizonpower.com.au/media/1287/hpc_9dj_01_0001_2012_tec.pdf.
Hossain, Md Rahat, Amanullah M. T. Oo, and Shawkat A. B. M. Ali. 2013. "Chapter 2: Smart
Grid." In Smart Grids. London: Green Energy and Technology, Springer-Verlag.
Hughes, Thomas P. 1983. Networks of Power: Electrification in Western Society, 1880-1930.
London: The Johns Hopkins University Press.
Katiraei, Farid, Konrad Mauch, and Lisa Dignard-Bailey. 2007. "Integration of Photovoltaic
Power Systems in High-Penetration Clusters for Distribution Networks and Mini-
Grids." Internation Journal of Distributed Energy Resources 3 (3).
Kersting, William H. 2007. Distribution System Modeling and Analysis. Boca Raton: CRC
Press LLC.
Lasster, Robert H. 2003. "Distributed Generation." Power Systems Engineering Research
Center. Department of Electrical and Computer Engineering University of Wisconsin-
Madison. September 04. Accessed October 15, 2016.
http://pserc.wisc.edu/documents/publications/special_interest_publications/grid_reliabi
lity/lasseter_dg_pserc_sep_2003.pdf.
Lo, Chun-Hao, and Nirwan Ansari. 2013. "Decentralized Controls and Communications for
Autonomous Distribution Networks in Smart Grid." IEEE Transactions on Smart Grid
4 (1): 66-77.
Lucidchart. 2015. "Flowchart Symbols'Meaning Explained." Lucidchart software inc 2015.
Accessed October 25, 2016. https://www.lucidchart.com/pages/flowchart-symbols-
meaning-explained.
Nexans Olex. n.d. "Nexans Olex Cable Catalogue." Aerial.Accessed October 05, 2016.
http://www.olex.com.au/Australasia/2012/OLC12641_AerialCat.pdf.
Office of the Chief Economist. 2016. "Australian Energy Update." Australian Government:
Department of Industry, Innovation and Science. October. Accessed October 09, 2016.
http://www.industry.gov.au/Office-of-the-Chief-
Economist/Publications/Documents/aes/2016-australian-energy-statistics.pdf.
Pannell, David J. 1997. "Sensitivity Analysis of Normative Economic Models: Theoretical
Framework and Practical Strategies." Elsevier: Agricultural Economics 16 139-152.
Purchala, K, R Belmans, L Exarchakos, and AD Hawkes. 2010. Distributed Generation and
the Grid Integration Issues. London: Imperial College, KU Leuven. Accessed October
12, 2016.
http:/www.eusustel.be/public/documents_publ/WP/WP3/WP%203.4.1%20Distributed
%20generation%20%20and%20grid%20integration%20issues.pdf.
Page 55
46
Rahman, Md Moktadir Ali Arefi, GM Shafiullah and Sujeewa Hettiwatte. 2016. "Penetration
Maximisation of Residential Rooftop Photovoltaic using Demand Response."
ICSGTEIS 2016.
Siemens AG. 2016. "Planning of Electric Power Distribution: Technical Principles." Siemens.
Accessed September 19, 2016.
https://w3.siemens.com/powerdistribution/global/EN/consultant-support/download-
center/tabcardpages/Documents/Planning-
Manuals/Planning_of_Electric_Power_Distribution_Technical_Principles.pdf.
U.S Department of Energy. 2016. "Advanced Metering Infrastructure and Customer Systems."
Office of Electricity Delivery & Energy Reliability. Accessed October 20, 2016.
https://www.smartgrid.gov/recovery_act/deployment_status/sdgp_ami_systems.html.
U.S Department of Energy. 2006. "Benefits of Demand Response in Electricity Markets and
Recommendations for Achieving Them." Office of Electricity Delivery & Energy
Reliability. February. Accessed October 20, 2016.
http://energy.gov/sites/prod/files/oeprod/DocumentsandMedia/DOE_Benefits_of_Dem
and_Response_in_Electricity_Markets_and_Recommendations_for_Achieving_Them_
Report_to_Congress.pdf.
U.S Department of Energy. 2016. "Smart Grid: Demand Response." Office of Electricity
Delivery & Energy Reliability. Accessed October 20, 2016.
http://energy.gov/oe/services/technology-development/smart-grid/demand-response.
Utility Magazine. 2015. "Peak versus Average Demand Loads: Managing the Load." Utility
Engineering, Construction & Maintenance. April 28. Accessed October 25, 2016.
http://www.utilitymagazine.com.au/peak-versus-average-demand-loads-managing-the-
load/.
Wayne H, Beaty. 1998. Electric Power Distribution Systems: A Nontechnical Guide.
Oklahoma: Pennwell Publishing.
WesternPower. 2016. "Technical Rules." Economic Regulation Authority. August 01.
Accessed October 09, 2016.
https://www.erawa.com.au/cproot/14411/2/EDM%2040518689%20-
%20TECHNICAL%20RULES%201ST%20AUGUST%202016%20PUBLISH%20VE
RSION%20-%20FRI%20-%20R....pdf.
WesternPower. 2015. "Western Australian Distribution Connections Manual 2015." Accessed
October 09, 2016. https://www.westernpower.com.au/media/1495/wa-distribution-
connections-manual.pdf.
Willis, Lee H. 2004. Power Distribution Planning Reference Book. 2nd. New York: Marcel
Dekker, Inc.
Wolfs, Peter, and Syed Islam. 2009. "Potential Barriers to Smart Grid Technology in
Australia." 19th Australasian Universities Power Engineering Conference. Adelaide:
University of Adelaide. 1-6.
Page 56
47
Appendix A: Project grid parameters
The following parameters were used in the construction of the grid.
Table 17: Project transformer parameters
Transformer parameters
Rated power 200 kVA
Voltage step 22 kV /415 V
Rating factor 98.94%
Vector group YN / D
Positive sequence impedance
Short-circuit Voltage uk 4.30%
Copper losses 2.963 kW
Zero sequence impedance
Short-circuit Voltage uk0 0.614%
SHC-Voltage uk0r 0%
Magnetising impedance
No load current 0.10%
No load losses 0.424 kW
Table 18: Project transformer parameters
Nexans Olex Mercury
Aluminium
Overhead 3 phase lines 7/4.50 mm
Rated current rural 196 A
Impedance 0.315 R + j 0.259 X
Overall diameter 13.5 mm
Cross sectional area 111 mm2
Page 57
48
Table 19: Project cable lengths
From To Length m
Feeder Bus A 1 100
Feeder Bus B 1 100
Feeder Bus C 1 100
Feeder Bus D1 100
Bus W 1 LV bus (A1) 50
Bus X 2 LV bus (B3) 50
Bus W 2 LV bus (A3) 50
Bus W 2 LV bus (A4) 50
Bus Y 3 LV bus (C5) 50
Bus X 3 LV bus (B5) 50
Bus X 3 LV bus (B4) 50
Bus W 1 LV bus (A2) 50
Bus W 3 LV bus (A5) 50
Bus Y 1 LV bus (C1) 50
Bus Y 1 LV bus (C2) 50
Bus W 4 LV bus (A6) 50
Bus W 4 LV bus (A7) 50
Bus X 4 LV bus (B6) 50
Bus Y 3 LV bus (C6) 50
Bus X 1 LV bus (B1) 50
Bus Y 4 LV bus (C7) 50
Bus Z1 LV bus (D1) 50
Bus Z 2 LV bus (D2) 50
Bus Y 2 LV bus (C4) 50
Bus Y 2 LV bus (C3) 50
Bus X 2 LV bus (B2) 50
Bus W 1 Bus W 2 100
Bus W 2 Bus W 3 100
Bus W 3 Bus W 4 100
Bus X 1 Bus X 2 100
Bus X 2 Bus X 3 100
Bus X 3 Bus X 4 100
Bus Y 1 Bus Y 2 100
Bus Y 2 Bus Y 3 100
Bus Y 3 Bus Y 4 100
Bus Z1 Bus Z 2 100
Page 58
49
Table 20: Project bus parameters
Bus
Nominal
voltage V
Phase technology
Max voltage limit %
Min voltage limit %
22 kV 22000 ABC-N 6 -6
Bus W 1 415 ABC-N 6 -6
Bus W 2 415 ABC-N 6 -6
Bus W 3 415 ABC-N 6 -6
Bus W 4 415 ABC-N 6 -6
Bus X 1 415 ABC-N 6 -6
Bus X 2 415 ABC-N 6 -6
Bus X 3 415 ABC-N 6 -6
Bus X 4 415 ABC-N 6 -6
Bus Y 1 415 ABC-N 6 -6
Bus Y 2 415 ABC-N 6 -6
Bus Y 3 415 ABC-N 6 -6
Bus Y 4 415 ABC-N 6 -6
Bus Z 2 415 ABC-N 6 -6
Bus Z 1 415 ABC-N 6 -6
Feeder 415 ABC-N 6 -6
LV bus (A1) 415 ABC-N 6 -6
LV bus (A2) 415 ABC-N 6 -6
LV bus (A3) 415 ABC-N 6 -6
LV bus (A4) 415 ABC-N 6 -6
LV bus (A5) 415 ABC-N 6 -6
LV bus (A6) 415 ABC-N 6 -6
LV bus (A7) 415 ABC-N 6 -6
LV bus (B1) 415 ABC-N 6 -6
LV bus (B2) 415 ABC-N 6 -6
LV bus (B3) 415 ABC-N 6 -6
LV bus (B4) 415 ABC-N 6 -6
LV bus (B5) 415 ABC-N 6 -6
LV bus (B6) 415 ABC-N 6 -6
LV bus (C1) 415 ABC-N 6 -6
LV bus (C2) 415 ABC-N 6 -6
LV bus (C3) 415 ABC-N 6 -6
LV bus (C4) 415 ABC-N 6 -6
LV bus (C5) 415 ABC-N 6 -6
LV bus (C6) 415 ABC-N 6 -6
LV bus (C7) 415 ABC-N 6 -6
LV bus (D1) 415 ABC-N 6 -6
LV bus (D2) 415 ABC-N 6 -6
Page 59
50
Table 21: Project loads and PV generators, locations and values
Located at
PV
generator
Gen. Active
power kW
Gen. P.f
absorbing Load
Load Active
power kW
Load P.f
absorbing
LV bus (A1) A-PV 1 10 0.8 A-Load 1 10 0.95
LV bus (A2) A-PV 2 10 0.8 A-Load 2 10 0.95
LV bus (A3) A-PV 3 10 0.8 A-Load 3 10 0.95
LV bus (A4) A-PV 4 10 0.8 A-Load 4 10 0.95
LV bus (A5) A-PV 5 10 0.8 A-Load 5 10 0.95
LV bus (A6) A-PV 6 10 0.8 A-Load 6 10 0.95
LV bus (A7) A-PV 7 10 0.8 A-Load 7 10 0.95
LV bus (B1) B-PV 1 10 0.8 B-Load 1 10 0.95
LV bus (B2) B-PV 2 10 0.8 B-Load 2 10 0.95
LV bus (B3) B-PV 3 10 0.8 B-Load 3 10 0.95
LV bus (B4) B-PV 4 10 0.8 B-Load 4 10 0.95
LV bus (B5) B-PV 5 10 0.8 B-Load 5 10 0.95
LV bus (B6) B-PV 6 10 0.8 B-Load 6 10 0.95
LV bus (C1) C-PV 1 10 0.8 C-Load 1 10 0.95
LV bus (C2) C-PV 2 10 0.8 C-Load 2 10 0.95
LV bus (C3) C-PV 3 10 0.8 C-Load 3 10 0.95
LV bus (C4) C-PV 4 10 0.8 C-Load 4 10 0.95
LV bus (C5) C-PV 5 10 0.8 C-Load 5 10 0.95
LV bus (C6) C-PV 6 10 0.8 C-Load 6 10 0.95
LV bus (C7) C-PV 7 10 0.8 C-Load 7 10 0.95
LV bus (D1) D-PV 1 10 0.8 D-Load 1 10 0.95
LV bus (D2) D-PV 2 10 0.8 D-Load 2 10 0.95
Page 60
51
Appendix B: Simulation parameters and results
Scenario 1
Table 22: Scenario 1 parameters and simulation results
Load scale factor 0.4 Component Loading %
PV Gen. factor 0 Transformer 47.012
Feeder A 20.920
Feeder B 17.932
Feeder C 20.920
Feeder D 5.977
Bus
Bus
Vp.u
Load
kW
Load
kVAr
Gen
kW
Gen
kVAr
LV bus (A1) 0.981 4.00 1.31 0.00 0.00
LV bus (A2) 0.981 4.00 1.31 0.00 0.00
LV bus (A3) 0.976 4.00 1.31 0.00 0.00
LV bus (A4) 0.976 4.00 1.31 0.00 0.00
LV bus (A5) 0.973 4.00 1.31 0.00 0.00
LV bus (A6) 0.972 4.00 1.31 0.00 0.00
LV bus (A7) 0.972 4.00 1.31 0.00 0.00
LV bus (B1) 0.982 4.00 1.31 0.00 0.00
LV bus (B2) 0.977 4.00 1.31 0.00 0.00
LV bus (B3) 0.977 4.00 1.31 0.00 0.00
LV bus (B4) 0.974 4.00 1.31 0.00 0.00
LV bus (B5) 0.974 4.00 1.31 0.00 0.00
LV bus (B6) 0.973 4.00 1.31 0.00 0.00
LV bus (C1) 0.981 4.00 1.31 0.00 0.00
LV bus (C2) 0.981 4.00 1.31 0.00 0.00
LV bus (C3) 0.976 4.00 1.31 0.00 0.00
LV bus (C4) 0.976 4.00 1.31 0.00 0.00
LV bus (C5) 0.973 4.00 1.31 0.00 0.00
LV bus (C6) 0.973 4.00 1.31 0.00 0.00
LV bus (C7) 0.972 4.00 1.31 0.00 0.00
LV bus (D1) 0.985 4.00 1.31 0.00 0.00
LV bus (D2) 0.984 4.00 1.31 0.00 0.00
Page 61
52
Scenario 2
Table 23: Scenario 2 parameters and simulation results
Load scale factor 1 Component Loading %
PV Gen. factor 0 Transformer 117.224
Feeder W 52.301
Feeder X 44.829
Feeder Y 52.301
Feeder Z 14.943
Bus
Bus
Vp.u
Load
kW
Load
kVAr
Gen
kW
Gen
kVAr
LV bus (A1) 0.953 10.01 3.24 0.00 0.00
LV bus (A2) 0.953 10.01 3.24 0.00 0.00
LV bus (A3) 0.941 10.01 3.22 0.00 0.00
LV bus (A4) 0.941 10.01 3.22 0.00 0.00
LV bus (A5) 0.935 10.01 3.21 0.00 0.00
LV bus (A6) 0.930 10.00 3.20 0.00 0.00
LV bus (A7) 0.930 10.00 3.20 0.00 0.00
LV bus (B1) 0.955 10.01 3.25 0.00 0.00
LV bus (B2) 0.944 10.01 3.23 0.00 0.00
LV bus (B3) 0.944 10.01 3.23 0.00 0.00
LV bus (B4) 0.937 10.01 3.21 0.00 0.00
LV bus (B5) 0.937 10.01 3.21 0.00 0.00
LV bus (B6) 0.935 10.01 3.21 0.00 0.00
LV bus (C1) 0.953 10.01 3.24 0.00 0.00
LV bus (C2) 0.953 10.01 3.24 0.00 0.00
LV bus (C3) 0.941 10.01 3.22 0.00 0.00
LV bus (C4) 0.941 10.01 3.22 0.00 0.00
LV bus (C5) 0.935 10.01 3.21 0.00 0.00
LV bus (C6) 0.935 10.01 3.21 0.00 0.00
LV bus (C7) 0.932 10.01 3.21 0.00 0.00
LV bus (D1) 0.964 10.01 3.26 0.00 0.00
LV bus (D2) 0.962 10.01 3.25 0.00 0.00
Page 62
53
Table 24: Scenario 2 Change in bus voltage due to change in scaling factor
V p.u
Scaling factor
1 0.99 0.98 0.97 0.96 0.95
LV bus (A1) 0.95273 0.95319 0.95365 0.95412 0.95458 0.95504
LV bus (A2) 0.95273 0.95319 0.95365 0.95412 0.95458 0.95504
LV bus (A3) 0.94136 0.94193 0.94250 0.94308 0.94365 0.94422
LV bus (A4) 0.94136 0.94193 0.94250 0.94308 0.94365 0.94422
LV bus (A5) 0.93455 0.93519 0.93583 0.93647 0.93710 0.93774
LV bus (A6) 0.93002 0.93070 0.93138 0.93206 0.93275 0.93343
LV bus (A7) 0.93002 0.93070 0.93138 0.93206 0.93275 0.93343
LV bus (B1) 0.95501 0.95545 0.95589 0.95633 0.95677 0.95720
LV bus (B2) 0.94363 0.94418 0.94473 0.94528 0.94583 0.94638
LV bus (B3) 0.94363 0.94418 0.94473 0.94528 0.94583 0.94638
LV bus (B4) 0.93682 0.93744 0.93805 0.93867 0.93928 0.93990
LV bus (B5) 0.93682 0.93744 0.93805 0.93867 0.93928 0.93990
LV bus (B6) 0.93455 0.93519 0.93583 0.93647 0.93710 0.93774
LV bus (C1) 0.95273 0.95319 0.95365 0.95412 0.95458 0.95504
LV bus (C2) 0.95273 0.95319 0.95365 0.95412 0.95458 0.95504
LV bus (C3) 0.94136 0.94193 0.94250 0.94308 0.94365 0.94422
LV bus (C4) 0.94136 0.94193 0.94250 0.94308 0.94365 0.94422
LV bus (C5) 0.93455 0.93519 0.93583 0.93647 0.93710 0.93774
LV bus (C6) 0.93455 0.93519 0.93583 0.93647 0.93710 0.93774
LV bus (C7) 0.93228 0.93294 0.93360 0.93426 0.93492 0.93559
LV bus (D1) 0.96413 0.96448 0.96483 0.96518 0.96553 0.96588
LV bus (D2) 0.96185 0.96222 0.96259 0.96296 0.96334 0.96371
Figure 6: Scenario 2 Voltage VS Scaling factor plot
0.92
0.93
0.93
0.94
0.94
0.95
0.95
0.96
0.96
0.97
0.97
1 0.99 0.98 0.97 0.96 0.95
Vo
lta
ge
p.u
Scaling factor
LV bus (A1) LV bus (A2) LV bus (A3) LV bus (A4) LV bus (A5) LV bus (A6) LV bus (A7) LV bus (B1) LV bus (B2) LV bus (B3) LV bus (B4) LV bus (B5) LV bus (B6) LV bus (C1) LV bus (C2) LV bus (C3) LV bus (C4) LV bus (C5) LV bus (C6) LV bus (C7) LV bus (D1) LV bus (D2)
Page 63
54
Table 25: Scenario 2 load flow sensitivities
Bus
dv/dP
(Vp.u/MW)
dv/dQ
(Vp.u/MVar)
LV bus (A1) 0.391 0.457
LV bus (A2) 0.391 0.457
LV bus (A3) 0.606 0.627
LV bus (A4) 0.606 0.627
LV bus (A5) 0.821 0.797
LV bus (A6) 1.034 0.966
LV bus (A7) 1.034 0.966
LV bus (B1) 0.388 0.455
LV bus (B2) 0.603 0.625
LV bus (B3) 0.603 0.625
LV bus (B4) 0.817 0.794
LV bus (B5) 0.817 0.794
LV bus (B6) 1.021 0.959
LV bus (C1) 0.391 0.457
LV bus (C2) 0.391 0.457
LV bus (C3) 0.606 0.627
LV bus (C4) 0.606 0.627
LV bus (C5) 0.821 0.796
LV bus (C6) 0.821 0.796
LV bus (C7) 1.025 0.961
LV bus (D1) 0.377 0.449
LV bus (D2) 0.572 0.607
Page 64
55
Scenario 3
Table 26: Scenario 3 parameters and simulation results
Load scale factor 0.4 Component Loading %
PV Gen. factor 0.3 Transformer 15.380
Feeder W 6.803
Feeder X 5.831
Feeder Y 6.803
Feeder Z 1.944
Bus
Bus
Vp.u
Load
kW
Load
kVAr
Gen
kW
Gen
kVAr
LV bus (A1) 1.002 4.00 1.31 3.00 2.25
LV bus (A2) 1.002 4.00 1.31 3.00 2.25
LV bus (A3) 1.002 4.00 1.32 3.00 2.25
LV bus (A4) 1.002 4.00 1.32 3.00 2.25
LV bus (A5) 1.002 4.00 1.32 3.00 2.25
LV bus (A6) 1.002 4.00 1.32 3.00 2.25
LV bus (A7) 1.002 4.00 1.32 3.00 2.25
LV bus (B1) 1.002 4.00 1.31 3.00 2.25
LV bus (B2) 1.002 4.00 1.32 3.00 2.25
LV bus (B3) 1.002 4.00 1.32 3.00 2.25
LV bus (B4) 1.002 4.00 1.32 3.00 2.25
LV bus (B5) 1.002 4.00 1.32 3.00 2.25
LV bus (B6) 1.002 4.00 1.32 3.00 2.25
LV bus (C1) 1.002 4.00 1.31 3.00 2.25
LV bus (C2) 1.002 4.00 1.31 3.00 2.25
LV bus (C3) 1.002 4.00 1.32 3.00 2.25
LV bus (C4) 1.002 4.00 1.32 3.00 2.25
LV bus (C5) 1.002 4.00 1.32 3.00 2.25
LV bus (C6) 1.002 4.00 1.32 3.00 2.25
LV bus (C7) 1.002 4.00 1.32 3.00 2.25
LV bus (D1) 1.002 4.00 1.31 3.00 2.25
LV bus (D2) 1.002 4.00 1.31 3.00 2.25
Page 65
56
Scenario 4
Table 27: Scenario 4 parameters and simulation results
Load scale factor 0.4 Component Loading %
PV Gen. factor 1 Transformer 95.799
Feeder W 42.816
Feeder X 36.699
Feeder Y 42.816
Feeder Z 12.233
Bus
Bus
Vp.u
Load
kW
Load
kVAr
Gen
kW
Gen
kVAr
LV bus (A1) 1.052 3.99 1.31 9.99 7.47
LV bus (A2) 1.052 3.99 1.31 9.99 7.47
LV bus (A3) 1.063 3.99 1.30 9.98 7.46
LV bus (A4) 1.063 3.99 1.30 9.98 7.46
LV bus (A5) 1.069 3.98 1.30 9.97 7.45
LV bus (A6) 1.073 3.98 1.30 9.96 7.44
LV bus (A7) 1.073 3.98 1.30 9.96 7.44
LV bus (B1) 1.050 3.99 1.31 9.99 7.47
LV bus (B2) 1.061 3.99 1.30 9.98 7.46
LV bus (B3) 1.061 3.99 1.30 9.98 7.46
LV bus (B4) 1.067 3.99 1.30 9.97 7.45
LV bus (B5) 1.067 3.99 1.30 9.97 7.45
LV bus (B6) 1.069 3.98 1.30 9.97 7.45
LV bus (C1) 1.052 3.99 1.31 9.99 7.47
LV bus (C2) 1.052 3.99 1.31 9.99 7.47
LV bus (C3) 1.063 3.99 1.30 9.98 7.46
LV bus (C4) 1.063 3.99 1.30 9.98 7.46
LV bus (C5) 1.069 3.98 1.30 9.97 7.45
LV bus (C6) 1.069 3.98 1.30 9.97 7.45
LV bus (C7) 1.071 3.98 1.30 9.97 7.45
LV bus (D1) 1.042 4.00 1.31 9.99 7.48
LV bus (D2) 1.044 3.99 1.31 9.99 7.47
Page 66
57
Table 28: Scenario 4 Change in bus voltage due to change in scaling factor
Scaling factor
V p.u 1 0.99 0.98 0.97 0.96 0.95
LV bus (A1) 1.05123 1.05054 1.04985 1.04916 1.04848 1.04779
LV bus (A2) 1.05123 1.05054 1.04985 1.04916 1.04848 1.04779
LV bus (A3) 1.06183 1.06100 1.06017 1.05934 1.05851 1.05768
LV bus (A4) 1.06183 1.06100 1.06017 1.05934 1.05851 1.05768
LV bus (A5) 1.06857 1.06766 1.06674 1.06582 1.06491 1.06399
LV bus (A6) 1.07363 1.07266 1.07169 1.07071 1.06974 1.06877
LV bus (A7) 1.07363 1.07266 1.07169 1.07071 1.06974 1.06877
LV bus (B1) 1.04885 1.04819 1.04753 1.04687 1.04621 1.04556
LV bus (B2) 1.05900 1.05820 1.05740 1.05660 1.05579 1.05499
LV bus (B3) 1.05900 1.05820 1.05740 1.05660 1.05579 1.05499
LV bus (B4) 1.06530 1.06441 1.06352 1.06264 1.06175 1.06086
LV bus (B5) 1.06530 1.06441 1.06352 1.06264 1.06175 1.06086
LV bus (B6) 1.06800 1.06708 1.06617 1.06525 1.06433 1.06342
LV bus (C1) 1.05100 1.05031 1.04963 1.04894 1.04825 1.04756
LV bus (C2) 1.05100 1.05031 1.04963 1.04894 1.04825 1.04756
LV bus (C3) 1.06138 1.06055 1.05972 1.05889 1.05806 1.05723
LV bus (C4) 1.06138 1.06055 1.05972 1.05889 1.05806 1.05723
LV bus (C5) 1.06790 1.06698 1.06607 1.06515 1.06424 1.06332
LV bus (C6) 1.06790 1.06698 1.06607 1.06515 1.06424 1.06332
LV bus (C7) 1.07094 1.06999 1.06905 1.06810 1.06716 1.06621
LV bus (D1) 1.04062 1.04008 1.03954 1.03899 1.03845 1.03790
LV bus (D2) 1.04255 1.04198 1.04141 1.04083 1.04026 1.03969
Figure 7: Scenario 2 Voltage VS Scaling factor plot
1.02
1.03
1.04
1.05
1.06
1.07
1.08
1 0.99 0.98 0.97 0.96 0.95
Vo
lta
ge
p.u
Generation scaling factor
LV bus (A1) LV bus (A2) LV bus (A3) LV bus (A4) LV bus (A5) LV bus (A6) LV bus (A7) LV bus (B1) LV bus (B2) LV bus (B3) LV bus (B4) LV bus (B5) LV bus (B6) LV bus (C1) LV bus (C2) LV bus (C3) LV bus (C4) LV bus (C5) LV bus (C6) LV bus (C7) LV bus (D1) LV bus (D2)
Page 67
58
Table 29: Scenario 4 load flow sensitivities
Bus
dv/dP
(Vp.u/MW)
dv/dQ
(Vp.u/MVar)
LV bus (A1) 0.320 0.395
LV bus (A2) 0.320 0.395
LV bus (A3) 0.483 0.526
LV bus (A4) 0.483 0.526
LV bus (A5) 0.645 0.657
LV bus (A6) 0.808 0.790
LV bus (A7) 0.808 0.790
LV bus (B1) 0.321 0.396
LV bus (B2) 0.484 0.527
LV bus (B3) 0.484 0.527
LV bus (B4) 0.647 0.659
LV bus (B5) 0.647 0.659
LV bus (B6) 0.814 0.796
LV bus (C1) 0.320 0.395
LV bus (C2) 0.320 0.395
LV bus (C3) 0.483 0.526
LV bus (C4) 0.483 0.526
LV bus (C5) 0.645 0.657
LV bus (C6) 0.645 0.657
LV bus (C7) 0.812 0.794
LV bus (D1) 0.327 0.402
LV bus (D2) 0.500 0.544
Page 68
59
Scenario 5
Table 30: Scenario 5 parameters and simulation results
Load scale factor 1 Component Loading %
PV Gen. factor 0.3 Transformer 78.881
Feeder W 35.160
Feeder X 30.137
Feeder Y 35.160
Feeder Z 10.046
Bus
Bus
Vp.u
Load
kW
Load
kVAr
Gen
kW
Gen
kVAr
LV bus (A1) 0.974 10.01 3.27 3.01 2.25
LV bus (A2) 0.974 10.01 3.27 3.01 2.25
LV bus (A3) 0.967 10.01 3.26 3.01 2.24
LV bus (A4) 0.967 10.01 3.26 3.01 2.24
LV bus (A5) 0.963 10.01 3.26 3.01 2.24
LV bus (A6) 0.960 10.01 3.25 3.01 2.24
LV bus (A7) 0.960 10.01 3.25 3.01 2.24
LV bus (B1) 0.975 10.01 3.27 3.01 2.25
LV bus (B2) 0.968 10.01 3.26 3.01 2.24
LV bus (B3) 0.968 10.01 3.26 3.01 2.24
LV bus (B4) 0.964 10.01 3.26 3.01 2.24
LV bus (B5) 0.964 10.01 3.26 3.01 2.24
LV bus (B6) 0.963 10.01 3.26 3.01 2.24
LV bus (C1) 0.974 10.01 3.27 3.01 2.25
LV bus (C2) 0.974 10.01 3.27 3.01 2.25
LV bus (C3) 0.967 10.01 3.26 3.01 2.24
LV bus (C4) 0.967 10.01 3.26 3.01 2.24
LV bus (C5) 0.963 10.01 3.26 3.01 2.24
LV bus (C6) 0.963 10.01 3.26 3.01 2.24
LV bus (C7) 0.961 10.01 3.25 3.01 2.24
LV bus (D1) 0.981 10.01 3.28 3.00 2.25
LV bus (D2) 0.979 10.01 3.28 3.00 2.25
Page 69
60
Scenario 6
Table 31: Scenario 6 parameters and simulation results
Load scale factor 1 Component Loading %
PV Gen. factor 1 Transformer 46.840
Feeder A 20.933
Feeder B 17.943
Feeder C 20.933
Feeder D 5.981
Bus
Bus
Vp.u
Load
kW
Load
kVAr
Gen
kW
Gen
kVAr
LV bus (A1) 1.023 9.99 3.29 9.99 7.50
LV bus (A2) 1.023 9.99 3.29 9.99 7.50
LV bus (A3) 1.027 9.99 3.29 9.99 7.51
LV bus (A4) 1.027 9.99 3.29 9.99 7.51
LV bus (A5) 1.029 9.99 3.30 9.99 7.51
LV bus (A6) 1.030 9.99 3.30 9.99 7.51
LV bus (A7) 1.030 9.99 3.30 9.99 7.51
LV bus (B1) 1.023 9.99 3.29 9.99 7.50
LV bus (B2) 1.026 9.99 3.29 9.99 7.51
LV bus (B3) 1.026 9.99 3.29 9.99 7.51
LV bus (B4) 1.028 9.99 3.30 9.99 7.51
LV bus (B5) 1.028 9.99 3.30 9.99 7.51
LV bus (B6) 1.029 9.99 3.30 9.99 7.51
LV bus (C1) 1.023 9.99 3.29 9.99 7.50
LV bus (C2) 1.023 9.99 3.29 9.99 7.50
LV bus (C3) 1.027 9.99 3.29 9.99 7.51
LV bus (C4) 1.027 9.99 3.29 9.99 7.51
LV bus (C5) 1.029 9.99 3.30 9.99 7.51
LV bus (C6) 1.029 9.99 3.30 9.99 7.51
LV bus (C7) 1.029 9.99 3.30 9.99 7.51
LV bus (D1) 1.020 10.00 3.29 9.99 7.50
LV bus (D2) 1.021 10.00 3.29 9.99 7.50
Page 70
61
Appendix C
Definitions:
PV penetration value
The ratio between the total peak PV power generated and the peak load apparent
power on a feeder in a network is referred to as the PV penetration value (Hoke,
2012).
Voltage drop
The reduction of voltage due a current moving through an element with electrical
impedance is referred to as voltage drop. In a distribution system, the voltage drop
at a node is proportional to the distance from the energy source (Willis, 2004).
Voltage rise
Opposite to a voltage drop, a voltage rise is referred to as the increase in voltage at
a given node in a circuit in comparison to the energy source. In a distribution
system, a voltage rise occurs at a node where an additional energy source is
connected (Willis, 2004).
Peak to average load demand ratio
The ratio between the hourly average power demand and the peak hourly power
demand of a given network is referred to as peak to average demand ratio. This
ratio is an indicator to the variability of the power demand in a network (Daintith,
2008)
AC power
Power in a power distribution system is measured by the rate of flow of energy
passing through a certain point. AC power in a grid is made up of active and
Page 71
62
reactive power. The active power, often referred to as real power (P), is measured
in Watts (W). The reactive power, often referred to as reactive power (Q), is
measured in reactive Volt-Amps (Var). Theses powers are added as vectors taking
into consideration a phase angle between them. The sum of these powers makes up
the overall power referred to as the apparent power (S) which is measured in Volt
Amps (VA). The absolute value of the cosine of the apparent power is referred to
as the power factor (P.f). For a well performing grid the aim is to increase the P.f as
close to 1 as possible by reducing the reactive power. This reduces the line losses
in the grid and thus betters the overall performance and efficiency (Beaty, 1998)
Bus voltage limit
The voltage at a bus in comparison to the voltage at the energy source is referred to
as bus voltage. Bus voltages must remain within a certain limit to maintain the
appropriate power quality in a power delivery system (Beaty, 1998).
Electrical impedance
The extent of opposition that a circuit imposes on an electrical current when a
voltage is applied is referred to as electrical impedance. Electrical impedance
occurs with the flow of DC or AC currents. For a DC current the impedance is
purely resistive as only impedance magnitude is present. For an AC current, the
impedance possesses a magnitude and a phase angle. This adds a reactive element
to the impedance know as reactance. The reactive element of the impedance is
influenced by the component the AC current passes through. If the AC current
passes through an inductor, the reactance is referred to as inductance and is
measured in the form of positive imaginary impedance. If the AC current passes
Page 72
63
through a capacitor, the reactance is referred to as capacitance and is measured in
the form of negative imaginary impedance (Beaty, 1998).
Newton-Raphson load flow equations
The Newton-Raphson load flow equations for active and reactive powers used by
DIgSILENT PowerFactory 15.2 to calculate the bus voltages.
∑ | | | | | |
∑ | | | | | |
Where Pi is the active power of the ith
bus,
Qi is the reactive power of the ith
bus,
Yin∠θin is the admittance of the line from the ith
to the nth
bus,
Vi is the voltage magnitude if the ith bus,
δi is the voltage angle of the ith
bus
Page 73
64
Jacobian Matrix load flow sensitivity equations
[
]
[
]
Where f is a function of x,
For the load flow sensitivity analysis, the change in voltage is a function of the
change in active and reactive power. For the sensitivity of a bus voltage the
expression of J in Eq.3 is set up using:
f = the active power (P) or reactive power (Q)
x = the bus voltage magnitude (V) and angle of voltage vector (δ)
The J matrix becomes:
[
| |
| |
| |
| |
| |
| |
| |
| |
]
Page 74
65
The inverse of the Jacobian matrix shown in Eq.4 is then combined with Eq5.
A simplified expression for the change in voltage magnitude as a function of the
change in active and reactive powers is shown in Eq.6.
[
| || |
| || | ]
[
]
| | ∑ | |
| |