connecting renewable generation with active network management

132
Connecting Renewable Generation with Active Network Management MSc Dissertation Student : Ian Moore – Heriot Watt University Date : May-August 2008 Course : Renewable Energy and Distributed Generation Supervisors : Keith Brown – Heriot Watt University Dr. Gareth Harrison & Dr Luis Ochoa – Edinburgh University Institute of Energy Systems Declaration I the undersigned, Ian Moore, confirm that this work submitted for assessment is my own and is expressed in my own words. Any uses made within it of other authors in any form e.g. ideas, equations, figures, text, tables, programs etc are properly acknowledged. A list of references employed is included.

Upload: ian15moore

Post on 27-Apr-2015

518 views

Category:

Documents


4 download

DESCRIPTION

MSc dissertation reviewing and modelling the connection of a wind farm using Active Network Managemenent techniques. Note the pictures on the front are nothing to do with the work I did. It was desk based simulation work. The software is available free for educational use.

TRANSCRIPT

Page 1: Connecting Renewable Generation with Active Network Management

Connecting Renewable Generation with Active Network

Management

MSc Dissertation

Student : Ian Moore – Heriot Watt University

Date : May-August 2008

Course : Renewable Energy and Distributed Generation

Supervisors :

Keith Brown – Heriot Watt University

Dr. Gareth Harrison & Dr Luis Ochoa – Edinburgh University Institute of Energy Systems

Declaration

I the undersigned, Ian Moore, confirm that this work submitted for assessment is my own and is

expressed in my own words. Any uses made within it of other authors in any form e.g. ideas,

equations, figures, text, tables, programs etc are properly acknowledged. A list of references

employed is included.

Page 2: Connecting Renewable Generation with Active Network Management

Abstract

This report describes the evaluation of Active Network Management (ANM) applied to connection of

renewable energy sources on the electricity distribution network. ANM is a broad subject area but

typically implements local measurement and control methods on what has traditionally been a passively

operated distribution network. A key driver for the development of ANM is to help facilitate increased

connection of Distributed Generation (DG) to meet renewable energy generation targets.

The project consisted of three basic phases being review of research, comparison of solutions and lastly

simulation of a case study.

The area of research of ANM was found very active varying from individual power system component

technologies such as intelligent generator exciter controls to system wide solutions using advanced

computer science such as Multiple Agent Systems (MAS).

Basic DG connection problems were reviewed and consideration of three technology solutions to tackle

the important and common problem of voltage rise were detailed. Circuit theory illustrated voltage rise to

control by both real and reactive power.

For the simulation phase an educational version of the PowerWorld simulation software package and a

previously published case scenario were used. This verified the correct functioning of the PowerWorld

simulation setup and modelling. The case scenario looked at a classic problem of multiple feeders sharing

a common substation busbar with a wind farm connected at the end of one of the feeders. Four ANM

solutions were applied to tackle the voltage rise problem. Real power curtailment, reduced power factor

operation, improved voltage control of the main transformer and addition of a voltage booster.

Results showed that a trebling in connection capacity from 6 MW to 20 MW was achievable by using the

voltage booster applied in the line where the DG was present.

Losses were analysed for all cases and a characteristic 'U' shape was seen at lower DG connection

capacities breaking even at around 5 MW. Large incremental increases in network losses could occur but

these were not financially significant. Simple financial analysis showed a ROI of around 25% for up to 6

MW DG capacity for all cases.

A key aspect for all results was that curtailment resulted in a rapid reduction in revenue and return on

investment. The voltage booster showed continued high financial performance up to the maximum 20 MW

capacity used in the case scenario.

Caveats of using the simulation software included implementing 'Dummy' loads to model curtailment and

also close attention needed to be paid to circuit parameters and setup sensitivities.

Review of industry reports revealed that many complexities existed in implementing ANM on the network.

The use of PowerWorld to simulate DG connection with ANM in the project was thought to be innovative

in its speed of development.

Further work suggested includes investigation of simulation scripting and possible modelling of more

complicated ANM solutions such Area Based Control.

Page 3: Connecting Renewable Generation with Active Network Management

Acknowledgements

I would like to thank the following people for their support on the project :-

Keith Brown from Heriot Watt University for kindly offering to supervise the project and his advice

throughout the project.

Dr. Gareth Harrison and Dr. Luis Ochoa from Edinburgh University Institute of Energy Systems for

being the 'customers' of the project and their constant thorough technical reviews.

Santiago Grijalda at PowerWorld support for general help and development of the curtailment

method.

Predrag Djapic at Imperial College London for help in evaluating the UKGDS AMPT software.

Page 4: Connecting Renewable Generation with Active Network Management

Table of Contents

Abbreviations & Glossary....................................................................................................................7

1 - Introduction.....................................................................................................................................9

1.1 Scope..........................................................................................................................................9

1.2 Definitions..................................................................................................................................9

1.3 Incentives for ANM..................................................................................................................10

2 - Background Work..........................................................................................................................12

2.1 Rational....................................................................................................................................12

2.2 Objectives.................................................................................................................................12

2.3 Simulation Tools Evaluation....................................................................................................14

2.3.1 PowerWorld.......................................................................................................................14

2.3.2 PSCAD.............................................................................................................................15

2.3.4 UK Generic Distribution System (UKGDS) Active Management Planning Tool (AMPT)

...................................................................................................................................................15

2.3.5 PSS/E ...............................................................................................................................15

3 - Electricity Networks......................................................................................................................16

3.1 Power System Infrastructure.....................................................................................................16

3.1.1 Generation.........................................................................................................................17

3.1.2 Transmission.....................................................................................................................17

3.1.3 Distribution.......................................................................................................................18

3.2 Power System Operation..........................................................................................................19

4 - Connection Issues..........................................................................................................................21

4.1 DG Connection Problems ........................................................................................................21

4.1.2 Voltage Rise......................................................................................................................21

4.1.2.1 Theory.......................................................................................................................22

4.1.3 Stability.............................................................................................................................25

4.1.4 Fault levels........................................................................................................................26

4.1.5 Protection and coordination..............................................................................................26

4.1.6 Power Quality...................................................................................................................27

4.1.7 Economics........................................................................................................................27

4.1.8 Other Impacts...................................................................................................................28

4.2 Standards..................................................................................................................................28

5 - Current Research...........................................................................................................................30

5.1 DG & ANM Research Projects................................................................................................30

5.2 Technology Research Areas.....................................................................................................31

5.2.1 Intelligent Generator Control............................................................................................31

5.2.2 Distributed versus centralised control of voltage.............................................................31

5.2.3 Active Power Flow Management (APFM).......................................................................31

5.2.4 Network Capacity Planning and OPF..............................................................................32

5.2.5 Multi-agent Systems (MAS)............................................................................................33

6 – Connection Solutions and Implementation..................................................................................36

6.1.1 Overview...........................................................................................................................36

6.1.2 Estimating Connection Limits..........................................................................................36

6.2 Solutions..................................................................................................................................36

6.2.1 Voltage Control Techniques..............................................................................................37

6.2.2 Comparison......................................................................................................................39

6.2.3 Implementation.................................................................................................................41

6.3 Products available or in development.......................................................................................41

6.3.1 Dynamic Line Ratings......................................................................................................41

Page 5: Connecting Renewable Generation with Active Network Management

6.3.2 Area Based Voltage Controllers.......................................................................................41

6.3.4 SVCs and STATCOMs.....................................................................................................43

6.3.5 Commercially Available Plant..........................................................................................43

6.4 Specific Implementation Issues...............................................................................................44

7 – Simulation.....................................................................................................................................46

7.1 Overview..................................................................................................................................46

7.2 Description of the Network......................................................................................................46

7.3.1 Steady State Circuit Analysis................................................................................................48

7.3.2 Applying ANM to the network.........................................................................................50

7.3.3 Steady State Simulation Results.......................................................................................53

7.4 Time Series Simulation............................................................................................................55

7.4.1 Overview...........................................................................................................................55

7.4.2 Curtailment Modelling.....................................................................................................55

7.4.3 Spreadsheet.......................................................................................................................60

7.4.4 Simulation Procedure.......................................................................................................60

7.4.5 WF Power Factor..............................................................................................................61

7.4.6 Wind Turbine Output Profile............................................................................................63

7.4.7 Demand Profile................................................................................................................64

7.5 Development Methodology......................................................................................................65

8 - Results...........................................................................................................................................66

8.1 Generation Spreadsheet............................................................................................................66

8.1.1 Curtailment.......................................................................................................................66

8.1.2 General Observations........................................................................................................67

8.2 Losses Spreadsheet..................................................................................................................70

8.2.1 Circuit Losses...................................................................................................................70

8.2.2 General Observations.......................................................................................................71

8.3 Finance Spreadsheet.................................................................................................................76

8.3.1 Calculation of costings.....................................................................................................76

8.3.2 Cash Flows.......................................................................................................................78

9 - Discussion.....................................................................................................................................80

9.1 Curtailment Results..................................................................................................................80

9.1.1 Comparison to Book Results.............................................................................................81

9.1.2 Explanation of Differences...............................................................................................81

9.1.3 Effect of Wind Profile......................................................................................................83

9.1.4 Coincidence of Demand and Generation..........................................................................84

9.2 Losses.......................................................................................................................................84

9.2.1 Comparison to Book.........................................................................................................84

9.3 Finance.....................................................................................................................................85

9.4 Overall Sensitivities.................................................................................................................86

9.5 Suggested Incremental Improvements.....................................................................................86

9.6 Simulation Capabilities............................................................................................................87

9.7 Future Aspects of DG Connection...........................................................................................88

10 - Conclusion...................................................................................................................................89

10.1 Summary of main findings.....................................................................................................89

10.2 Contribution to subject field..................................................................................................89

10.3 Possible further work.............................................................................................................90

10.4 Overall evaluation of the project............................................................................................90

11 - Bibliography................................................................................................................................91

Appendix A1 – Extra Simulation Information...................................................................................94

A1.1 Steady State Circuit Screen-shots..........................................................................................94

Page 6: Connecting Renewable Generation with Active Network Management

A1.1.1 Original Case..................................................................................................................94

A1.1.2 PF 0.95...........................................................................................................................99

A1.1.3 LDC..............................................................................................................................104

A1.1.4 LDC & Voltage Booster...............................................................................................109

A1.2 – Crib Sheet.........................................................................................................................114

A1.3 – Individual Week Wind Profiles.......................................................................................115

A1.4 - Simplified Generation & Load Profile..............................................................................116

A1.5 – PW setup screenshots........................................................................................................117

A1.6 – Curtailment results for previous simulation......................................................................119

A1.7 – Loss Results and one-line diagram using reversed X/R...................................................120

Appendix A2 – Case study source....................................................................................................122

A2.1 - Wind Power in Power Systems Chapter 21.......................................................................122

A2.2 - UKGDS AMP Tool EHV-1 Circuit...................................................................................133

Appendix A3 – Project Errata..........................................................................................................134

A3.1 – CD File Listing.................................................................................................................134

A3.2 - Project Gantt Chart...........................................................................................................135

A3.3 - Email Dialogue with PowerWorld Support.......................................................................136

Appendix A4 – Miscellaneous..........................................................................................................144

A4.1 – Typical G59 Protection.....................................................................................................144

Page 7: Connecting Renewable Generation with Active Network Management

Abbreviations & Glossary

ANM – Active Network Management

AVC – Automatic Voltage Control. Used to control OLTC's on transformers for the purpose of

keeping voltages on the network within acceptable limits as load and/or GSP voltage varies.

Auto-recloser – Protection device used on rural circuits which tries to minimise interruptions by

automatically closing a tripped circuit breaker after a time delay to clear a transitory fault.

Capacity Factor - Applied to a Wind Farm means equivalent output in 1 yr as a factor of

theoretical generation for 1 yr at maximum wind speed.

CI – Customer Interruptions

Circle Curve – Graphical plot indicating real power output versus reactive power consumption for

an electrical machine

CML – Customer Minutes Lost

.COM – A windows based interface for accessing programs using a custom interface e.g Visual

Basic

DFIG – Doubly Fed Induction Motor

DG – Distributed Generation

DMS – Distribution Management System

DN – Distribution Network

DNO – Distribution Network Operator

DSM – Demand Side Management

EG – Embedded Generation/Generator, synonymous with DG

FACTS – Flexible AC Transmission

GSP – Grid Supply Point

H.V – High Voltage

I2R – Component of real power loss which is the result of the current squared times the resistance.

IFI – Innovation Funding Initiative

IG – Induction Generator (may also be termed Asynchronous generator)

IGBT – Insulated Gate Bi-Polar Transistor

IVR – In-line Voltage Regulator

L.V – Low Voltage

MAS – Multi Agent Systems

N-1 Contingency – Typical power system outage planning criteria which states that any single

circuit failure should not cause a knock-on failure in any other part of the network.

Normalised – Data series which has been proportionally resized to vary between the limits of zero

and one.

NPV – Net Present Value

Nuisance tripping – Operation of protective relays under non-genuine circuit faults. Typically

transient spikes.

OLTC – On Load Tap Changer, Transformer capable of changing turns ratio whilst still in circuit

O&M – Operation and Maintenance

OPF – Optimal Power Flow

PEI – Power Electronics Interface

PES – Public Electricity Suppliers

Page 8: Connecting Renewable Generation with Active Network Management

PFC – Power Factor Control

PQ – Power Quality

PSS/E – Simulation software package for power system analysis

PW - PowerWorld

ROC – Renewable Obligation Certificate, extra unit payment received for generating Renewable

electricity

ROI – Return On Investment

RTU – Remote Terminal Unit

RPZ - Registered Power Zone

State Estimation – Common technique used to predict analogue values without measurement.

SVC – Static Variable Compensator

TSS – Time Series Simulation, part of PowerWorld simulation package

Unbalanced Currents – Condition in a three-phase network when a single phase connected circuit

load causes cause imbalance.

VAr – Volt Amp Reactive, this refers to the reactive component of electrical power which serves

only to provide magnetising currents for inductive components. The negative quantity of this is

capacitance reactance. Both are 90 degrees out of phase with the active power component.

VC – Voltage Control

Weak Network – Rural network which as high impedance plant and circuits. This leads to small

fault currents which result in poor discrimination and also high voltage rises when connecting DG.

This term is closely related to 'fault level'.

Page 9: Connecting Renewable Generation with Active Network Management

1 - Introduction

This project provides a review of problems of connecting Distributed Generation (DG) to the

distribution network using Active Network Management (ANM). Also current research into this

area is highlighted and a simulation case study is undertaken to demonstrate the application of

ANM solutions.

1.1 Scope

In order to give the project sufficient focus in what is large subject area voltage management is

taken as the main topic. The subject of faults is only given a brief mention along with management

of thermal constraints.

1.2 Definitions

Distributed Generation (DG)

Basically DG is electrical generation plant of any type which is not connected to the transmission

network but connected at the distribution network level. This type of generation is in contrast to the

vast majority of the UK generating capacity which connects directly to the transmission system for

transfer of bulk power to load centres on the distribution network. The term DG is synonymous with

Embedded Generation (EG).

DG can exist for many forms and for different reasons including the following :

� Standby Diesel Generation

� Micro Hydro

� Wind Farms

� Domestic P.V

� Co-generation for industrial of heating schemes

Recommend size definitions are taken from [1] are shown in the table 1.1 below

Definition Size

Micro 1 W to 5 kW

Small 5 kW to 5 MW

Medium 5 MW to 50 MW

Large 50 MW to 300 MW

Table 1.1 – DG size definitions

1-Introduction

Page 10: Connecting Renewable Generation with Active Network Management

Active Network Management (ANM)

A good definition of ANM and its distinction from another closely related area of distribution

network and control is as below

“ Abstract1-- Active Network Management (ANM) has emerged

as the common, collective term in the UK for the automatic

management and control of distributed energy resources (DER)

in distribution networks. DER is the main driver of research and

development in automation in the UK at present but network

reliability performance enhancement remains the most common

use of automation in power network applications.”[2]

In the US ANM is driven by issues of reliability and is known under the banner of Advanced

Distribution Automation (ADA). [2]

1.3 Incentives for ANM

Increasing the quantity of DG which can connect to the network and also maximising its energy

output onto the network without having to invest in expensive line, equipment and transformer

upgrades/changes is a key driver for the adoption of ANM. By using measurement and control of

voltages and powerflows on the distribution network to control the above values the network can

integrate a greater number and quantity of energy from DG than it's normally passive topology.

The Registered Power Zone (RPZ) scheme and the Innovation Funding Initiative (IFI) in the UK are

being used to encourage DNO participation in ANM and Distribution Automation (DA) research

and development. The importance of the longer term R&D work is highlighted as

“ Of these 11 project areas, the longer-term issues of

Network Design and Active Management are regarded as

central to the future of distribution systems. In these two

specific areas, activities are underway to map out future

configurations of the distribution network itself and the

control and communications infrastructure required to

manage more complex future situations.”[2]

1-Introduction

Page 11: Connecting Renewable Generation with Active Network Management

Automation of the DN does already exist to some extent but this is a different type of area to ANM

although a mutually beneficial cross over is expected to develop.

“ In the short to medium term, active network management

solutions are likely to be rolled out into distribution network

as problems arise with distributed generation. In the longer

term it is likely that distribution network operators will

consider combining the functions of traditional distribution

automation with those of active network management for

distributed generation. This will occur as necessitated by the

interaction and overlap in control at a local level. However,

combination and accumulation of functionality will also occur

as the benefits of coordinated action from active network

management and distribution automation schemes becomes

more apparent and accessible.”[2]

1-Introduction

Page 12: Connecting Renewable Generation with Active Network Management

2 - Background Work

2.1 Rational

Increasing demand globally for energy, the connection between fossil fuel and global warming

together with energy security issues is driving the requirement for increased use of renewable

energy sources. The nature of RE is that it is essentially dispersed. Electrical power generation by

methods such as wind farms, district heating CHP, biomass and other future technologies such as

wave and tidal all require a method of distribution to consumers.

Integration of these dispersed generation sources into the existing transmission and distribution

network with its existing centralised power generation topology presents a number of engineering

challenges.

Business, organisational and social challenges also exist in facilitating the production and

distribution of energy resources. The essentially monopolistic nature of the business and capital

infrastructure of the grid and the dependence of day to day business and lifestyles on the continuous

and reliable nature of the existing electricity supply make changes such as increasing the penetration

of DG particularly interesting.

With capital assets averaging 50 years and some as old as 100 years and the high cost of

replacement of these the industry is looking for solutions to all of the above problems whilst

maximising the utility of this existing infrastructure.

ANM is targeted to help tackle some of these problems by facilitating the easy connection of DG

using minimum resources and where applicable innovative technology.

2.2 Objectives

These were initially stated in a portfolio submitted at the beginning of the project and are repeated

again here :-

2 - Background

Page 13: Connecting Renewable Generation with Active Network Management

2 - Background

The three main objectives are to -

Provide a brief overview of :

� Research being done in the area of ANM

� Connection problems associated with DG and in particular

Renewables

� Methods/Solutions for their connection

� Nature of market place and touch upon some issues of T&D planning

Compare and contrast the connection solutions in terms of their :

� Performance

� Cost

� Maturity

� Complexity etc

Illustrate where possible by undertaking simple simulations the problems and

available solutions to DG connection using relevant software tools particularly :

� Local Voltage Control

� Power Flow management

� including Intermittent generation from Renewables

Page 14: Connecting Renewable Generation with Active Network Management

2.3 Simulation Tools Evaluation

2.3.1 PowerWorld

This computer program from PowerWorld Corporation of Illinois USA is available as a demo

version which has a limitation of a maximum of 12 buses instead of its full capability of 100,000.

The demo version is licensed for educational and evaluation use only.

The basic functions of this simulator are to find the steady state power flow solution to a power

system network which has been input graphically. Multiple generators, loads, transmission lines,

transformers and other system components such as compensators are drawn. The simulator will then

find the operating point of the system. For basic theory of solving power flow (or load flow) see

[3] .

The program is not a transient simulator for analysis on a sub second basis but provides simulations

which are useful for power system operational decisions on an hour by hour basis or for planning of

new lines or generation capacity. For example how should generator units be scheduled under

different pricing scenarios. The program is targeted for power system operators and has extensive

capability for analysing operational costs and other economic factors for large power systems which

have multiple business operators.

Summary of features

Basic capabilities are:

� Load Flow solution

� Contingency analysis (e.g what happens if a generator suddenly trips)

� Determination of fault currents

� Time Series Simulation for solving profile based problems

Power system components functionality includes :

� Remote voltage regulation by OLTC's, generators and switched shunts

� Shared voltage regulation by generators

� Phase-shifting transformer control

Financial functions include :

� Marginal Pricing for each bus

� Generator Cost Models

� Participation Factor control of generators

� Dispatching of power pools

The following Add-on modules are however also included in this demo version :

ATC – Available Transfer Capability for determining incremental MW transfer

PVQV – For studying voltage stability problems

OPF – Optimal dispatching of system to achieve a certain goal e.g lowest cost

SCOPF – Optimal dispatching including constraints cases

2 - Background

Page 15: Connecting Renewable Generation with Active Network Management

Automation features:

� A COM or .NET interface is available for direct control of the simulator called AutoSim

(this is not present on the demo version).

� Some scripting through basic interface is available.

Basic tutorials were completed for the following

� Creating a new Case

� Starting with an existing case

� OPF

� Contingency Analysis

More information and the demo program is available from [4].

2.3.2 PSCAD

This is a transient analysis electrical circuit simulation tool for use in power systems. Sub micro-

second level simulation including harmonic and control system level modelling is possible. Similar

to PowerWorld circuits are entered graphically. For transient analysis of faults or control system

schemes this program may prove useful. However for this projects investigation into ANM the

steady state problem domain is more relevant so PSCAD will not considered suitable. Information

and a demo version of this program is available from [5].

Note this tool was previously used by the author to model the effect of the addition of wind-farms

(consisting of distributed Induction Generators) on 3-phase network faults levels. Hence it's basic

capabilities are known.

2.3.4 UK Generic Distribution System (UKGDS) Active Management

Planning Tool (AMPT)

The UKDGS project provides a library of distribution network models for use with the AMPT

simulator. The tool is in fact an excel spreadsheet with attendant VB code and a DLL which

contains an optimal power flow 'engine'. Both of these resources are for public use and are freely

available via the internet [6]. This tool is directly applicable to the problem domain of this project as

it was developed specifically for the analysis of active distribution networks. Initial simulation runs

using a 16 bus example gave seemingly odd results.

2.3.5 PSS/E

(Power System Simulator for Engineering). This is an industry standard simulation tool from Power

Technologies International (PTI) a group within Siemens. It is one of the simulation tools used by

Edinburgh University Institute of Energy Systems. It is not available as a demo version. See [7] for

more information about this tool. It was considered too complex to use for the time frame of this

project.

Choice of Simulation Tool for the project

PowerWorld was chosen as this proved simple to use and gave immediate results.

2 - Background

Page 16: Connecting Renewable Generation with Active Network Management

3 - Electricity Networks

3.1 Power System Infrastructure

This section is based on information from [3,ch.1.7] unless where otherwise referenced. The

electricity system is basically made up of generation, transmission, distribution and loads.

Interconnections enable efficient scheduling of generation capacity, sharing of spinning reserve and

security in the event of failures of plant or lines.

Some single line representations of some typical parts of three-phase power systems are shown in

the following figures. Figure 3.1 below shows standard symbols for power system components.

Fig. 3.1 – Three phase power system component symbols [3, pg.40]

3 – Electricity Networks

Page 17: Connecting Renewable Generation with Active Network Management

Fig. 3.2 – Section of a typical power system [3,pg.42]

3.1.1 Generation

Electrical power is normally produced at centralised power stations of the order of 1GW or greater.

Each power station may have a number of turbine/generator units of around 250MW each outputting

at a voltage of 11-25kV. These generators directly connect to the transmission system via

appropriate step-up transformers.

3.1.2 Transmission

Bulk transmission of power is through high voltage normally overhead lines at voltages of 400kV or

250kV. A sub-transmission network also exists which uses 132kV. High voltages are used to

minimise power losses whilst also minimizing conductor sizes.

3 – Electricity Networks

Page 18: Connecting Renewable Generation with Active Network Management

3.1.3 Distribution

Nearer the point of use voltages are stepped down to 11kV and finally 400V. Domestic customers

use a single phase to neutral of the 400V giving 230V. Industrial and commercial users take 400V

three-phase and larger users may take it at 11kV stepping the voltage down as required on site.

Distribution systems have a huge number of transformers and circuits compared to the transmission

system.

Rural Distribution

11kV overhead lines are stepped down at single points by pole-mounted transformers which then

feed radial circuits. Transitory flashover faults on these circuits requires the use of auto-reclosers to

minimise blown fuses and resulting circuit outages. Loads are usually small and well dispersed.

Figure 3.3 below shows a typical rural circuit.

Fig. 3.3 – Typical rural distribution system [3,pg.45]

Suburban Distribution

These originally developed from radial circuits but include interconnections at the lower voltage

points. Sectionalization also allows containment of short circuit levels and better management of

protection. Extra capacity can be added by extra step-down transformers without the need for cable

upgrade works. Instead of overhead lines H.V underground cables may be used and in highly

populated areas like cities substations might also be located underground. Overlapping of circuits

can provide extra security of supply. Figure 3.4 below shows an example of L.V cables running

down both sides of a set of roads.

3 – Electricity Networks

Page 19: Connecting Renewable Generation with Active Network Management

Fig. 3.4 – Typical Suburban Distribution system [3,pg.47]

3.2 Power System Operation

This section is based on information from [8,ch.13.3.2]

Technical

Centralised control of the electricity system is based around the transmission system. This has a

number of methods of control from tap-changing transformers, use of FACTS devices and also

sequencing and operation of circuit switching devices. The transmission system also has an

attendant amount of measurement systems which its enable effective control. Balancing of the

system is achieved through generator scheduling, operation of a small quantity of pumped storage

and occasionally load shedding.

Large power stations are equipped with synchronous generators which are capable of operation at

constant power factor and also sometimes frequency control (variation of real power by turbine

governor). All of this contributes to a manageable electrical power system.

In contrast the distribution system has little control or measurement and is essentially 'Passive' in its

operation. Automatic voltage control through transformers equipped with OLTC's occurs on 11kV

and 33kV circuits however local voltage adjustment is done via off-load tap-changing (I.e circuit

outage required)[8]. In addition powerflows are 'uni-directional' nearly always flowing from the

GSP transformer through the distribution system to the load (some reverse flow in meshed circuits

might be expected).

3 – Electricity Networks

Page 20: Connecting Renewable Generation with Active Network Management

Voltage limits for the UK are shown table 3.1 below.

Location Operating Limits

Transmission 132kV +/-10%

H.V Distribution [9] 11kV +/- 6%

L.V Distribution [9] 230V +10/-6%

Table 3.1 – UK Operational Voltage Limits

Market

For the majority of load and generation the energy purchasers typically a local network operator

company will buy power directly from the power station operating company. This 'bulk' power

market operates on a long term basis and is called the 'bilateral market' and thus scheduling of

demand and supply occurs without the intervention of the transmission operator.

Balancing of demand and supply on a closer to real-time basis is done in the 'balancing market' over

the time frame of one hour. This market is participated in by the Transmission System Operator

(TSO).

Other issues

Optimisations which might be resolved in typical power systems at the transmission operational

level can include minimisation of losses, ensuring security, safety and economic dispatch of plant.

Distribution network operators have a requirement to maintain and operate their network to ensure

minimum standards of availability and power quality. Customer Minutes Lost (CML) and Customer

Interruptions (CI) are an example of two measures used to quantify the network performance.

3 – Electricity Networks

Page 21: Connecting Renewable Generation with Active Network Management

4 - Connection Issues

4.1 DG Connection Problems

The general characteristics of DG are :

� De-centralised

� Located on the distribution network which has limited control and monitoring

capabilities

� Large number of small installations (or potential for)

� Large number of nodes on distribution network compared to transmission network

� Changes power flow direction with implications for equipment

� No standard electrical machine or network connection is applied

� A new market player in the production and sale of electricity

The important problems are discussed in the following sections.

4.1.2 Voltage Rise

The voltage rise scenario is most severe when DG is connected at the ends of long radial feeders

where voltages may be already close to the limits. Voltage regulation set points for rural distribution

transformers would be adjusted to regulate their terminal output in the upper limits of accepted

tolerance. This should mean that the last user on a radial circuit should have a high enough voltage

and also I2R losses would be marginally reduced reduced. 'Weak' rural networks will typically

exhibit these properties. These variations in voltages along a circuit are characterised as 'voltage

profiles'.An illustrative example is shown below

Fig. 4.1 – Voltage profile along heavily loaded line taken from [10]

4 - Connection Issues

Page 22: Connecting Renewable Generation with Active Network Management

With DG connected near the ends of feeders two scenarios give rise to the possibly of the voltages

exceeding the statutory limits.

� Minimum load and maximum generation -

� No generation and maximum load – This case is would be most relevant when substation

terminal voltage setpoints have been set lower than normal to compensate for a permanently

operating DG plant under normal generation conditions.

These are illustrated in the below figure which shows the effect of connecting an EG a distance of

12km from the substation in the previous example.

Fig 4.2 – Voltage profile with EG taken from [10]

4.1.2.1Theory

Since voltage rise is such an important issue the theory of how it relates to power flows is explained

here.

Solution of Power flows

Consider a two bus power system as below

Fig. 4.3 – Two bus power system

4 - Connection Issues

G Load

Z = R + jX

Pg + jQg

I

VgVload

Page 23: Connecting Renewable Generation with Active Network Management

The apparent power transferred is given by

S = Pg + jQg (4.1)

Dividing both sides by voltage

I = (Pg-jQg) (4.2)

Vg

Voltage at the load must be the Voltage at Generator less the voltage drop across the line

Vload = Vg – (R+jX)I (4.3)

Combining the above equations for Vload and I

Vload = Vg – (R+jX)(Pg-jQg) (4.4)

Vg

Multiplying out the top line of this equation

Vload = Vg – (RPg+XQg) – j(XPg-RQg) (4.5)

Vg Vg

It can be shown that vice-versa for the generator the voltage in terms of load and load voltage is

Vg = Vload – (RPload + XQload) + j(XPload-RQload) (4.6)

Vload Vload

It must be noted that some of the real and reactive power entering the line will be consumed by the

line itself. Using real and reactive power at the load we can derive another equation using the same

process as above.

S = Pload + jQload (4.7)

4 - Connection Issues

Page 24: Connecting Renewable Generation with Active Network Management

Dividing both sides by Voltage as in (4.2) and inserting into equation (4.3) we get

Vload = Vg – (R+jX)(Pload-jQload) (4.8)

Vload

The above equation is used in power simulation software programs for solving power flows.

Significance of power flows

Another way of writing (4.6) is

Vload = Vg - ∆V - jV (4.9)

Where the in phase component is

∆V = RPg + Xqg (4.10)

Vg

and the quadrature component is

V = XPg + Rqg (4.11)

Vg

These relationships can seen be in the following vector diagram figure 4.4.

Fig. 4.4 – Vector representation of line voltage drop

Power Systems typically operate at load angles below 30 degrees for various reasons including

stability.

It can be seen that the voltage drop between the different bus bars ∆V is caused by a combination of

the product of real power and resistance and the product of reactive power and reactance.

Conversely for the phase angle difference is caused by a combination of the product of real power

and reactance and the product of reactive power and resistance.

These comparisons are more useful when considering transmission lines which have considerably

higher reactances than resistance. Hence they can be simplified to the below approximations

4 - Connection Issues

Load Angle

- jV

- ∆V

Vg

Vload

Page 25: Connecting Renewable Generation with Active Network Management

∆V = Xqg (4.12)

Vg

V = XPg + Rqg (4.13)

Vg

Thus for transmission systems voltage drop is largely determined by reactive power transfer and load

angle is determined largely by real power transfer.

Application to DG

Since DG is likely to have an equal contribution to voltage drop from real and reactive power flows

both curtailment and reactive power control can be effective. The effect of controlling the net flows

of real and reactive power on a bus can be approximated in p.u by dividing (4.10) by Vbase.

∆Vp.u = RPg + XQg

Vg/Vbase

∆Vp.u = RPg + XQg (4.14)

4.1.3 Stability

Issues here can be briefly divided on the size of the DG schemes.

Small DG

These typically do not present a problem as small installations of DG will respond to network faults

by disconnecting themselves if network conditions exceed operational limitations. When normal

conditions resume the DG should automatically reconnect. However a possible issue is interaction

between machine inertia and tripping times which may cause instability. See [11,pg.77]

Large DG

If the installation is large enough then an interaction on the overall stability of the power system

may occur. Two examples are illustrative of possible problems.

Nuisance Tripping - Sudden loss of a large generation plant may causes a transient drop in system

voltage. If protection designed to detect islanding for a WF operates this will further suppress the

network voltage by removing real power generation from the system.

Post Fault Restoration – On restoration of an outage if large amounts of load were previously served

by the DG this load will need to be supplied via the grid until DG comes back on-line (DG will wait

until network is properly restored).

4 - Connection Issues

Page 26: Connecting Renewable Generation with Active Network Management

Other

Induction Generators (IG) can draw large reactive currents on overspeed and contribute to

instability.

4.1.4 Fault levels

A property of rotating machinery is that when short circuits occur very high currents can flow into

the fault. This is analogous to high currents taken when starting rotating machines. IG fault currents

decay rapidly with collapse in rotor magnetising current. Synchronous machines will contribute

sustained currents largely dependent on the capability of the rotor excitation equipment.

Overall increases in fault levels are largely a problem in heavily meshed urban networks where

contribution of current into faults from DG may mean contact breaker and other such equipment

ratings are exceeded.

4.1.5 Protection and coordination

Three general areas of protection are relevant :

Internal Generator Faults

This is often easily dealt with by standard motor/generator protection relays which operate by

current supplied from the network the only possible problem being lack of sufficient fault current to

operate the protective relays/fuses etc on weak rural networks.

Network Faults & Islanding Protection

This presents a particular problem due to the inability of induction generators and conventionally

excited synchronous machines to supply large over-currents when close-up faults occur on the

network. In order to stop the DG plant continuing to supply current into the fault the network

protection must first isolate the fault and then the DG must isolate itself. This sequencing is done by

over/under voltage, over/under frequency or Loss-Of-Mains protection gear on the generator.

Operation in an 'islanded' mode is not common as safety, regulatory, administrative, neutral

grounding issues and sequencing where auto-reclosers are present make it a complicated issue.

Distribution System Protection

Existing network protection systems can be affected by introduction of DG due to change to fault

currents sizes and directions. Some protection relays can be assigned zones of operation and are

known as 'distance' relays. Effectively they measure the impedance of the line during the fault to

ascertain the location of the fault and operate accordingly. Changes in fault currents due to DG can

effect the operation of these relays by effectively changing the impedance of the circuit. Another

issue is some relays are uni-directional and may need to be replaced if power flow directions are

changed by introduction of DG. An example of typical hardware used in protecting a DG

installation is included in the appendix A4.1.

4 - Connection Issues

Page 27: Connecting Renewable Generation with Active Network Management

4.1.6 Power Quality

Change in transient voltage variations and harmonic distortions represent the two main effects DG

can have on Power Quality.

Large current changes due to connection and disconnection of EGs can lead to unacceptable spikes

in voltage especially when single machines are located on 'weak' (high impedance) networks.

Careful sequencing of induction generator starts together with soft-start capacitors and properly

synchronisation of alternator type machines can limit starting transients however tripping of

generators on full load may still cause large transients although less frequent. Cyclic disturbances

called 'flicker' from asynchronous equipped wind turbine can occur due to tower-blade interactions.

Importantly in some circumstances transient noise reduction can actually be reduced by connection

of DG in some networks as the low impedance of the generator leads to an overall reduction in

impedance of the network and hence lower voltage rise associated with any currents flowing.

Harmonics can be introduced by poorly designed Power Electronic Interface (PEI) units on

generators but also reduced by directly connected stators which once again reduce impedance levels.

Induction generators however are in this respect particularly susceptible to unbalanced currents

which might commonly be present in a rural network and cause overheating of the windings.

4.1.7 Economics

Network Losses

Change in power flow directions and the relative location of DG to load centres will affect the size

of network losses. DNOs are incentivised to minimise losses. Also DG operators which cause losses

have to pay for the increased losses which they cause. See case study in later chapter for example of

how losses can vary when DG is connected.

Connection charging

A pertinent issue to the financial viability of any DG scheme is how connection costs are paid for.

'Deep' charges refer to when a developer is expected to pay for the full cost of any works associated

with the connection to the network and also any extra operational costs incurred. 'Shallow'

connection means that only a portion of the costs are paid for by the DG developer. The issues

surrounding this are that these upgraded assets which become the property of the DNO may benefit

other customers by creating easy access for new load points, new DG connections. Other possible

benefits are improved PQ and early refurbishment of assets leading to reduced CML and CI. Hence

'Shallow' connection charging could be deemed fairer in terms of apportionment of costs although

the DNO would have to bear these upfront costs.

Note that as part of the de-regulated electricity market since 1990 DNO's are required by the market

regulator to provide equal connection opportunities to would be generators.

4 - Connection Issues

Page 28: Connecting Renewable Generation with Active Network Management

Ancillary Services

Generation participation in control of network voltage is normally only undertaken by certain large

generators on the transmission system. Operation of DG to contribute to this control through use of

variable reactive power consumption or production is a future service which DG might contribute

too. In general DG is viewed as 'negative load' on the network but in future it may be recognised as a

part of the generating capacity. Similar potentials apply to contribution to network security.

4.1.8 Other Impacts

As greater levels of penetration occur some impacts on the overall UK power system will occur.

� Increased in variance of power output of centralised generation

� Decrease in the mean power output of centralised generation

For the UK some particularly relevant technical issues are penetration of DG and connection of

large schemes to rural networks. For more information on the first of these see [12]. Regarding

managing high penetrations of DG Danish CHP is a current example the dispersed generation being

28% of total capacity [11,pg.5]. Although the easy access to the European electricity grid compared

to the UK would be expected to facilitate high penetration levels.

4.2 Standards

Guidelines exist for the connection of DG and are set out in the table 4.1 below. However it may be

noted that the regulations are not subject to regular updates so may be advancing at a slower rate

than the implementation of DG.

4 - Connection Issues

Page 29: Connecting Renewable Generation with Active Network Management

Name Issuing Body Comment

Distribution Code Local DNO DNO regulations which relate to

G59/1 Connection of Embedded

Generators

Electricity Association Recommendations for the

connection of embedded generating

plant to the PES distribution

system.

G75 Electricity Association Recommendations for the

connection of EG plant to PES

distribution systems above 20kV

with outputs over 5MW.

G83/1 <16A per phase Electricity Association Small scale domestic applications

P2/6 Security Standard Requirements more voltages and

safety

ETR 113 Protection of Embedded

Generators < 5MW

Energy Networks Association Notes of guidance for the protection

of private generating sets up to

5MW in parallel with PES

distribution systems.

ETR 124 – Guidelines for actively

managing power flows associated

with the connection of a single

distributed generation plant

Energy Networks Association

Electricity at Work Regulations

(1989)

Government Applies to all electrical installations

in the UK

Table 4.1 – Applicable grid regulations & guidelines for DG connections in the UK

4 - Connection Issues

Page 30: Connecting Renewable Generation with Active Network Management

5 - Current Research

5.1 DG & ANM Research Projects

The area of research of ANM is very large and it is useful to provide an overview of some of the

'umbrella' research projects which are being currently undertaken in the field of electrical power

systems. A wealth of reports on the subject area of ANM are available from these organisations

which are detailed below. Their objectives are best illustrated by directly quoting the projects

mission statements.

Autonomous Regional Active – Network Management System (Aura-NMS)

Aims :

“The 3-year project aims to demonstrate the benefits of integrating network

management & energy storage technologies onto the networks.”[34]

SuperGen

This project aims to :

“help the UK meet its environmental emissions targets through a radical

improvement in the sustainability of power generation and supply.”[35]

Asset Management and Performance of Energy Systems (AMPerES)

This project which is part of 'SuperGen' has a mission to :

“provide the tools to enable reliability of energy supply at minimum cost

in the context of ageing plant, a drive to deploy renewable and distributed

generation and an environmental imperative.”[35]

Electricity Networks Strategy Group (ENSG)

Aim :

“The aim of the ENSG is to identify, and co-ordinate work to address

the technical, commercial, regulatory and other issues that affect the

transition of electricity transmission and distribution networks to a

low-carbon future.”

One of the outputs of the work done by this organisation which is supported by the DTI and

OFGEM is a 'register' of projects and activities for Active Management which includes over 100 of

which are in the UK [26]. The most mature of these are in the area of distributed coordinated

voltage control and active power flow management [22].

Other Research Programs

FENIX [36], FutureNET and FlexNET are three other programs although the last two are within the

SuperGEN project mentioned above.

5 – Current Research

Page 31: Connecting Renewable Generation with Active Network Management

5.2 Technology Research Areas

5.2.1 Intelligent Generator Control

Automatic Voltage Power Factor Control (AVPFC) and Fuzzy Logic Power Factor Control (FLPFC)

are two methods of generator excitation control which seek to maintain better control over voltage at

the ends of rural networks for the purpose of allowing increased DG connection. Most DNO's

require that synchronous generators connected on their network use constant PFC generator control.

This enables easy control and administration of the network. The AVPFC hybrid control system

proposed by Edinburgh University upon reaching local voltage limits transitions from a constant

PFC mode into a Voltage Control (VC) mode hence increasing or decreasing the production of

VARs to control the local voltage at the generator connection point as required to keep within

network limits. Where the VC mode might conflict with local voltage regulation an alternative

modified PFC method is proposed using Fuzzy Logic called FLPFC. This control method

dynamically adjusts the PFC setpoint to give an optimum PFC setting and can cope better with

varying network conditions as might be found when other DG generators are connected [13].

5.2.2 Distributed versus centralised control of voltage

Edinburgh University compares the use of generator and other VAr control methods for the task of

maximising DG penetration. Co-ordination of these VAr controls gives similarly good results when

performing the OPF though centralised or distributed methods. Note that losses were found to be

an issue[14].

5.2.3 Active Power Flow Management (APFM)

Strathclyde University's work on the Scottish Orkney Islands RPZ is an example of APFM on trial

in a network constrained by its undersea cable links to the mainland. Cable resistance and

capacitance being a dominant factor here as opposed to long overhead radial feeders where where

the effect of reactance dominates.

VAr control to help manage voltage levels is already exercised by a dynamic reactive compensation

device (DVAR) located at the village of Scorradale. See 6.3.4 for more information on these

devices.

The existing generation capacity on the island is divided into Firm Generation (FG) and Non-Firm

Generation (NFG). This comprises a mixture of Wind, Marine (test sites) and Gas. This division of

capacity in needed as a result of the N-1 contingency condition if one of the two sub-sea cables

which feed the island is lost.

These are known as 'pre-fault constraints'. Hence network capacity in a conventional sense has

already been reached, the NFG subject to a hardwired 'direct intertrip' on the First Circuit Outage

(FCO) for this contingency. These types of connections are already offered by some DNO's.

Strathclyde University detail a method to introduce New Non-Firm Generation (NNFG) by using

output regulation (trimming) and tripping to enable the dominant thermal limits of the mainland

cable to be overcome and hence obtain more DG capacity.

Essentially a zoned area control system is introduced with attendant sensors and communications

which firstly regulates output of NNFG and if insufficient then trips the units. A Last In First Out

(LIFO) buffer is used as the method of selection with the generators ranked according to priority.

Aside from possible Power Quality (PQ) issues from switching a main issue for generator operators

5 – Current Research

Page 32: Connecting Renewable Generation with Active Network Management

would be loss of revenue from curtailment. Strathclyde University present an analysis on the amount

of curtailment using a simulated data time series [15] (I.e same as chapter 7 in this report).

The Orkney APFM control logic is illustrated below. In this figure the acronym RNFG means

Regulated Non Firm Generation [16].

Fig 5.1 – Orkney APFM control logic taken from [15]

5.2.4 Network Capacity Planning and OPF

Due to the interconnected nature of the distribution network, placement of DG resources can

become a strategic issue with effective 'sterilisation' of the network to new connections occurring.

Edinburgh University has undertaken a number of studies regarding how placement and control of

DG effects this overall DG penetration level. Time series power flow analysis for intermittent DG,

Fault Level Constraint analysis, effect of commercial incentives are all considered. Multi-objective

OPF has been investigated also [17],[18],[19]. A script driven simulation suite using PSS/E is the

basis for the simulations [20].

5 – Current Research

Page 33: Connecting Renewable Generation with Active Network Management

5.2.5 Multi-agent Systems (MAS)

MAS is an area of computer science which is being applied to the problem domain of ANM.

Fundamental Concepts

What is a MAS? In [21] Wooldridge gives a basic definition of an agent as

“ merely 'a software (or hardware) entity that is situated in some environment and is

able to autonomously react to changes in that environment.' “

In the context of MAS 'intelligent' agents are used which have the characteristics of Reactivity, Pro-

activeness and Social ability. These key abilities distinguishing a MAS system from traditional

hardware software systems. Fundamentally a MAS is a collection of two or more of these agents.

Communication is implicit between agents but not necessarily directly although Power Engineering

applications are thought to warrant these.

The application of MAS to PE applications is useful as MAS exhibit the qualities of robustness,

flexibility and extensibility. Four areas are identified. Protection, Modelling and Simulation,

Distributed Control, and finally Monitoring and Diagnostics. Local management of DNs using

MAS is put forward as an alternative solution to problem of increasingly complex centralised

control.[21]

AuRA-NMS and use of MAS

The AuRA-NMS project (see section 5.1) proposes the use of MAS to implement a network

automation system. The difference from previous network automation proposals for MAS being that

“it does not map an agent architecture onto the existing topology of the

power system, i.e. in AuRA-NMS agents do not represent specific generators,

circuit breakers, feeders or busbars: MAS technology is used simply as a

flexible, extensible,distributable software integration framework.”[22]

The complimentary benefits that this new type of control system are hoped to bring are due to it's

ability to address a broad range of issues regarding DN operation. Three main issues are identified

as the following.

Network Access and Connection Agreements. This derives from increasing demands for

DG. Connection of DG can alter control schemes on the DN and thus flexibility and

adaptability of these controls are key to enabling increased connections.

Complexity of control schemes. Layering of control schemes as networks develop can have

the unfortunate effect of making interactions more complicated and root causes of

operational problems difficult to determine. Thus risking adverse network performance with

attendant penalties.

Network Performance. DG connection incentives aside DNOs are obliged to meet

performance targets such as minimising Customer Minutes Lost (CML), Customer

Interruptions (CI) and also minimising losses.

AuRA-NMS see all of the above issues as being ideally addressed by the use of MAS.[22]

5 – Current Research

Page 34: Connecting Renewable Generation with Active Network Management

DNO requirements for a Network Management System

Six items are identified.

Safety and Security. This is considered paramount.

Flexibility and extensibility. Easy reconfiguration for plant rating changes, new

connections, disconnections, changes to measurement and monitoring etc. Coping with

addition of new functionality or improved algorithms etc.

Tolerance of Failure. e.g To communications, hardware, software, plant, EMI.

Graceful Degradation. Maintaining fulfilment of important goals e.g keeping supply on.

Integration with the existing DMS. The DNO's Control centre must still have ability to

override any automation.

Interfacing with existing equipment. Existing and future measurement and control

equipment should be easily connectable. IEC 61850 is identified as a suitable standard to

comply with. [22]

Implementation for AuRA-NMS

AuRA-NMS implements network control and management using a distributed hardware platform.

Management and control is done by 'agents' which would typically be located at substations and

would communicate where necessary with software running in other locations. The hardware

platform selected is the ABB COM600 which already has applications in the substation automation

field. IEC 61850 is supported by this device in addition to some other common protocols.

Functionality

This has been basically split into two functions .'Reactive' control being a short term response for

example to faults and 'Pro-Active' being a longer term pre-emptive response to for example goals to

minimise network losses in response to load forecasts.

It's core functions in initial deployment are stated as [22]

� Management of steady state voltage

� Automated restoration

� Operation of the network within thermal limits e.g power flow management

� Management of constrained connections

� Proactive network optimisation strategies e.g minimisation of losses

A key aspect of the functioning of the system is to use “selectively devolved goal driven network

control”. Obviously communication to the control engineer of the 'goal' which the agent is pursuing

would be essential for understanding what the ANM system was doing in an operational context.

[22]

5 – Current Research

Page 35: Connecting Renewable Generation with Active Network Management

Fig 5.2 - Example layout of network containing ANM platforms taken from [22]

MAS Development Standards

Some industry standards and tools are worth mentioning in relation to, or which might be applied to

the AuRA-NMS system. These are briefly [22]

a) Foundation for Intelligent Agents (FIPA). This is an agent management reference model as

shown below. The Agent Management System (AMS) handles the agents when they connect

or disconnect. The Directory Facilitator (DF) registers the services which the agents can

offer.

Fig 5 – FIPA agent management reference model taken from [22]

b) Java Agent Development Environment (JADE). A software platform for agent

development.

c) Common Information Model (CIM). This is a data model which can be used to capture

details of a network

5 – Current Research

Page 36: Connecting Renewable Generation with Active Network Management

6 – Connection Solutions and Implementation

6.1.1 Overview

For the basic connection problems identified in chapter 4 a brief discussion of connection solutions

is given here. From these basic connection problems voltage control is chosen and specific

implementation issues for three of these voltage control techniques is detailed.

6.1.2 Estimating Connection Limits

Rule of thumb connection limits for passive connections to the network are shown in table 6.1

below. If larger capacities are required relevant system studies need to be done to ascertain the

actual connection capacity available. However any 'passive' connection technique will use

conservative values. This is known in the industry as the 'fit and forget' approach.

Network Location Maximum Capacity of

Embedded Generator

On 400 V 50 kVA

At 400 V busbars 200-250 kVA

On 11kV 2-3 MVA

At 11kV busbars 8 MVA

On 15kV/ 20kV or at busbars 6.5-10 MVA

On 63kV/90kV 10-40 MVA

Table 6.1 – Indicative design rules used DG connection limits [11,pg.13]

The purpose ANM connection techniques is to increase DG connection capacities beyond these

existing connection limitations without incurring large capital outlays for plant upgrades.

In terms of network design some 'rules of thumb' exist in terms of optimum location of DG on the

network. One such guideline is the '2/3 rule' which states that on a radial circuit a DG of 2/3rds of

the line load situated at 2/3rds of the line length gives optimum reduced losses. However for DG

such as wind and hydro the primary plant location is determined by the geographics of the energy

resource.

6.2 Solutions

This section is based on [11] unless otherwise referenced which is based on survey data from 8

DNO companies.

Solutions to DG connection problems can be broadly divided into three categories :

� Voltage control

� Power flow management

� Fault Level management

6 - Connection Solutions and Implementation

Page 37: Connecting Renewable Generation with Active Network Management

For the purposes of this report only the first category voltage control will be dealt with in depth as

this is directly relevant to the case study in the later chapter and is commonly a key limitation in

determining connection capacity. Although many of the above issues may be interrelated.

Power flow management is outlined as a subject area in chapter 5. These three problem domains can

be addressed with the solutions shown in table 6.2 below.

Category Solution

Conventional (C) ANM (A)

Problem

Voltage Control Line re-conductoring or upgrade(C)

New dedicated line (C)

Generator Real/Reactive power control (A)

Line Voltage regulator install (C/A)

Improved OLTC Control (C/A)

Area Based OLTC Control (A)

Voltages must be kept within

statutory limits.

Power Flow

Management

Network Re-inforcement (C)

Pre-Fault Constraint (C)

Post-Fault Constraint (A)

Post-Fault Constraint (dynamic rating)(A)

During circuit outage conditions

DG can cause power-flows to be

exceeded.

Fault Level

Management

Switch gear upgrade (C)

Increase Impedance (C/A)

Generator Convertor Improvement (C/A)

Network Re-configuration (C/A)

Operational procedures (C/A)

Under transient fault conditions

lack of fault current or increased

fault current from DG can cause

mal-operation or violation of

equipment ratings.

Table 6.2 – Broad categorisation of DG connection problems and solutions

6.2.1 Voltage Control Techniques

As explained in chapter X the voltage difference between two points on a circuit is chiefly

determined by the resistance, reactance, real and reactive power flows. Hence control of these

parameters either at the planning stage or in real-time are the basis of the methods along with

exploitation of existing transformer OLTC's controls.

Line re-conductoring

An increase in the cross-sectional area of conductors or the addition of a second circuit result in a

reduction in resistance and hence a lower voltage gradient on the circuit.

Upgrading of the line to a higher voltage may also result in lower voltage drops as less current is

needed to carry the same power.

New Dedicated Line

This presents the opportunity to use appropriately sized conductors and also to connect directly to a

substation at a suitable voltage. Operation with voltages outside of statutory limits at the DG end of

the line is a possibility if no customer loads are to be present on this new line.

6 - Connection Solutions and Implementation

Page 38: Connecting Renewable Generation with Active Network Management

Generator Reactive Power Control

Reactive power control is employed on the transmission network the high X/R ratio of long lines

making it particularly effective. If allowed by a DNO control of reactive power at a DG site or a

common point of connection for a group of DG's may allow some flattening of the voltage on the

network to the benefit of DG penetration. A number of different methods exist for control of PF

including switched capacitor banks and SVC's.

Generator Real Power Control

Ability to reduce a generators real output will reduce voltage rise as explained in chapter 4. Since

typical capacity factors for wind farms are of 40% or less then some loss of energy revenues in

periods where high wind and low local demand exist may be financially feasible for a corresponding

general increase in output at other times. This 'curtailment' of real power is used in the case study

later in this report.

Line Voltage Regulator/Booster

Modification of voltage on a feeder by a dedicated transformer can achieve improvements in voltage

profile at problem locations. These transformers have essentially 1:1 ratio windings with seasonal

adjustment of tappings or OLTC control enabling more accurate dynamic regulation of the voltage.

A single stepped control would enable coarse control. This single step control is known as a booster

type regulator [3,pg.212]. In the case example in this report a booster transformer is used rather to

'suppress' the voltage on the problem DG Bus. The improvement effect of an IVR on a radial circuit

with DG connected at is shown in figure X below. (DG located at 10km and IVR at 5km)

Fig. 6.1 – Example of DG on Network voltage profile with and without an IVR [23,pg.18]

LDC Cancellation CT

Transformers using LDC control can suffer from inaccurate voltage control when the DG on a

feeder is large compared to the overall transformer throughput. Varying DG output can cause the

effective power factor on the bus to change.

A cancellation CT device can exclude the contribution of the DG feeder to the LDC control input

and remove the inaccuracy thus improving the voltage control and presumably the available DG

headroom.

6 - Connection Solutions and Implementation

Page 39: Connecting Renewable Generation with Active Network Management

Advanced OLTC Schemes

See section X commercially available products.

Area Based Voltage Control

This method uses a single or a number of distributed controllers to effect ideal OLTC control on the

network for a single or multiple number of transformers. A mixture of remote measurement and

estimation of voltages based on models using known network parameters results in a good

approximation of real voltages on the network. Thus an optimum combination of control settings

can be used to maximise DG penetration.

6.2.2 Comparison

Note that any of the solutions detailed here could be incrementally applied in any combination an

example is given below.

Fig. 6.2 – Incremental application of voltage control solutions [102.60]

The below table gives a summary of connection issues for the above solutions for the voltage control

problem. Costings are taken from [11,pg.69] the most common response in the survey being taken

as the most realistic cost figure.

6 - Connection Solutions and Implementation

Page 40: Connecting Renewable Generation with Active Network Management

Solution Advantages Disadvantages Stakeholder Implications Cost (Benchmark) Comments

DG Operator DNO Customer

Line Re-

conductoring

Improves P.Q

Reduces CML & CI

Is a Standard Solution

Expensive Deep connection charging

possible

Lower Losses Possible disruption

for installation

(11kV) < £20k/km

(33kV) £20k-50k/km

(132kV) £50k-200k/km

Cheaper if done

during scheduled

asset replacement

New Line Same as above Very expensive

Planning consultation

Same as above.

Significant increase in capacity

Extra connection capacity

available for other parties.

Fault level increase.

Negligible (11kV) £20k-50k/km

(33kV) £50k-200k/km

(132kV) £+200k/km

Connection to HV

bus would be

preferred DNO

solution.

Generator

Reactive Power

Control

General increased

VAr flows & losses

Operational costs for importing

VArs

Review of the network

voltage control scheme

needed

Negligible (11kV) £20k-50k

(33kV) £50k-200k

(132kV) N/A

Generator Real

Power Control

Large plant investment

not required

Change in Control system

Some Operational cost &

complexity.

Reduced connection cost

Negligible NA Good opportunity in

future if combined

with storage

Line Regulator Large increase in

capacity

Deep connection charging

possible

Extra losses

O & M costs

May affect protection

schemes

Installation

disruption.

Increased component

count reducing

reliability

£40k for a 2MW wind

farm example given

[11,pg.37]

LDC

cancellation

CT's

Improves exisitng

OLTC control

Good Value Substation work.

Need voltage control re-

evaluation of Network

Negligible NA

Advanced

OLTC

Schemes

Similar to above NA

Area Based

OLTC schemes

Maximises use of

existing network assets

Added complexity Monitoring & control hardware

needs installing

Monitoring hardware

O&M

May need on-site

monitoring

equipment

£20k-50k

Table 6.3 – Summary of connection solution issues

Page 41: Connecting Renewable Generation with Active Network Management

6.2.3 Implementation

This section is based on [23]. Important information is highlighted by specific page references.

Other sources of information are referenced appropriately.

6.3 Products available or in development

6.3.1 Dynamic Line Ratings

These systems are based on the fact that lines and transformers are rated for worst case thermal

limits. Hence it is possible to operate some assets at higher capacity when thermal conditions are

not at their most severe. A particularly good example of this is overhead lines which are connected

to wind generation. In windy conditions cooling of the lines is improved and hence higher currents

can be achieved. Some form of measurement of temperature or for example line sag is used which is

then sent as a signal to indicate extra available capacity and reduction of a curtailment constraint.

6.3.2 Area Based Voltage Controllers

These are sometimes referred to as micro-grid controllers [23,pg.11]. There purpose is to effect

optimum control of voltages across the network rather than at just one bus bar which would be the

function of a conventional AVC at a substation. An example is the GenAVC by Econnect which is

shown in the figure below controlling the OLTC at a primary DN substation. A combination of

remote measurement and 'state estimation' is used to model the the operating scenarios of the

network. See [24] for a more in depth overview of development and functioning of this controller.

Fig. 6.3 – GenAVC active voltage controller [24,pg.6]

An important consideration with any new type of control on the network is its integration and

interaction with existing control or communication systems. DNOs use SCADA based management

systems sometimes known as Distribution Management Systems (DMS). These have limited control

and measurement in the DN most of it confined to primary substations (Secondary substations are

typically the last I.e. 11kV/400V). Some automation does exist the purpose being to improve

restoration of faults. Figure below shows a typical DN support system.

6 - Connection Solutions and Implementation

Page 42: Connecting Renewable Generation with Active Network Management

Fig. 6.4 – Typical DMS [25,pg.26]

Centralised Versus Distributed Control

Since this is a developing area of technology some consideration of choice of topology is useful.

The main choice being using centralised control at the the already existing SCADA level or using

devolved control at the substation level. Important considerations are as follows

� Controller configuration after changes to network

Distribution level network changes are relatively frequent and any distributed ANM

controller would likely need to be updated by maintenance staff. Information regarding these

network changes are known at the centralised level for the purposes of safety and

coordination hence modification of ANM control schemes would seem easier if carried out

at this higher level too.

� Product standards

For ease of maintenance and simplification of working procedures standardised hardware,

software and protocols may make commercial sense in the long-term.

� Installation time

DMS are large and where ANM control might only be useful in a few isolated locations a

distributed solution may be more readily achievable in terms of development time-scales.

6 - Connection Solutions and Implementation

Page 43: Connecting Renewable Generation with Active Network Management

� Communication requirements

Remote installations may benefit from distributed controllers due to savings in

communication infrastructure. On the other hand more advanced automation of the network

would require some data communication so a comprehensive centralised system could have

overall benefits in terms of data management and overall maintenance of the system.

Another issue is response time for control loops, localised control obviously offering

immediate response for control actions such as voltage regulation.

In general reliability of communications for both topologies of control are essential for

minimising operation of equipment in fail-safe mode (e.g would typically mean increased

curtailment for DG).

6.3.4 SVCs and STATCOMs

These modern shunt connected devices use a power electronics converter or inductors and

capacitors or combination of to locally absorb or inject reactive power. Hence local control of

voltage and power flow is possible with possible applications for improvement of stabilisation and

harmonic filtering. The Shunt Variable Compensator (SVC) is currently most widespread[8,pg.383].

However their use is almost exclusively on the transmission network and even then not that

common. Some installations are at industrial sites to improve power quality issues for large or

sensitive industrial users.

An example 'converter' based variable compensator is included in the appendices. These have very

fast operating times so are often termed 'dynamic'.

[23,pg.22] defines STATCOMS as using the newer IGBT technology whilst SVCs use thyristors.

Obviously these devices can form part of an ANM system if given suitable localised control

function and communications.

Note the PEIs on these devices share some similarities in construction and function as PEIs on some

Wind Generator DFIG/Synchronous units.

6.3.5 Commercially Available Plant

Plant may be generally classified in to primary plant consisting of large pieces of hardware such as

In-line Voltage Regulators SVC's etc and secondary plant consisting of small pieces of hardware

such as micro-grid controllers.

A number of products are now currently available or nearing commercial status. An excellent source

for information regarding the development of ANM technology is the ANM Register [26]. Table 6.4

below shows items which are available or are though to be near commercial production.

6 - Connection Solutions and Implementation

Page 44: Connecting Renewable Generation with Active Network Management

Product/Item Type Stage

Cancellation CT - VATech Cancellation CT Commercial

MicrTAPP - VATech Regulation for remote side of

transformer

Commercial

Gen-AVC - Econnect Area Based OLTC Preliminary Commercial Trial

ENMAC DPA(Distribution

Power Analysis) module – GE

Energy

Area Based OLTC and general DN

operation SCADA tool

Commercial

Power Donut2- USi-Power Dynamic Line Rating Commercial

D-VAR – American

Superconductors

D-STATCOM – Areva

(+/-10MVAr)

MINICOMP – ABB (up to

+/-20MVAr)

Distribution Network sized

STATCOM

Distribution Network sized

STATCOM

SVC

Commercial

(Used on Orkney for sub-sea Cable

Volt rise mitigation)

Commercial Development

Commercial

APFM Power flow management Research - Trial of APFM on

Orkney for renewables capacity

constraint problems

COM600 - ABB Research - Substation Industrial

computer platform used by

Strathclyde University research into

use of MAS for ANM. Part of Aura-

NMS project.

Table 6.4 – Commercialised (or near commercialised) ANM technologies

Note some solutions are considered 'enablers' rather than direct solutions. e.g fault limiters act to

control faults during transitory conditions which is the responsibility of distribution automation.

However these devices can provide solutions to DG connection fault issues.

6.4 Specific Implementation Issues

The following three technologies reviewed in table 6.5 are deemed “largely available now but for

various reasons are not being used by the DNOs” [23,pg.14]. These provide a revealing insight into

practical integration issues in what is considered to be a very conservative industry regarding uptake

of new technologies. Note these technologies are present on a number of RPZ schemes. Dynamic

Line Ratings involve optimisation of thermal constraints so were not included.

6 - Connection Solutions and Implementation

Page 45: Connecting Renewable Generation with Active Network Management

Technology Technical Operational Planning Other

In-line Voltage Regulators

(An example trial is

present on a North Wales

DG

installation using a Cooper

Power

Systems IVR [23,pg.18])

Safety & Environment : Same as conventional transformers

(Fire/Oil etc)

Functionality : Might be used as last case solution if AVC

optimisation/LDC/cancellation CT's aren't sufficient. Possible

low voltage downstream if DG trips off line.

System Modelling : Increased time required by DNO to

include these unfamiliar IVRs into system models.

Communication : Could be stand-alone or communicate with

main substation and/or Area based voltage controller.

Very similar to

conventional

transformers and

OLTC's in terms of

maintenance.

Novel method requires

extra time in during

initial voltage control

studies.

Reliability & Consequence of Failure : Similar

high reliability as of normal transformers.

Fitting of by-pass switches can provide

isolation of the IVR without loss of the circuit.

Adaptability : Large voltage variation between

each side of the IVR and sensitivity to tapping

range require careful studies to ensure the asset

doe s not become 'stranded' after future

changes on the network.

Costs: Autotransformer construction means

low cost compared to normal transformers.

Extra installation of IVR's on

the UK network may not

qualify for RPZ technology

funding unless combined with

other innovative technologies.

Synchronous Compensators

-

SVCs & STATCOMs

(D-VAr on Orkney is

example of UK use on a

DN)

Safety & Environment : Standard issues as with similar oil

filled equipment and/or Power Electronic installations.

Functionality : Noted for ability to solve general problems on a

network of PQ stability etc.

Communications : Standard alarms required. Could function

stand-alone or as part of area coordinated approach.

Complex but might be

considered as 'black-

boxes'.

Reliability & Consequence of Failure :

Shunt connection implies relatively safe failure

scenario.

Future Proofing : Shunt connection and

possible scalability may make movement of

the asset practicable.

Cost : Distribution sized STATCOMs are a

relatively new product so typical costs are

unknown.

May qualify for IFI/RPZ

funding as Orkney installation

was a subsea cable volt-rise

problem rather than a DG

connection issue.

Active Voltage Controllers

(Example Martham RPZ

trial of Econnects GenAVC)

Safety & Environment : Voltage damage to consumer

equipment through mal-operation of voltage control is a prime

concern although this risk is considered not arguably greater

than existing passive control risks.

Functionality : Recognised as providing effective levels of

increased DG connection. Integration with other ANM

methods may be beneficial.

Centralised versus Distributed Control : See separate section.

Use of standard

hardware and

communications systems

should ensure

integration. Changes to

configuration of the

network present concern

in terms of making

controller changes. With

care CML & CI should

not be negatively

impacted, could even be

opportunity for an

improvement.

Reliability & Consequence of Failure :

Incorrect setup or modelling, feedback errors

and unexpected network events might cause a

voltage excursion. Due to complexity open-

loop in situe testing has been done on the two

existing ANM controllers. This has shown

successful operation.[23,pg.30]

Cost : Initial ANM controller implementations

will force DNO changes to operational

procedures and SCADA systems. This has

implications but the savings on negating line

upgrades are thought significant.

Other manufacturers ANM

controllers or DNOs might be

eligible for IFI/RPZ funding.

Table 6.5 – Appraisal of specific implementation issues

Page 46: Connecting Renewable Generation with Active Network Management

7 – Simulation

7.1 Overview

In order to illustrate implementation of ANM a simple case study is undertaken using the chosen

software tool PowerWorld as evaluated in chapter 2.

In order to make the project simpler and also to provide results which can be verified the network

case scenario is taken from a published book [27,ch.21] a copy of which is included in appendix

A2.1. Note that this network case study also appears in a number of journal papers [28] and a report

[29]

Two types of analysis are undertaken for each of the four ANM scenarios applied to the network.

Firstly steady state operating characteristics are determined for all four combinations of maximum-

minimum loads and WF outputs. Secondly a time series analysis is run for each of the four ANM

cases which simulates hourly operation over a year.

Fig 7.1 – Network case scenario taken from book [27,pg.467]

7.2 Description of the Network

This is detailed in [27,pg.466] as :

� 33 kV distribution network fed from a GSP at 132kV

� Busbar 2 load “represents the aggregated loads of the remaining part of the system”.

� Loads represent mixture of commercial/residential and industrial

� Distributed wind generation on bus 6 with power factor correction

7 - Simulation

Page 47: Connecting Renewable Generation with Active Network Management

The network is understood to be a mixed urban rural circuit of unknown line lengths and an

unknown mixture of cable or overhead line constructions. Since no loads exist on Circuit 5 to 6 this

would be called an extension which would be built specifically to facilitate the Wind Farm

construction.

In order to obtain the real and reactive losses and bus angles as shown presented in the book (Fig.

7.1) the reactance and resistance columns in the branch parameters table [27,pg.467] were swapped

and used as below. Simulation of the network using the values as presented in the book results in

disproportionately high losses in the GSP transformer and yearly circuit losses nearly ten times

greater than that presented in the book. This circuit is included in the appendices A1.7 for reference.

Correct operation of the circuit with the table values swapped is also confirmed by the annual

losses with DG of 3363MWh (see chapter 8) which compares well to results presented in the book.

From – To Reactance (p.u) Resistance (p.u) X/R

ratio

Rating (MVA)

1-2 0.17726 0.01869 9.48 60

2-3 0.20174 0.15500 1.30 30

3-4 0.15770 0.09000 1.75 30

2-5 0.30000 0.40000 0.75 30

5-6 Line 0.40000 0.35700 1.120 30

5-6 Booster

(see section X)

0.72600 0.40000 1.815 30

Table 7.1- Circuit Parameters (On 100 MVA base)

Note : MVA ratings are estimated from the 'analogue gauges' shown on figure 7.1.

Bus Loads (MW) Max / Min Loads (MVAr) Max / Min

3 3.60 / 0.70 1.44 / 0.28

4 9.10 / 1.80 3.64 / 0.72

(Feeder A) 12.7 / 2.5 5.08 / 1.00

5 (Feeder B) 3.30 / 0.60 1.32 / 0.24

2 (Feeder C) 31.50 / 6.30 12.60 / 2.52

Total 47.50 / 9.40 19.00 / 3.76

Table 7.2 – Maximum & minimum real and reactive loads

Maximum and minimum loads are also taken from figure 7.1 and are shown in table 7.2 above. The

real and reactive ratios equating to load power factors of 0.98 lagging and also a minimum load to

maximum load ratio of 0.4.

Bus voltage limits of +/- 3% are used the same as that stated for use in the book analysis.

7 - Simulation

Page 48: Connecting Renewable Generation with Active Network Management

7.3.1 Steady State Circuit Analysis

Figure 7.2 shows the PowerWorld one-line diagram for the network case scenario. The main

difference between this one-line diagram and that presented in the book version (Fig. 7.1) are the

'dump' loads included for modelling of curtailment and a dump load 'Load Scale Factor' control for

their easy adjustment.

On the one-line diagram the numbers above the circuit lines are MW and MVAr power flows along

the line. The number s below are MW and MVAr losses in the line. Bus voltages are indicated in

per-unit and also the angle with respect to the GSP Bus which is taken as the reference Bus. This

bus also known as a 'Slack' Bus provides the balancing power in the system. The values above the

pie charts are apparent power pie-charts indicating the percentage of the MVA peak rating of the

line which is being used.

Basic circuit characteristics

The notable elements of this circuit are that a large load of approx 9MW is located on Bus 4 and the

DG is located on a 'line extension' to bus5 where a relatively small load of approximately 3MW is

located. Also there a essentially three radial feeders having their voltage controlled by the OLTC

GSP transformer.

The network can be easily characterised by how it behaves under the following combinations of load

and generation. Worst case conditions will normally coincide with one of the below extremes of

operation :-

� Maximum load and maximum generation

� Maximum load and minimum generation

� Minimum load and maximum generation

� Minimum load and minimum generation

Examining bus voltages for the basic passive voltage control method I.e holding Bus 2 at a constant

1.0 Vpu under the four conditions shown above are shown in table X. It can be seen that Bus 4

suffers from being on the limit of under voltage at 0.971 pu under conditions of maximum load. The

limitation on DG size occurs under this maximum load condition also.

Under increasing amounts of wind generation the power flow through line 2-5 reduces and

consequently the voltage on bus 5 rises. Eventually after the bus 5 load is supplied entirely from the

WF and power is exported along line 2-5 back into the rest of the network the voltage on bus 5 rises

above that of bus 2.

Hence the circuit voltage profile under conditions of DG output takes on a characteristic 'K' shape.

Note that this explanation is true when the power flows are largely real in nature. Large reactive

power flows may modify this as explained in chapter 4.

With the network as presented peak DG outputs of 6MW under maximum load and 6MW under

minimum load are possible. The minimum load maximum generation case should arguably be

nearer to 10MW as presented in the book. However OLTC transformer settings used in the

simulation here result in the maximum DG penetration being lower in this example simulation due

to variation in Bus 2 tolerances. This is discussed later.

7 - Simulation

Page 49: Connecting Renewable Generation with Active Network Management

Fig. 7.2 – PowerWorld one-line diagram of circuit as presented in book

Page 50: Connecting Renewable Generation with Active Network Management

7.3.2 Applying ANM to the network

The previous basic analysis allows the specification of a number of basic ANM solutions to try to

increase DG capacity on this particular network. The following ANM techniques are applied to the

circuit to ascertain their effectiveness.

� Real power flow management – Curtailment

� Reactive power flow management

� Area based Voltage Control

� Use of Voltage Regulator/Booster

These were easily modelled and show application of technologies which are at a commercial stage.

Although included in analysis in the book difficulties in simulating the use of SVC and STATCOM

devices resulted in these being left out of the analysis.

Optimisation of OLTC voltage setting

Note that as an improvement for the two passive voltage control case scenarios the OLTC control of

bus 2 was set to use 1.005Vpu as a target instead of 1.000Vpu. This brings all of the buses well

inside the limits of +/- 3% under the worst case scenario which is of maximum load. This actually

has a slightly negative effect on DG penetration as it reduces the voltage headroom available for Bus

6.

Real Power Curtailment

This is implemented as explained in section 7.4.2. All of the simulations presented depend on

curtailment of real power to prevent over-voltage on the DG Bus 6. Thermal limits are not

encountered.

Reactive Power Management

Operation of bus 6 at a reduced power factor of 0.95 consuming instead of 0.98 consuming. This is

effected by changing the generator capability curve in the relevant PW setup dialogue. See section

7.4.5 for an explanation of this.

Area based OLTC (LDC for bus2)

This control method involves measuring or estimating voltages at various locations on the network

and controlling the OLTC to keep all voltages within limits whilst maximising DG penetration.

For this network 'area' based voltage control was not thought necessary in that use of LDC control

for bus 4 seemed an optimal voltage control solution without measuring or controlling any other

circuit elements. Note also that PW may only be able to model Area Based Voltage control if its

simulation automation package was added which is outside of the scope of this project.

Line Drop Compensation is method of OLTC control which adjusts the local bus voltage setting to

take into account increased voltage drops in radial lines serving other loads when under heavy

7 - Simulation

Page 51: Connecting Renewable Generation with Active Network Management

loads. Hence for this example bus 2 will have its voltage automatically raised during heavy loads

which will raise the bus 4 voltage also. This control method is also known as current compounding.

See [23,pg.257] or [3,pg.147] for a good explanation of LDC control.

Voltage Regulator/Booster

Placement of a transformer with tap changing capability on the line from the WF provides the

ability to reduce the voltage on Bus 6 under times of high output. This allows the Bus 6 voltage to

track that of Bus 5 hence allowing headroom for more DG generation. The limiting factor on output

is now the voltage rise due to the impedance of line 2-5 only. Note that thermal constraints may

become a problem at outputs beyond 20MW for lines 2-5 and 5-6.

For modelling simplicity a voltage 'Booster' rather than regulator is used. This is essentially a

transformer with single tapping which gives a 1:1 ratio and a single boosted ratio. In terms of setup

a caveat exists in that the tap ratio for the booster needs to be so that the voltage on Bus 5 does not

go above Bus 6. This is achieved by using a tap ratio of 1.025 is instead of an ideal setting of 1.035

which could be used with a voltage regulator. This reduces the available increased headroom

slightly compared to the case of a voltage regulator under the minimum load maximum generation

combination. See Fig. A1.14 in appendices for illustration.

For simplicity the increased reactance due to the insertion of the voltage regulator is lumped in with

the reactance of the line 5-6. The resistance of the regulator is assumed to be negligible. The new

reactance of line 5-6 is calculated as below :

System Base 100MVA

Regulator Reactance is taken as 0.1pu on 30MVA

Zpu (new) = Zpu(given) x Sb(chosen) / Sb(given)

Zpu of line 5-6 on its original base of 30MVA is :

Zpu (new) = 0.4x30/100 = 0.12

Total Zpu of combined line and transformer = 0.1 + 0.12 = 0.22

Xpu of combined line and transformer on 100MVA = 3.3 x 0.22 = 0.726

The four case scenarios for the simulation study are shown below in figures 7.3 to 7.6

7 - Simulation

Page 52: Connecting Renewable Generation with Active Network Management

Fig. 7.3 – Simulation case Passive voltage control Fig. 7.4 – Simulation case Passive voltage control with Bus 6 PF 0.95

Fig. 7.5 – Simulation case Load Drop Compensation Fig. 7.6 – Simulation case Voltage Booster & Load Drop Compensation

OLTC

Transformer

Bus 2

Bus 1 - GSP

132 kV

Bus 4Bus 3

33 kV

Bus 6Bus 5

DG

Load

PF 0.98 Consuming

OLTC

Transformer

Bus 2

Bus 1 - GSP

132 kV

Bus 4Bus 3

33 kV

Bus 6Bus 5

DG

Load

PF 0.95 Consuming

OLTC

Transformer

Bus 2

Bus 1 - GSP

132 kV

Bus 4Bus 3

33 kV

Bus 6Bus 5

DG

Load

PF 0.98 Consuming

OLTC

Transformer

Bus 2

Bus 1 - GSP

132 kV

Bus 4Bus 3

33 kV

Bus 6Bus 5

DG

Load

PF 0.98 Consuming

Current Transducer

Voltage Booster

Page 53: Connecting Renewable Generation with Active Network Management

General simulation settings for the four cases are set out in the table 7.3 below.

Case OLTC Bus / Target PF Other settings

Original (Passive) Bus 2 1.005 +/-

0.005pu

0.98 Lag OLTC (target middle)

PF 0.95 “ “ 0.95 Lag “ “

LDC Bus 4 0.970 +/-

0.005pu

0.98 Lag “ “

Voltage Booster &

LDC

“ “ “ “ Booster Bus 6 / 1.000 / 1.025

(max/min type target)

All cases (except

Booster)

Tap step 0.00625 pu

Max 1.1 Min 0.9 pu

(middle type target)

Table 7.3 – Summary of PW settings for cases

7.3.3 Steady State Simulation Results

Case Load/Gen

min/max?

Bus 2

(V p.u)

Bus 3

(V p.u)

Bus 4

(V p.u)

Bus 5

(V p.u)

Bus 6

(V p.u)

Maximum

WF Size

(MW)

Passive Voltage

Control at PF 0.98

L-Min/G-Min 1.002 0.992 0.988 0.996 0.996 -

L-Min/G-Max 1.001 0.991 0.987 1.015 1.031 6

L-Max/G-Min 1.009 0.983 0.971 0.993 0.993 -

L-Max/G-Max 1.009 0.983 0.971 1.013 1.029 6

PF 0.95 L-Min/G-Min 1.009 0.999 0.994 1.002 1.002 -

L-Min/G-Max 1.006 0.996 0.991 1.019 1.035 7

L-Max/G-Min 1.002 0.976 0.965 0.987 0.987 -

L-Max/G-Max 1.006 0.980 0.968 1.015 1.034 9

LDC L-Min/G-Min 0.984 0.974 0.969 0.978 0.978 -

L-Min/G-Max 0.982 0.971 0.967 1.006 1.032 10

L-Max/G-Min 1.009 0.983 0.972 0.994 0.995 -

L-Max/G-Max 1.008 0.982 0.971 1.016 1.034 7

VBoost & LDC L-Min/G-Min 0.984 0.974 0.969 0.978 0.979 -

L-Min/G-Max 0.986 0.976 0.971 1.029 1.033 20

L-Max/G-Min 1.009 0.983 0.972 0.994 0.995 -

L-Max/G-Max 1.012 0.986 0.974 1.035 1.034 14

Table 7.4 – Summary of steady state simulation results

7 - Simulation

Page 54: Connecting Renewable Generation with Active Network Management

Table 7.4 above shows bus voltages plus DG penetration for the four ANM scenarios and each

possible load generation combination.

Passive Voltage Control

As explained in section X Bus 6 voltage rise limits the DG capacity and maintenance of Bus 4

(ideally always at 0.970pu) determines the optimum setting for Bus 2.

Referring back to equation (4.14) we can apply this to the general problem of voltage rise for Bus 6.

∆Vp.u = RPg + XQg

For line 5-6 the voltage rise due to 6MW of DG is

0.357 x 6/100 = 0.021Vpu (7.1)

At a PF of 0.98 the reactive the volt drop on the line is

0.4 x 1.2/100 = 0.005 Vpu (7.2)

This gives an overall voltage rise for 6MW @ PF 0.98 of

0.021 – 0.005 = 0.016 Vpu (7.3)

I.e approximately 0.003 Vpu per MW of DG output at PF 0.98.

Taking the base case scenario results from table 7.4 under maximum DG output conditions the volt

rise due to DG on line 5-6 is

1.029 – 1.013 = 0.016 Vpu (7.4)

This correlates with the above calculation (7.3).

7 - Simulation

Page 55: Connecting Renewable Generation with Active Network Management

PF 0.95

For PF 0.95 the reactive consumption is 2.0MVAr for 6MW DG output. Hence the increase of

reactive consumption of 0.8MVAr gives a decrease in Bus 6 voltage of

0.4 x 0.8/100 = 0.003 Vpu

This allows approximately one extra MW of DG capacity per six DG MW installed.

LDC

During minimum load cases Bus 2 voltage is allowed to drop and hence allow more headroom for

generation. For maximum load the results are similar to the original PF 0.98 case. This is because

the Bus 2 voltage is similar in both maximum load cases.

Voltage Booster & LDC

This ANM method results in a best case trebling of connectable DG capacity by allowing flattening

of the voltage profile between Bus 6 and Bus 5.

General Observations

The variation of the regulation on Bus 2 presents a problem in evaluating the results. The regulation

settings of the Bus were all +/- 0.005 (as specified in table 7.3). This variation of up to 0.010 pu

results in a variation in possible DG capacity of 3MW and illustrates the sensitivity of this circuit to

variation in voltage control of only +/- ½%.

7.4 Time Series Simulation

7.4.1 Overview

In order to ascertain the more realistic operational performance of the ANM an hour by hour

simulation series was ran for each case. This is because occurrences of the combinations of

maximum and minimum conditions for load and generation are statistically unlikely and so these

possible limiting conditions will have a smaller impact on yearly DG output than might first be

expected.

Minimising curtailment the main 'objective' is achieved by setting the OLTC ctrl to maximise DG

headroom at all times. The exact same settings and ANM cases specified in the Steady State

simulations of section X were used for these Time Series Simulations also.

7.4.2 Curtailment Modelling

For modelling of the WF a 'bus injection group' consisting of the DG generator which can be varied

in size along with a number of 'dump loads' are grouped together on Bus 6. Definition as an

'injection group' in PW enables easier management of data when copying results data to the

spreadsheet. The number of loads switched on not being relevant to the analysis only the 'effective'

export output of the WF.

7 - Simulation

Page 56: Connecting Renewable Generation with Active Network Management

Initial modelling of curtailment was done directly via 'Post Flow Action' control of generator output.

However this was found to not function as expected and a simpler setup using individual 'dummy'

loads was used at the suggestion of PW support. See appendix figure A1.23 and PW manual [30] for

more info on the functioning of this simulator feature and appendix A3.3 for email correspondence

regarding development of the curtailment with Post Flow Actions.

Note each load has a reactive compensator providing more VArs to the generators as their real

output rises. This keeps the WF power factor constant as the generator reactive consumption

increases. Dump load real and reactive values used are shown in the results spreadsheet section X.

Note some of the load dumps on the one-line drawings are hidden for clarity.

Operation

Each dummy load is assigned a voltage level. During simulation the network power flow solution is

found. Under excessive WF generation this can result in a very high voltage on Bus 6. If this occurs

the Post Flow Action turns on an appropriate number of the dump loads and resolves the power flow

solution with these new dump loads in circuit. Figure 7.7 below illustrates the logic.

Fig. 7.7 – Open loop proportional control scheme for curtailment

7 - Simulation

Volt Bus 6 > 1.035

Load 1 ON

ELSE

Load 1 OFF

Dummy Load 1

Volt Bus 6 > 1.043

Load 2 ON

ELSE

Load 2 OFF

Dummy Load 2

Volt Bus 6 > 1.051

Load 3 ON

ELSE

Load 3 OFF

Volt Bus Wind Farm > Threshold

Load X ON

ELSE

Load X OFF

Dummy Load X

Dummy Load 3

“Open Loop Proportional Control Scheme” for

Bus 6 Wind Farm Curtailment Logic using

dummy loads. (0.008 pu resolution)

Page 57: Connecting Renewable Generation with Active Network Management

Refinement

As a good compromise between accuracy and complexity of setup steps of 0.008 Vpu were used

between dump load switch-on thresholds. Although the calculated Bus 6 voltage rise of

approximately 0.003pu per 1MW the curtailment dump loads were set to between 1.5 and 2.2 MW

per 0.008.

The slight difference in voltage rise on Bus 6 occurs between the different four cases and also as the

WF size changes. To adjust for this a 'Zone load scale control' is included on the one-line diagram.

This enables easy global adjustment of the size of real dump loads. The 'model explorer' is used to

globally adjust the size of the reactive dump loads. See appendix figure A1.22 for a model explorer

screen-shot.

Some extra adjustment of curtailment results are required as the curtailment is 'open-loop' in nature.

This fine tuning is done by an error adjustment column in the spreadsheet. See figure 7.8 below for

the logic.

Fig. 7.8 – Logic for curtailment error correction in spreadsheet

The resulting over-curtailments and under-curtailments are summed and an adjustment made in a

'corrected' curtailment results column.

Example of Curtailment Operation

Figures 7.9 to 7.11 below show the operation of curtailment for the cases of passive voltage control

at 10MW DG capacity and LDC control at 10MW capacity for the second week I.e timepoints 169

to 336. Flattening of the Bus 4 voltage due to LDC control (light-blue coloured plot) clearly results

in increased headroom and thus decreased curtailment.

7 - Simulation

IF Curtailment occurred

(i.e Wind Farm Output > Bus Export)

THEN

Next

IF Bus 6 Voltage > 1.035

THEN

Undercurtailment TRUE

(Increase Mwh Curtailed, reduce MWh generated)

ELSIF Bus 6 Voltage < 1.025

THEN

Overcurtailment TRUE

(Decrease Mwh Curtailed, Increase MWh generated)

Logic for curtailment error correction

in spreadsheet

Page 58: Connecting Renewable Generation with Active Network Management

Fig. 7.9 – Curtailment operation for Power Factor 0.98 ( 10MW size Wind Farm )

Fig. 7.10 - Example of increased headroom from LDC (10MW size Wind Farm)

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

10:00:00

20:00:00

06:00:00

16:00:00

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

10:00:00

20:00:00

06:00:00

16:00:00

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

10:00:00

20:00:00

06:00:00

16:00:00

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

10:00:00

20:00:00

06:00:00

16:00:00

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

10:00:00

20:00:00

06:00:00

16:00:00

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

0.94

0.96

0.98

1

1.02

1.04

1.06

0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

18.00

20.00

Network Bus Voltages & Curtailment

1 PU Volt 2 PU Volt 3 PU Volt 4 PU Volt 5 PU Volt 6 PU Volt 1 TO 2 CKT 1 Tap Ratio Bus Voltage Max Bus Voltage Min Curt

Timepoint

V p

.u &

Bu

s 2

OL

TC

Rati

o

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

10:00:00

20:00:00

06:00:00

16:00:00

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

10:00:00

20:00:00

06:00:00

16:00:00

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

10:00:00

20:00:00

06:00:00

16:00:00

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

10:00:00

20:00:00

06:00:00

16:00:00

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

10:00:00

20:00:00

06:00:00

16:00:00

02:00:00

12:00:00

22:00:00

08:00:00

18:00:00

04:00:00

14:00:00

00:00:00

0.94

0.96

0.98

1

1.02

1.04

1.06

0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

18.00

20.00

Network Bus Voltages & Curtailment

1 PU Volt 2 PU Volt 3 PU Volt 4 PU Volt 5 PU Volt 6 PU Volt 1 TO 2 CKT 1 Tap Ratio Bus Voltage Max Bus Voltage Min Curt

Timepoint

V p

.u &

Bu

s 2

OL

TC

Rati

o

Page 59: Connecting Renewable Generation with Active Network Management

10MW DG with LDC control 10MW DG with passive voltage control

Curtailment levels of 1.5, 3.0 and 4.5 MW. Light blue line is Bus 4 voltage.

Fig. 7.11 - Decreased curtailment resulting from LDC control compared to base case for timepoints 169 to 336 with 10MW Wind Farm

Page 60: Connecting Renewable Generation with Active Network Management

7.4.3 Spreadsheet

A spreadsheet was used to create load data and WF output data. This data was then copied into PW

for use by the TSS. Results were then copied back into the spreadsheet for analysis and presentation.

A second spreadsheet was used to collate general results from each of the 4 week simulation runs

with a worksheet summarising curtailment, losses and financial results.

7.4.4 Simulation Procedure

Fig. 7.12 - Basic 4 week TSS procedure for a single ANM case

In order to reduce the quantity of data involved for the simulation and reduce the length of the

simulations only 672 time periods are solved instead of a full 8760 (the number of hours in a year).

This equates to four weeks of data. To make the data more realistic the four weeks of data are made

up of single weeks from four different seasons of the year. The same was done with the demand

data. Table 7.5 below shows the data timepoints used

7 - Simulation

PowerWorld (PW) : Setup

chosen ANM scenario

Spreadsheet (SS) :

Choose Wind Farm

Size

SS : Export Wind Farm

Output profile to PW

PW : Run Simulation

PW : Export Results

to SS

SS : Copy individual

results into Aggregated

results SS

Repeat

for 0 to 20 MW

in 2 MW steps

Simulations

Complete

Page 61: Connecting Renewable Generation with Active Network Management

TSS timepoint UKGDS wind data profile

timepoint

National Grid Demand data

timepoint (2007)

(Wk 1) 1 - 168 1 - 168 Jan 1st to 7th

(Wk 2) 169 - 336 2160 - 2327 Apr 1st to 7th

(Wk 3) 337 - 504 4344 - 4511 Jul 1st to 7th

(Wk 4) 505 - 672 6552 - 6719 Oct 1st to 7th

Table 7.5 – Timepoints used for data input

The results from these '4 week' runs were then multiplied by 365/28 to give approximated results

for the year.

11 simulations were done with WF sizes from 0 to 20 MW in 2MW steps for each of the four case

scenarios making a total of 44 simulation runs.

The time required to run a complete set of the simulations (on a 1.5GHz Pentium with 512MB Ram)

was :

1 single TSS run of 1min + 9 min spreadsheet copy & pasting = 10min

11 TSS runs for each of 0-20MW WF sizes = 11 x 10min = 110 min

4 ANM case scenarios = 4 x 110 = 440 min total = 8hrs

7.4.5 WF Power Factor

Initially a circle curve approximating an induction generator was used. However for simplicity the

generator setup was changed to an approximately constant power factor. In addition the WFs

capability of operation at constant power factor is thought to be a more realistic requirement. Hence

the generators are assumed to be neither induction or synchronous. The Power Factors were

calculated as below

PF 0.98

PF = 1

Sqrt (Q� + P�)

Q = Sqrt ( 1 -1 )

PF�

= 0.203 MVAr per MW

7 - Simulation

Page 62: Connecting Renewable Generation with Active Network Management

PF 0.95

Using the same equation above for PF 0.95 reactive power consumption is

= 0.329 MVAr per MW

This gives a consumption for a 20MW WF of

MVAr = 4.4 @ PF 0.98

and

MVAr = 6.7 @PF 0.95

PW modelling limitations appear to restrict the generator 'capabilty curve' from passing through 0,0

in other words constant Power Factor operation down to zero output was not possible. To try and

account for this +0.1 MVAr of capacitive reactance is switched in under non zero WF output

conditions to improve the bus PF at low outputs. Hence the circle curve for the bus is effectively the

generator circle curve 'pulled down' by 0.1 MVAr. The circle curve achieved is shown in figure 7.13

below, the ideal PF 0.98 line being shown in dashed.

Fig. 7.13 – Generator and injection group circle diagram for 8MW size Wind Farm at PF 0.98

7 - Simulation

0 1 2 3 4 5 6 7 8 9

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Generator & Injection Group (Bus 6) - Circle Diagram

Generator MVAr Bus 6 MVAr

Real Power Output (MW)

Re

ac

tiv

e P

ow

er

Co

ns

um

pti

on

(M

VA

r)

Page 63: Connecting Renewable Generation with Active Network Management

The resulting variation of power factor is shown in figure 7.14 below. Some variation occurs

especially at very low WF outputs due to the non-linearity of the circle curve. However it is not

thought significant.

Fig. 7.14 – Real power, apparent power and power factor for bus 6 with 8MW Wind Farm at PF 0.98

7.4.6 Wind Turbine Output Profile

A normalised wind farm power output series was taken from UKGDS AMP tool [6] which contains

a years worth of data. Four single weeks were selected to make up the 672 timepoints of data for

input to the TSS the below table 7.6 shows their respective capacity factors.

Week Capacity Factor

1 18

2 56

3 21

4 29

4 Week Series 31

Table 7.6 – Wind Output capacity factors

It is interesting to note that the capacity factor for the complete year of UKGDS data was 21%

whereas the selected weeks used in this simulation when assembled gave a higher value of 31%.

7 - Simulation

Page 64: Connecting Renewable Generation with Active Network Management

The zeros in this data indicate times when the wind speed is below turbine cut-in speed or above

wind turbine cut-out speed. The ones indicate where wind turbine output has reached its maximum.

See appendix A1.3 for individual graphs of these 4 weeks. The complete set of assembled four

weeks of data are shown below in figure 7.15 below.

Fig. 7.15 – Normalised Wind Farm power output for the 4 weeks

7.4.7 Demand Profile

Data downloaded from the National Grid website [31] was used to model loads in the simulation.

The source data used was for the total UK electricity consumption so was thought be a

representative data for the specified mixed industrial, commercial and domestic load on the case

study network. The minimum to maximum ratio of the data is 0.41 which is close to that used in the

book of 0.4. The complete assembled four weeks of data are shown in figure 7.16 below.

Fig. 7.16 – Normalised demand profile for the 4 weeks

7 - Simulation

2 42 82 122

12

22

32 52

62

72 92

102

112 132

142

152

162

172

182

192

202

212

222

232

242

252

262

272

282

292

302

312

322

332

342

352

362

372

382

392

402

412

422

432

442

452

462

472

482

492

502

512

522

532

542

552

562

572

582

592

602

612

622

632

642

652

662

672

0

0.2

0.4

0.6

0.8

1

1.2

Normalised Wind Gen Output - 4 weeks

Timepoint

2 42 82 122

12

22

32 52

62

72 92

102

112 132

142

152

162

172

182

192

202

212

222

232

242

252

262

272

282

292

302

312

322

332

342

352

362

372

382

392

402

412

422

432

442

452

462

472

482

492

502

512

522

532

542

552

562

572

582

592

602

612

622

632

642

652

662

672

0.000

0.200

0.400

0.600

0.800

1.000

1.200

Normalised Demand

Timepoint

Page 65: Connecting Renewable Generation with Active Network Management

7.5 Development Methodology

As a summary the overall simulation process followed the stages below

� PW Evaluation

� Basic circuit evaluation under steady state conditions

� Short TSS with a basic load and generation profiles (See appendix A1.4)

� Development of curtailment and data presentation in spreadsheet for input/output from/to

PW

� Four week long TSS

� Steady State ANM setups

� Four week long ANM TSS including finance

7 - Simulation

Page 66: Connecting Renewable Generation with Active Network Management

8 - Results

For the four ANM cases and each wind farm capacity between zero and twenty MW the results are

compiled into three spreadsheets Generation, Losses and Finance.

8.1 Generation Spreadsheet

Note the first five columns were used for fine adjustment of the simulation cases and can be ignored

for the analysis of the results.

For each ANM case and wind farm size table 8.1 the generation worksheet shows yearly quantities

of

� Net Generation - DG output exported onto the network

� Generation Curtailed – DG output prevented from export

� Wind Curtailment (%) - Percentage of DG output prevented from export

� Wind Energy Supplied (% Load) – Percentage of total network load which DG

supplies

� Wind Energy Peak Output – Maximum value of export of DG to nework

8.1.1 Curtailment

Plots of gross generation showing the component of curtailed generation are shown in figures 8.1 to

8.4 for the four ANM cases.

Passive Voltage Control

For this base case a curtailment (at 2%) is seen to begin at 6 MW with a net production of around

16,000 MWh/year. For a 8MW DG curtailment is 12.5% and yearly net production 19,000 MWh. A

maximum net production of approximately 27,000 MWh is produced for 20MW DG at a

curtailment level of 49%.

PF 0.95

For this base case curtailment (at 6%) is seen to begin at 8 MW with a net production of around

20,000 MWh/year. For a 10MW of DG curtailment is 14.6% and yearly net production 23,000

MWh. A maximum net production of approximately 32,000 MWh is produced for 20MW of DG at

a curtailment level of 41%.

This ANM case essentially allows an extra 2MW of DG capacity by shifting the 'curtailment curve'

upwards.

8 - Results

Page 67: Connecting Renewable Generation with Active Network Management

LDC

For this base case curtailment (at 2%) is seen to begin again at 8 MW with a net production of

around 21,000 MWh/year. For 10MW of DG curtailment is 11% and yearly net production 24,000

MWh. A maximum net production of approximately 33,000 MWh is produced for 20MW of DG at

a curtailment level of 39%.

This ANM case is very similar to that of the PF 0.95 case but with marginally better generation

output. At 10 MW curtailment is reduced from 14.6% to 11%.

LDC & Voltage Booster

This case shows a huge improvement in wind farm net production with only 2 and 5% curtailment at

18 and 20 MW respectively.

8.1.2 General Observations

Curtailment

Comparing with the steady state analysis of chapter 7 these results show a close correlation. I.e

taking crude averages of the DG capacity limit from table 7.4 for the steady state cases we get

capacities of 6, 8 , 8.5 and 17 respectively. If a commercial constraint of a maximum curtailment of

10% was taken then the TSS cases would give approximately the same optimum size of DG as the

non-curtailment steady state analysis.

Wind Energy Supplied

The energy penetration onto the network is limited to between 4 and 12% for the first three cases.

The Voltage Booster with LDC case allows nearly 20%. This is a significant figure in terms of

yearly penetration and would be accompanied by peak penetrations much higher with implications

for other generation on the network.

8 - Results

Page 68: Connecting Renewable Generation with Active Network Management

Table 8.1 – Wind farm generation results spreadsheet

WF Size Net Curt error Dump Scale Net Generation Generation Curtailed Wind Curt % Wind Energy SuWind Energy Pk Output

of tot curt (Corrected) supplied peak output

(MW) (%) (%) of Wind Gen (%) of total load (MW)

PF 098 4 0.0 1.6 -0.32 1.000 0.992 10871 -1 0.0 3.8 4.0

6 5.3 1.6 -0.32 1.000 0.989 15930 375 2.3 5.6 6.0

8 6.6 1.7 -0.34 1.000 0.988 19026 2713 12.5 6.7 8.0

10 4.9 1.7 -0.34 1.000 0.986 21127 6047 22.3 7.5 7.3

12 2.4 1.7 -0.34 1.000 0.986 22921 9688 29.7 8.1 7.8

14 1.5 1.7 -0.34 1.000 0.985 24421 13623 35.8 8.6 7.7

16 0.3 1.7 -0.34 1.000 0.984 25569 17910 41.2 9.0 7.5

18 -0.3 1.7 -0.34 1.000 0.984 26739 22174 45.3 9.4 7.7

20 -1.2 1.7 -0.34 1.000 0.983 27643 26706 49.1 9.7 7.6

PF_095 4 0.0 2 -0.66 1.000 0.979 10871 -1 0.0 3.8 4.0

6 0.0 2 -0.66 1.000 0.973 16305 -1 0.0 5.8 6.0

8 0 2 -0.66 1.000 0.969 20357 1382 6.36 7.18 8.00

10 0.0 2 -0.66 1.000 0.966 23211 3963 14.6 8.2 8.2

12 -0.5 2 -0.66 1.000 0.964 25721 6888 21.1 9.1 9.0

14 -4.8 2 -0.66 1.000 0.963 27852 10192 26.8 9.8 8.9

16 -6.1 2 -0.66 1.000 0.961 29348 14131 32.5 10.3 10.0

18 -4.7 2 -0.66 1.000 0.960 30822 18092 37.0 10.8 10.0

20 -3.7 2 -0.66 1.000 0.959 32201 22147 40.8 11.3 10.0

LDC 4 0.0 1.6 -0.32 1.000 0.991 10871 -1 0.0 3.8 4.0

6 0.0 1.6 -0.32 1.000 0.989 16305 -1 0.0 5.8 6.0

8 5.6 1.6 -0.32 1.000 0.987 21384 355 1.6 7.5 8.0

10 4 1.7 -0.34 1.000 0.986 24211 2963 10.90 8.53 8.75

12 1.1 1.7 -0.34 1.000 0.985 26494 6115 18.8 9.3 9.4

14 0.3 1.7 -0.34 1.000 0.985 28740 9304 24.5 10.1 9.5

16 -0.3 1.7 -0.34 1.000 0.984 30520 12958 29.8 10.7 9.2

18 -1.5 1.7 -0.34 1.000 0.984 32006 16907 34.6 11.3 9.5

20 -1 1.7 -0.34 1.000 0.983 33378 20970 38.58 11.74 10.11

LDC + VB 4 0.0 2.2 -0.44 1.000 0.991 10871 -1 0.0 3.8 4.0

6 0.0 2.2 -0.44 1.000 0.989 16305 -1 0.0 5.8 6.0

8 0.0 2.2 -0.44 1.000 0.987 21739 0 0.0 7.7 8.0

10 0.0 2.2 -0.44 1.000 0.986 27174 0 0.0 9.6 10.0

12 0.0 2.2 -0.44 1.000 0.985 32609 0 0.0 11.5 12.0

14 0.0 2.2 -0.44 1.000 0.985 38044 0 0.0 13.3 14.0

16 64.3 2.2 -0.44 1.000 0.984 43469 10 0.0 15.2 16.0

18 4.7 2.2 -0.44 1.000 0.984 48011 902 1.8 16.7 18.0

20 -13.9 2.2 -0.44 1.000 0.983 51442 2906 5.3 17.9 20.0

MVAr nom Max PF Av PF

Aka Wind Inj Grp Corre

(MWh/yr) (MWh/yr)

Page 69: Connecting Renewable Generation with Active Network Management

Fig. 8.1 – Curtailment and generation for PF 0.98 case Fig. 8.2 – Curtailment and generation for PF 0.95 case

Fig. 8.3 – Curtailment and generation for LDC case Fig. 8.4 – Curtailment and generation for LDC & Voltage Booster case

4 6 8 10 12 14 16 18 20

0

10000

20000

30000

40000

50000

60000

PF 0.98

Net Generation Generation Cur-

tailed

Wind Farm Size (MW)

MW

h/y

r

4 6 8 10 12 14 16 18 20

0

10000

20000

30000

40000

50000

60000

PF 0.95

Net Generation Generation

Curtailed

Wind Farm Size (MW)

MW

h/y

r

4 6 8 10 12 14 16 18 20

0

10000

20000

30000

40000

50000

60000

LDC

Net Generation Generation Cur-

tailed

Wind Farm Size (MW)

MW

h/y

r

4 6 8 10 12 14 16 18 20

0

10000

20000

30000

40000

50000

60000

LDC and Voltage Booster

Net Generation Generation

Curtailed

Wind Farm Size (MW)

MW

h/y

r

Page 70: Connecting Renewable Generation with Active Network Management

8.2 Losses Spreadsheet

For each ANM case and wind farm size table 8.2 shows yearly losses for

� Area - Sum of all losses occurring on the one-line diagram

� Individual Circuits – Individual loss for particular line or transformer

� Area Peak & Minimum – Instantaneous greatest and smallest loss which occurred

� Total Network Peak & % of load peak – Estimated loss including 'feeder C'

� Incremental Area – Increase in loss over no DG base case

It is noted in the network case scenario description that the load located on Bus 2 represents

'aggregated' loads on the rest of the network. Since this load is nearly 2/3 of the total load flowing

through the transformer it is interesting to consider its possible effect regarding this network case

scenario. Examination of feeder A reveals no change in losses due to DG so it can be assumed that

the same will apply to feeder C. For comparison with general distribution network loss figures an

estimation is made for losses due to feeder C and this is added to the area losses and referred to as

'total network' losses. This is as opposed to 'area' losses which concern the full circuits drawn on the

one-line diagram.

8.2.1 Circuit Losses

No DG case

For the original circuit with no DG connected a yearly loss of 3439MWh occurs.

PF 0.98

At 2MW DG area losses are reduced by 103 MWh. This is a 3% reduction on total area losses.

Losses then steadily rise until at 8 MW DG the loss is 113 MWh, a 3% increase in total area losses.

Figure 8.5 shows that as DG output rises losses in the GSP transformer reduce as the power flow

through it is reduced. Circuits 2-5 and 5-6 contribute all of the increased losses when the DG output

rises above 4MW. Circuit 2-5 contributes to a reduction in losses at below 4MW DG as power

flowing from the DG feeds the load on Bus 5 and displaces current previously flowing in circuit

2-5.

PF 0.95

This case has losses with characteristics similar to the base case above with the exception of higher

losses on lines 2-5 and 5-6. This must be from the increased transmission of real power due to the

slightly reduced curtailment but should also be in part from the increased transmission of reactive

power along these lines to serve the increased VAr consumption of the wind farm.

8 - Results

Page 71: Connecting Renewable Generation with Active Network Management

LDC

Figure 8.7 shows an almost identical shape to that of figure 8.6 for the PF 0.95 case. Inspection of

table 8.2 reveals that losses are slightly less for the LDC case due to lines 2-5 and 5-6 even though

the net DG production is slightly higher for the LDC case.

Voltage Booster & LDC

Due to the effectiveness of this method in providing DG connection capacity a huge increase of

losses occurs of approximately 3600 MWh for the year which doubles the area losses on the

network. This is due simply to ohmic losses rising as the square of the current (at 10MW the losses

are only 800 MWh).

As a proportion of the real power transmitted through line 5-6 which is 27,174 MWh for 10 MW of

DG the losses form 2.9%.

8.2.2 General Observations

As a useful ballpark figure in characterising the network total losses can be worked out as a

percentage of load served. An estimation of the circuit loss due to feeder C would be approximately

2.5 times that of feeder A. Hence with no DG output overall circuit losses on the total network

would be

% Network Loss = (Area Loss + Feeder C Estimation )

Total Network Load

= 3439 + 3610

279983

= 2.5%

8 - Results

Page 72: Connecting Renewable Generation with Active Network Management

Table 8.2 - Spreadsheet results sheet for Losses

279983

No DG Area Ls 3439 Individual Circuit Losses Total Network As as % of

3610 of Area

7049 0.35 47.5 Loss

1 to 2 2 to 3 2 to 5 3 to 4 5 to 6 L 5 to 6 Booster Area Pk Area Min Pk Loss area loss

WF Size (MW) (%) (%) loss (%)

0 3439 1809 1111 189 333 0 0 0.85 0.13 1.725 3.6 0 0.0 0.0 2.5

2 3336 1746 1111 119 333 19 0 0.85 0.13 1.725 3.6 -103 -3.0 -1.5 2.5

PF98 4 3362 1689 1111 120 333 111 0 0.85 0.13 1.725 3.6 -77 -2.2 -1.1 2.5

6 3486 1638 1111 174 334 232 0 0.85 0.13 1.725 3.6 47 1.4 0.7 2.5

8 3553 1609 1111 198 334 298 0 0.85 0.13 1.725 3.6 114 3.3 1.6 2.6

10 3590 1589 1110 216 333 338 0 0.85 0.13 1.725 3.6 151 4.4 2.1 2.6

12 3643 1571 1109 239 333 391 0 0.85 0.13 1.725 3.6 204 5.9 2.9 2.6

14 3689 1555 1109 260 333 435 0 0.85 0.13 1.725 3.6 250 7.3 3.5 2.6

16 3729 1545 1109 281 332 467 0 0.85 0.13 1.725 3.6 290 8.4 4.1 2.6

18 3777 1531 1108 304 332 503 0 0.85 0.13 1.725 3.6 338 9.8 4.8 2.6

20 3828 1520 1108 327 332 537 0 0.85 0.13 1.725 3.6 389 11.3 5.5 2.7

0 3439 1809 1111 189 333 0 0 0.85 0.13 1.725 3.6 0 0.0 0.0 2.5

2 3342 1747 1111 123 333 25 0 0.85 0.13 1.725 3.6 -97 -2.8 -1.4 2.5

PF95 4 3381 1691 1110 132 333 113 0 0.85 0.13 1.725 3.6 -58 -1.7 -0.8 2.5

6 3548 1640 1109 208 333 255 0 0.85 0.13 1.725 3.6 109 3.2 1.5 2.6

8 3715 1605 1109 284 333 386 0 0.85 0.13 1.725 3.6 276 8.0 3.9 2.6

10 3826 1579 1109 335 333 473 0 0.85 0.13 1.725 3.6 387 11.3 5.5 2.7

12 3946 1556 1109 390 332 558 0 0.96 0.13 1.835 3.9 507 14.7 7.2 2.7

14 4092 1534 1109 459 332 656 0 0.94 0.13 1.815 3.8 653 19.0 9.3 2.8

16 4194 1520 1109 507 332 724 0 1.05 0.13 1.925 4.1 755 22.0 10.7 2.8

18 4271 1507 1109 547 332 779 0 1.05 0.13 1.925 4.1 832 24.2 11.8 2.8

20 4347 1496 1109 583 332 832 0 1.07 0.13 1.945 4.1 908 26.4 12.9 2.8

0 3491 1835 1129 192 339 0 0 0.84 0.14 1.715 3.6 52 1.5 0.7 2.5

2 3391 1771 1129 120 339 19 0 0.84 0.14 1.715 3.6 -48 -1.4 -0.7 2.5

LDC 4 3417 1712 1129 121 339 112 0 0.84 0.14 1.715 3.6 -22 -0.6 -0.3 2.5

6 3561 1661 1129 187 339 250 0 0.84 0.14 1.715 3.6 122 3.5 1.7 2.6

8 3794 1614 1129 291 339 414 0 0.84 0.14 1.715 3.6 355 10.3 5.0 2.6

10 3888 1588 1129 334 339 493 0 0.84 0.14 1.715 3.6 449 13.1 6.4 2.7

12 3977 1566 1129 378 339 567 0 0.84 0.14 1.715 3.6 538 15.6 7.6 2.7

14 4089 1547 1129 429 339 645 0 0.84 0.14 1.715 3.6 650 18.9 9.2 2.7

16 4181 1531 1129 476 339 710 0 0.84 0.14 1.715 3.6 742 21.6 10.5 2.8

18 4269 1516 1129 517 339 767 0 0.84 0.14 1.715 3.6 830 24.1 11.8 2.8

20 4340 1503 1128 552 339 817 0 0.85 0.14 1.725 3.6 901 26.2 12.8 2.8

0 3491 1835 1129 192 339 0 0 0.84 0.14 1.715 3.6 52 1.5 0.7 2.5

2 3391 1771 1129 121 339 0 19 0.84 0.14 1.715 3.6 -48 -1.4 -0.7 2.5

LDC & V Booste 4 3419 1713 1129 122 339 0 112 0.84 0.14 1.715 3.6 -20 -0.6 -0.3 2.5

6 3567 1661 1128 190 339 0 252 0.84 0.14 1.715 3.6 128 3.7 1.8 2.6

8 3849 1612 1129 313 339 0 457 0.84 0.14 1.715 3.6 410 11.9 5.8 2.7

10 4239 1568 1129 498 339 0 704 1.04 0.14 1.915 4.0 800 23.3 11.3 2.8

12 4742 1528 1129 740 339 0 1008 1.31 0.14 2.185 4.6 1303 37.9 18.5 3.0

14 5356 1493 1129 1033 339 0 1361 1.64 0.14 2.515 5.3 1917 55.7 27.2 3.2

16 6068 1461 1129 1380 338 0 1762 2 0.14 2.875 6.1 2629 76.4 37.3 3.5

18 6695 1435 1129 1686 338 0 2105 2.41 0.14 3.285 6.9 3256 94.7 46.2 3.7

20 7203 1412 1129 1934 338 0 2390 2.83 0.14 3.705 7.8 3764 109.5 53.4 3.9

Total Netw Load (MWh/yr)

ie + Contrib from

Est Fdr C 0MW Ls Fdr C of appx 2.5 x Fdr A % of Max Load Total netw Load

0MW Est Tot Netw Ls Incr

Area Netw Tot Netw Pk Ls incr tot netw Total Netw Ls

(MWh/yr) (MWh/yr) (MWh/yr) (MWh/yr) (MWh/yr) (MWh/yr) (MWh/yr) MWh MWh (MWh) MWh/yr (avg)

Page 73: Connecting Renewable Generation with Active Network Management

Fig. 8.5 – Circuit losses case PF 0.98 Fig. 8.6 – Circuit losses case PF 0.95

Fig. 8.7 – Circuit losses case LDC Fig. 8.8 – Circuit losses case Voltage Booster & LDC

0 2 4 6 8 10 12 14 16 18 20

0

500

1000

1500

2000

2500

3000

3500

4000

4500

PF 0.98

1 to 2 2 to 3 2 to 5

3 to 4 5 to 6 L 5 to 6 Booster

Wind Farm Size (MW)

Cir

cu

it L

oss (

MW

h/y

r)

0 2 4 6 8 10 12 14 16 18 20

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

PF 0.95

1 to 2 2 to 3 2 to 5

3 to 4 5 to 6 L 5 to 6 Booster

Wind Farm Size (MW)

Cir

cu

it L

osses (

MW

h/y

r)

0 2 4 6 8 10 12 14 16 18 20

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

LDC

1 to 2 2 to 3 2 to 5

3 to 4 5 to 6 L 5 to 6 Booster

Wind Farm Size (MW)

Cir

cu

it L

osses (

MW

h/y

r)

0 2 4 6 8 10 12 14 16 18 20

0

1000

2000

3000

4000

5000

6000

7000

8000

LDC & VBooster

1 to 2 2 to 3 2 to 5

3 to 4 5 to 6 L 5 to 6 Booster

Wind Farm Size (MW)

Cir

cu

it L

os

se

s (

MW

h/y

r)

Page 74: Connecting Renewable Generation with Active Network Management

Overall Circuit Losses

The incremental change in area losses for all of the ANM case scenarios is shown in figure 8.8

below. A break even point for losses for all cases appears at about 5 MW DG capacity. Up unto this

point no curtailment takes place. Notably as net generation increases losses rise accordingly.

Comparing losses between the power factor 0.95 case and the LDC case the former is lower. This

may be because the net generation and therefore power flows are slightly less for the PF 0.95 case.

On the other hand some effect due to the LDC generally holding Bus 4 and therefore other voltages

on the network lower may increase losses due to the higher required currents. Both of the cases

using LDC have higher incremental losses than the non-LDC control cases.

Fig. 8.8 – Overall losses

Note the voltage booster case in figure 8.8 above has a maximum incremental loss of 3764 MWh

per year.

Losses as a function of time

Due to the peaky nature of the wind profile under high DG penetration the losses essentially follow

the wind profile. At times of low DG output the losses follow the demand profile. Figure 8.9 below

shows these trends. The yellow plot is total area losses and the brown and dark blue plots are circuits

2-5 and 5-6 respectively.

8 - Results

0 2 4 6 8 10 12 14 16 18 20

-200

0

200

400

600

800

1000

Change in Total Area Losses

above 3363MWh/yr No DG base case

PF98 PF95 LDC LDC & V Booster

Wind Farm Size (MW)

MW

h/y

r

Page 75: Connecting Renewable Generation with Active Network Management

Fig. 8.9 - General Variation of Losses with time for 16MW Wind Farm case LDC & Voltage Booster

02:0

0:00

00:00:

00

22:0

0:00

20:0

0:00

18:0

0:00

16:00

:00

14:0

0:00

12:0

0:00

10:00:

00

08:0

0:00

06:0

0:00

04:00:

00

02:0

0:00

00:0

0:00

22:0

0:00

20:00:

00

18:0

0:00

16:0

0:00

14:00:

00

12:00:

00

10:0

0:00

08:0

0:00

06:00:

00

04:0

0:00

02:0

0:00

00:0

0:00

22:00:

00

20:0

0:00

18:0

0:00

16:00:

00

14:00

:00

0

0.5

1

1.5

2

2.5

0.9

0.95

1

1.05

1.1

Line & Xfrmr Losses

1 TO 2 CKT 1 MW Loss 2 TO 3 CKT 1 MW Loss 2 TO 5 CKT 1 MW Loss 3 TO 4 CKT 1 MW Loss 5 TO 6 CKT 1 MW Loss 1 Loss MW 5 TO 6 CKT 2 MW Loss 1 TO 2 CKT 1 Tap Ratio

Timepoint

MW

Page 76: Connecting Renewable Generation with Active Network Management

8.3 Finance Spreadsheet

For each of the ANM cases and DG wind farm sizes the financial worksheet shown in table 8.3

gives yearly values for :-

� Gross Revenue – Income from sale of electricity

� Loss Penalty – Fee incurred to pay for incremental network losses

� Cost of ANM – Upfront capital investment for ANM solution

� Capital Investment - Total upfront investment cost DG

� Operation & Maintenance (O&M) Cost – Ongoing yearly payment for upkeep of DG

� Net Income – Yearly cash flow to DG developer

� Return on Investment (ROI) – Percentage of initial investment which is returned as

income each year

8.3.1 Calculation of costings

A simplified method is used to calculate the financial viability of the DG investment. All capital

costs are assumed to be paid in full at the beginning of the first year. Also no discounting of

revenues or expenses takes place. Thus project lifetime or value of the DG at the end of this lifetime

is not considered in the analysis.

The following relationships show how the financial results are obtained :-

Gross Revenue = Net Generation x Electricity Sell Price(includes ROC)

Loss Penalty = Net Generation x Incremental Losses x DNO Loss Penalty Price(Excludes ROC)

Net Income = Gross Revenue – Loss Penalty – O&M Cost

Return on Investment = Wind Farm Cost + ANM Cost x 100%

Net Income

8 - Results

Page 77: Connecting Renewable Generation with Active Network Management

Table 8.3 – Financial Results spreadsheet

WF Size Net Generation Generation Curtail Revenue Incr Loss Penalty ANM WF Cost Cap O&M Cost Total Net Invest Return

Aka Wind Inj Grp Corrected (Corrected) gross area loss Cost (thou £/MW) (thou £ / MW) Exp Income Cost

90 48 800 32 (thou'£) %

(MW) (MWh/yr) (MWh/yr) (thou £'s) (thou £'s) MWh/yr (thou £'s) (thou'£) (thou £) (thou £) (thou £)

2 5434 0 0 489 -103 -5 0 1600 1600 64 59 430 0.29 27

PF 098 4 10871 -1 0 978 -77 -4 0 3200 3200 128 124 854 0.29 27

6 15930 375 34 1434 47 2 0 4800 4800 192 194 1239 0.30 26

8 19026 2713 244 1712 114 5 0 6400 6400 256 261 1451 0.34 23

10 21127 6047 544 1901 151 7 0 8000 8000 320 327 1574 0.38 20

12 22921 9688 872 2063 204 10 0 9600 9600 384 394 1669 0.42 17

14 24421 13623 1226 2198 250 12 0 11200 11200 448 460 1738 0.46 16

16 25569 17910 1612 2301 290 14 0 12800 12800 512 526 1775 0.50 14

18 26739 22174 1996 2407 338 16 0 14400 14400 576 592 1814 0.54 13

20 27643 26706 2403 2488 389 19 0 16000 16000 640 659 1829 0.58 11

2 5434 0 0 489 -97 -5 2 1600 1602 64 59 430 0.29 27

PF_095 4 10871 -1 0 978 -58 -3 2 3200 3202 128 125 853 0.29 27

6 16305 -1 0 1467 109 5 2 4800 4802 192 197 1270 0.29 26

8 20357 1382 124 1832 276 13 2 6400 6402 256 269 1563 0.31 24

10 23211 3963 357 2089 387 19 2 8000 8002 320 339 1750 0.34 22

12 25721 6888 620 2315 507 24 2 9600 9602 384 408 1907 0.37 20

14 27852 10192 917 2507 653 31 2 11200 11202 448 479 2027 0.40 18

16 29348 14131 1272 2641 755 36 2 12800 12802 512 548 2093 0.44 16

18 30822 18092 1628 2774 832 40 2 14400 14402 576 616 2158 0.47 15

20 32201 22147 1993 2898 908 44 2 16000 16002 640 684 2215 0.50 14

2 5434 0 0 489 -48 -2 25 1600 1625 64 62 427 0.30 26

LDC 4 10871 -1 0 978 -22 -1 25 3200 3225 128 127 851 0.30 26

6 16305 -1 0 1467 122 6 25 4800 4825 192 198 1270 0.30 26

8 21384 355 32 1925 355 17 25 6400 6425 256 273 1652 0.30 26

10 24211 2963 267 2179 449 22 25 8000 8025 320 342 1837 0.33 23

12 26494 6115 550 2384 538 26 25 9600 9625 384 410 1975 0.36 21

14 28740 9304 837 2587 650 31 25 11200 11225 448 479 2107 0.39 19

16 30520 12958 1166 2747 742 36 25 12800 12825 512 548 2199 0.42 17

18 32006 16907 1522 2881 830 40 25 14400 14425 576 616 2265 0.45 16

20 33378 20970 1887 3004 901 43 25 16000 16025 640 683 2321 0.48 14

2 5434 0 0 489 -48 -2 50 1600 1650 64 62 427 0.30 26

LDC + VB 4 10871 -1 0 978 -20 -1 60 3200 3260 128 127 851 0.30 26

6 16305 -1 0 1467 128 6 70 4800 4870 192 198 1269 0.30 26

8 21739 0 0 1957 410 20 80 6400 6480 256 276 1681 0.30 26

10 27174 0 0 2446 800 38 90 8000 8090 320 358 2087 0.30 26

12 32609 0 0 2935 1303 63 100 9600 9700 384 447 2488 0.30 26

14 38044 0 0 3424 1917 92 110 11200 11310 448 540 2884 0.30 25

16 43469 10 1 3912 2629 126 120 12800 12920 512 638 3274 0.30 25

18 48011 902 81 4321 3256 156 130 14400 14530 576 732 3589 0.30 25

20 51442 2906 262 4630 3764 181 140 16000 16140 640 821 3809 0.31 24

Cost Curt (£

(£/MWh) (£/MWh) Inv

(thou'£/MWh)

Page 78: Connecting Renewable Generation with Active Network Management

The values used as parameters in the finance calculations and the costs used for implementing the

ANM solutions are indicated in the table 8.4 and 8.5 below.

Parameter Value Comment

Initial Capital Investment £800,000 / MW DG capacity

O & M 4% of Initial Capital Investment

[32,ch.8]

£32,000/MW capacity using

above.

Electricity Sell Price £45 + ROC at £45 = £90/MWh Opportunity cost of curtailment

is same

DNO Loss penalty £48/MWh Note ROC issue

Table 8.4 – Parameters for financial analysis

ANM Case Cost Comment

Passive None

PF 0.95 £2,000 Flat rate Administration

LDC £25,000 Flat rate Includes O&M

Voltage Booster & LDC £40,000 + £5,000/MW Gives £50-140k for 2-20MW DG

Table 8.5 – Upfront capital cost for ANM solutions

8.3.2 Cash Flows

An overall view of the financial performance of the DG can be seen by looking at the ROI which is

shown in figure 8.10.

Passive Voltage Control

This case shows a marked drop in ROI after 6MW of DG capacity. This is as a result of reduction of

revenue from electricity sale due to curtailment. In addition the O&M costs become a relatively

larger proportion of costs. At 2MW DG O&M costs are £64,000 for £430,000 of electricity sales. At

20MW DG O&M costs are £640,000 for £1.83 million of electricity sales.

PF 0.95 case and LDC case

This is much the same as the passive voltage control base case. Slightly increased revenues are

obtained through marginally reduced curtailments giving approximately 3% better ROI.

8 - Results

Page 79: Connecting Renewable Generation with Active Network Management

Voltage Booster & LDC

This case shows significant improvements at higher DG capacities due to the almost non-existence

of curtailment. This clearly demonstrates the significance of the value of the electricity produced.

At 10MW DG capacity the electricity revenue is approximately two million pounds per year. The

up-front capital investment required for the ANM solution of £90,000 is only a small proportion of

this which would indicate that the Voltage Booster is a very cost effective method for increasing DG

connection capacity whilst maintaining a high ROI.

Fig. 8.10 – Return on investment

General Observations

Sizing of the wind farm to minimise network losses in all of the cases would show little financial

benefit. This is because savings in losses are small compared with the overall income stream

received from electricity exported. For example the loss penalties and electricity income varies for

the whole set of cases from -£5,000 penalty with £430,000 income at 2MW DG in the passive

voltage control case to +£181,000 penalty with £3.8 income at 20MW DG in the Voltage Booster

with LDC case.

The return of 25% on an investment would seem an attractive value. For wind farm sizes between

zero and six no ANM solution is required. For between six and eight MW the PF 0.95 and LDC

ANM solutions offer increased financial returns. For over nine MW the Voltage Booster ANM

solution is optimum.

8 - Results

2 4 6 8 10 12 14 16 18 20

0

5

10

15

20

25

30

35

40

Return on Investment

PF 098 PF_095 LDC LDC + VB

Wind Farm Size (MW)

(%)

Page 80: Connecting Renewable Generation with Active Network Management

9 - Discussion

Relevance

The simulation detailed in chapters seven and eight demonstrates the problem where a number of

feeder voltages are affected by the control of a single main bus at a substation. This would be a

fairly common circuit topology on rural distribution networks which is where much wind generation

resources exist.

9.1 Curtailment Results

Effectiveness of ANM

The four ANM cases for controlling voltage all show improvements in connection capacity above a

passive non-curtailment approach. The reactive power control and area based control showing

similar modest increases in pre-curtailment capacity levels of around 2MW allowing approximately

8MW of unconstrained capacity onto the network (see section 8.1.1). Obviously the reactive power

control method would work proportionally better on DG connected through circuits with higher

values of line reactance as equation 4.14 shows. The Voltage Booster method allows a considerable

increase from the passive non-curtailment level of 6MW to 18 MW.

Optimum Wind Farm Size

Larger connection capacities with levels of curtailment of 10% and above are possible but the

increase in yearly exported power is rapidly decreases. Hence financially and in terms of an efficient

use of resources operation at these high levels of curtailment would be unlikely. Increases in

generation output at less than peak wind speeds by having a greater installed capacity does not

appear to work for the case studied.

This might be intuitively explained by considering the equation of power output for a wind turbine

P = �Cp�V�A [11,pg.31]

This states that the power output 'P' of a turbine is proportional to the cube of the wind 'V'. Hence it

could be speculate that a large proportion of the energy produced will be at relatively high wind

speeds when the turbine is closer to maximum output. Week two of the wind profile input (figure

7.15) in particular exhibits this characteristic.

Estimation of Connection Capacity

As mentioned in 8.1.2 the simpler and cruder steady state analysis reveals similar connection

capability capacities for the different ANM cases as the TSS method. Ball park connection figures

stated in Table 6.X give 6.5-10 MVA as an approximate DG connection capacity limit for 15-20kV

networks. This is reasonably close to the 6MW passive voltage control base case DG limit

approximated in 7.3.3.

9 - Discussion

Page 81: Connecting Renewable Generation with Active Network Management

9.1.1 Comparison to Book Results

Figures 9.1 to 9.4 show the results for generation curtailment from the book.

For all cases a basic difference in the capacity factor is observed. For the book the maximum

generation that can occur with no curtailment and 20 MW generation is approximately 37,500

MWh per year. For the simulation done here the maximum is approximately 55,000 MWh per year.

This would indicate the books capacity factor of being 0.68 of what was used in the simulation here

(see Table 7.6).

0.68 x 31% = 21%

Hence a more general comparison of results is appropriate as some differences due to the wind

profile used will be expected.

Passive Voltage Control

Book : Curtailment begins at 6MW and generation output flattening off to a maximum at 14 MW of

DG.

PowerWorld : Curtailment also begins at 6MW with generation reducing thereafter but does not

reach an ultimate maximum.

PF 0.95

Book : Curtailment begins at 8MW with generation output reducing but still increasing generation

output at higher DG.

PowerWorld : Same shape curtailment curve as book.

Area Based Voltage Control (LDC)

Book : A very large capacity of 16MW is reached before curtailment occurs.

PowerWorld : A small improvement over the PF 0.95 case is seen however the same curtailment

profile occurs. The huge improvement in as in the book case is not seen.

Voltage Regulator (Booster) with LDC

Book : A marginal improvement over the previous Area Based Voltage control is seen. Virtually no

curtailment occurs.

PowerWorld : Very little curtailment occurs. This compares well with the book case.

9.1.2 Explanation of Differences

Apart from the wind profile capacity factor difference a number of other possible reasons to do with

the simulation setup parameters are presented below.

9 - Discussion

Page 82: Connecting Renewable Generation with Active Network Management

Fig. 9.1 – PF 0.98 book case (Appendix A2.1) Fig. 9.2 – PF 0.95book case (Appendix A2.1)

Fig. 9.3 – Area Based Voltage Control book case (Appendix A2.1) Fig. 9.4 – Area Based Voltage Control & Voltage Reg. book case (Appendix A2.1)

Page 83: Connecting Renewable Generation with Active Network Management

Variation in Reactive Loads

Book : Exact variation not stated but said to be less than real power variation.

PowerWorld : Reactive loads vary the same as Real Loads.

Bus 2 Voltage Regulation Target

Book : This is specified 1.0 Vpu for the passive voltage control cases.

PowerWorld : 1.005 +/- 0.005 was used.

Comment : For the PW case operating with Bus 2 at a slightly higher voltage will have reduced

available DG headroom by approximately 1.5 MW (see 7.3.3)

Area Based Voltage Control Method

Book : The exact method is unknown but would be expected to be capable of optimisation

PowerWorld : LDC control of Bus 4 was used as this seemed to be an optimum voltage control

strategy.

Voltage Regulator Operation

Book : Used a voltage regulator

PowerWorld : Used a Booster which only has one step.

Comment : Using a Booster instead of a Regulator results in less than optimum voltage control (see

section 7.3.2)

9.1.3 Effect of Wind Profile

Possible effects on results may occur depending on the profiles of load and demand and their

interaction. Typical wind generation capacity factors are shown below in figure 9.5.

Fig. 9.5 – Monthly capacity factors of typical UK wind farms [11, pg.36]

Table 7.6 reveals that the capacity factor used in the PW simulation is unrealistically high for the

9 - Discussion

Page 84: Connecting Renewable Generation with Active Network Management

'Spring' week and unrealistically low for the 'Winter week'. This may have implications for

interaction between load and demand which is dealt with in the next section.

9.1.4 Coincidence of Demand and Generation

For all of the 672 timepoints total network demand and WF generation was 'binned' into four ranges

according to its normalised value. The number of occurrences of the combinations of these values is

tabulated below in table 9.1.

Table 9.1 – Incidence of total Network Demand and WF generation

The most common occurrence is when demand lies between 50 and 75% with WF output lying

between 0 and 25%. Intuitively for balancing of a power system on a network it would be expected

that coincidence of high demand with high output would be most beneficial. However for the case

scenario here DG penetration is increased when conditions of high DG output coincide with low

load. The coincidence results show that the most common occurrence was high demand with low

generation.

9.2 Losses

9.2.1 Comparison to Book

Yearly Losses without DG

For the book yearly losses without DG are 2860 MWh and as stated in section 8.2.1 for the PW

simulation are 3439 MWh. This indicates that the basic circuit parameters and load profiles for both

simulations approximately match.

Losses for ANM cases

Figure 9.6 below shows losses for the four ANM cases including a case for use of an SVC. Although

DG capacities of zero and 2MW are not shown a characteristic 'U' shape would be expected with

incremental losses returning to zero for no DG. This basic shape is present in the PW results shown

in figure 8.8. Peak incremental losses of approximately 2250 MWh for the book compare to 3764

MWh for the PW case (see 8.2.2). This difference may be due to the increased general losses which

the PW case shows and also it could be due to a higher impedance being used for the Voltage

Booster.

9 - Discussion

76-100 ��� �� �� ��

51-75 ��� � � ��

Demand % 26-50 �� � �

0-25

0-25 26-50 51-75 76-100

Gen Output %

Page 85: Connecting Renewable Generation with Active Network Management

Regarding the 'break even point' in terms of loss reduction due to DG the book has a value of

approximately 10MW whilst the PW case has a value of approximately 5MW. A portion of this

difference likely comes from the fact that the PW case uses DG with a higher capacity factor and

hence a generally higher level of output and therefore losses will occur for each level of DG

capacity.

Fig. 9.6 – Losses for various voltage controls for book case (Appendix A2.1)

9.3 Finance

Since an NPV analysis was not done for the PW case an exact comparison cannot be made with the

financial results presented in the book case figure 9.7 below.

However general observations are that for the non voltage regulator/booster cases profitability drops

off as the DG size grows past optimum curtailment.

Fig. 9.7 – NPV analysis from book case (Appendix A2.1)

9 - Discussion

Page 86: Connecting Renewable Generation with Active Network Management

9.4 Overall Sensitivities

For the particular network scenario studied the Bus 2 voltage setting (see 7.3.3) when under non-

LDC control has a large effect on the DG connection capacity. Use of relatively tight voltage limits

on the network of +/- 3% must be partly responsible for this sensitivity.

The sensitivity of the curtailment results to variations in circuit parameters and simulation settings

is also evidenced by a previous set of TSSs done with different simulation and circuit settings. (See

Appendix A1.6)

9.5 Suggested Incremental Improvements

In order to improve accuracy and for further validation of the PW simulation capabilities a number

of refinements and improvements are suggested in table 9.2 below. Some of these suggestions could

be implemented for the purposes of investigating connection of DG to distribution circuits in

general.

Improvement Comment Purpose

Sensitivity Analysis

Tap step settings &

OLTC target voltage tolerance

Move DG

Change load

Try DG on Bus 5 or 4

Add bigger load to Bus 5 or add load

to 6

Determine realistic settings for Bus 2

regulation and investigate how Area

Based Voltage control might be

implemented

General Investigation of DG

connection.

“ “ “

Use wind data more typical Try swapping the individual week

positions in the series

More realistic and using a profile

with 21% should results similar to

the book.

Use real wind data and a turbine

model

Relatively simple Can apply analysis to a particular

geographical location

Extend data to use 8760 Relatively easy. Will take extra 10min

per run I.e 44 x 10min =

8hr extra + 8hr = 16hrs total

Improved accuracy. Note that this

might be more applicable if a more

automated simulation procedure was

possible.

Try more ANM scenarios

Voltage Regulator

SVC

Line Upgrade

Compare results to book case

Use less variable Reactive loads I.e for example use 0.6/1.0 for

reactive min/max ratio instead of

0.4/1.0

More realistic.

(Could try 0.1 or 0.2/1.0 for real load

also)

Table 9.2 – Suggested incremental improvements to simulation

9 - Discussion

Page 87: Connecting Renewable Generation with Active Network Management

9.6 Simulation Capabilities

Procedural

In terms of running simulations and visualising the results PW was easy to use with graphical real-

time feedback and on-screen controls which could dynamically adjust input parameters. This

enabled easy development of the initial circuit and the later the application of the ANM cases.

For the TSSs the large proportion of time was spend moving data from the relevant results tabs into

the spreadsheets.

Processing of the results in the spreadsheet was simple although the final version of the spreadsheet

was fairly complex and modification of this might lead to some debugging being necessary.

Accuracy

Open-loop implementation of curtailment modelling in PW was checked in the spreadsheet and

appeared acceptable. Comparison with the book results were also reasonably close which verifies

the overall simulation setup as being usefully accurate. Some room for introducing errors exists

when transferring data back to the spreadsheet.

Modelling of ANM

Development of the curtailment method was initially difficult but with the kind help of technical

support from PW in the U.S a solution was found using some more advanced functionality in PW

using 'post flow actions'. Application of an SVC and the Voltage Regulator was not obviously easy.

A compromise Voltage Booster was instead used which gave similar results.

Shortcomings

PowerWorld is not a simulation tool which is targeted to modelling of distribution networks rather

its main application is for transmission systems and operation of centralised generation(see

Appendix A3.3). Hence its use here to model DG connection on a distribution network is novel and

some difficulties were expected.

Use of the scripting and .COM interface additional software called AutoSim which can be purchased

from PW would enable greater control of the underlying simulation engine in PW and the

possibility of implementing more complex control strategies . Obviously this would have a attendant

learning curve associated with it.

9 - Discussion

Page 88: Connecting Renewable Generation with Active Network Management

9.7 Future Aspects of DG Connection

Since DG and ANM are a relatively new area in the electrical power systems industry a few topics

of interest are worth mentioning which nay affect DG and the use ANM in the future.

Operation

Currently DG connected to the network has no value of contribution to security attached to it by the

Network operators. Maintenance of adequate levels of security are mandatory and greater

penetration of DG in the future may mean a level of contribution of security being apportioned to

them. For a discussion of this topic see [11, ch.8].

Participation in ancillary service markets for supply of reactive power and responsibility for voltage

control on the network may be other services which DG can take part in.

Increasing penetration of DG from various intermittent and non-intermittent generators may lead to

increased variance and balancing problems on the network. Heat led CHP in Denmark is an example

of high levels of penetration of such DG. Advances in DSM or storage for modifying the demand

and supply curve may have a role to play in this balancing.

Planning

The analysis done for the case scenario in chapters 7 and 8 demonstrate the complexities involved in

trying to optimise DG connection capacity without resorting to traditional network upgrade

solutions. Use of strategic planning of generation as reviewed in chapter 6 would be expected to

have a large beneficial effect on potential DG penetration and its incremental addition.

Another area of interest is how does ANM fit in with projected infrastructure plans and asset

replacement. Use of ANM might enable a 'leaner' network in terms of traditional plant

infrastructure, less 'heavy' plant items likely having a smaller associated carbon footprint.

Monitoring of Progress

Current use and development of ANM can be seen by viewing a register which is maintained by the

Electrical Network Steering Group [26]

9 - Discussion

Page 89: Connecting Renewable Generation with Active Network Management

10 - Conclusion

10.1 Summary of main findings

Research review and connection solutions

A simple analysis of the common voltage rise problem revealed that this was due equally to the

contributions of firstly real power flow and resistance and secondly reactive power flow and

reactance. This principle leads to an easy understanding of how voltage rise occurs on networks and

principle methods for its reduction.

The early work done at the research review stage indicated the breadth of ANM solutions from

individual power system component level of generator exciters through to optimum planning and

system distributed 'agent control'. Requirements stated for future intelligent area based control can

be seen reflected in issues highlighted in chapter 6.

Various connection solutions were reviewed which used a number of types of voltage control

techniques. However literature containing survey responses from DNO operators [23]&[33]

appeared to show complications by many commercial issues limiting the application of ANM to the

distribution networks. Note only three Registered Power Zones currently exist in the UK.

Simulations

PW was found relatively easy to learn and modelling of curtailment was successfully developed and

reasonably accurate. Four ANM solutions were evaluated using a simple steady state scenario and a

set of more complex Time Series Simulations were ran also.

Use of a case scenario presented in a book showed PowerWorld capable of evaluating the

application of four basic ANM solutions. Subject to some sensitivities on circuit topology and

simulation setup inputs an optimum ANM solution was found using the Voltage Booster. Network

losses did not appear to affect the financial viability of any of the the ANM solutions for the case

scenario. The main factor was level of curtailment. Over-sizing the Wind Farm to obtain larger

outputs at lower wind speeds was seen not to work.

10.2 Contribution to subject field

The simulation case undertaken demonstrates the three key interrelated concerns of curtailment,

losses and finance for DG connection.

Simulation of a previously published ANM case scenario example resulted in a relatively easy

development path for use of PowerWorld in the DG/ANM problem domain. Use of PowerWorld was

thought innovative as there doesn't appear to be any alternative software tools available which have

a reasonably quick learning curve for use on a such a short project.

In a non-automated format the simulation setup as it stands would not likely be robust enough for

circuit analysis in commercial connection studies as mistakes through user operation may result and

also the complete simulations take a day to run.

PW proved effective for visualising and quantifying connection issues and hence evaluating an

optimum connection solution using ANM for this case scenario.

10 - Conclusion

Page 90: Connecting Renewable Generation with Active Network Management

10.3 Possible further work

A number of different investigations might be relevant for future work.

Analysis with Different Network Case

An example of another case scenario which could modelled is provided in [29,pg.23] . Also the

UKGDS EHV-1 ANM circuit (see Appendix A2.2) might be considered as well as further

evaluation of the UKGDS AMP tool itself to ascertain its pertinent features.

Different ANM Methods

Investigation of Area Based Voltage Control or even hybrid AVPFC if PowerWorld simulation

capabilities permitted.

Financial Modelling

Exploration of the operational modelling capabilities of PW might be useful for investigation of

how future DG could participate in ancillary services or balancing markets.

Scripting& other capabilities

Scripting and use of the .COM interface in the AutoSim add-on program will likely reveal much

more advanced functionality and hence allow more sophisticated and optimum analysis of case

studies. Exploration of the fault modelling capabilities may also be useful.

An interesting feature of PowerWorld in its standard demo version are the 'make-up' loads feature

which might have application in DG balancing and curtailment modelling for any future scenario of

energy storage.

10.4 Overall evaluation of the project

Project objectives stated in section 2.2 were satisfactorily achieved.

As the project was initially exploratory in nature the successful development of the PowerWorld

simulation functionality and application to a published network case scenario was thought highly

successful.

10 - Conclusion

Page 91: Connecting Renewable Generation with Active Network Management

11 - Bibliography

[1] Distributed generation: a definition Authors: Ackermann T.1; Andersson G.; Soder L.

Source: Electric Power Systems Research, Volume 57, Number 3, 20 April 2001 , pp. 195-204(10)

Publisher: Elsevier

[2] UK research activities on advanced distribution automation, Ault, G.W.; Foote, C.E.T.; McDonald,

J.R.;Power Engineering Society General Meeting, 2005. IEEE 12-16 June 2005 Page(s):2616 - 2619 Vol. 3

[3] B. M. Weedy and B. J. Cory, Electric Power Systems, 4th ed. Chichester, U.K.: Wiley, 1998

[4] http://www.powerworld.com/

[5] http://pscad.com/

[6] http://monaco.eee.strath.ac.uk/ukgds/

[7] http://www.siemens.com/

[8] D.F.Warne, Electrical Power Engineers Handbook, 2nd

edition, Oxford UK: Newnes 2005

[9] Network voltage controller for distributed generation ; Hird, C.M.; Leite, H.; Jenkins, N.; Li, H.

Generation, Transmission and Distribution, IEE Proceedings-

Volume 151, Issue 2, 2 March 2004 Page(s): 150 - 156 Glover

[10] C. L. Masters “Voltage rise: The big issue when connecting embedded generation to long 11 kV

overhead lines,” Power Eng. J., vol. 16, pp. 5, Feb. 2002.

[11] Jenkins, N., Allan, R., Crossley, P., Kirschen, D., Strbac, G.; ‘Embedded Generation’; IEE, Stevenage,

UK, 2000

[12] Impact of wind generation on the operation and development of the UK electricity systems – Goran

Strbac et al – ScienceDirect – 17 Oct 2006

[3] A. E. Kiprakis and A. R. Wallace “Maximising energy capture from distributed generators in weak

networks,” Proc. Inst. Elect. Eng., Gen., Transm., Distrib., vol. 151, pp. 611, Sep. 2004.

[14] Centralized and Distributed Voltage Control: Impact on Distributed Generation Penetration

Vovos, P.N.; Kiprakis, A.E.; Wallace, A.R.; Harrison, G.P.;

Power Systems, IEEE Transactions on

Volume 22, Issue 1, Feb. 2007 Page(s):476 - 483

[15] http://www.2ndintay Egrationconference2008.com/posters.asp Downloaded 26 M Robert Currie, Institute

for Energy and Environment, University of Strathclyde, United Kingdom, Active Power Flow Management to

Facilitate Increased Connection of Renewable and Distributed Generation to Rural Distribution Networks

[16] Active power-flow management utilising operating margins for the increased connection of distributed

generation; Currie, R.A.F.; Ault, G.W.; Foote, C.E.T.; McDonald, J.R.;

Generation, Transmission & Distribution, IET

Volume 1, Issue 1, January 2007 Page(s):197 - 202

11 - Bibliography

Page 92: Connecting Renewable Generation with Active Network Management

[17] Exploring the Tradeoffs Between Incentives for Distributed Generation Developers and DNOs

Harrison, G. P.; Piccolo, A.; Siano, P.; Wallace, A. R.;

Power Systems, IEEE Transactions on

Volume 22, Issue 2, May 2007 Page(s):821 - 828

[18] P. N. Vovos, G. P. Harrison, A. R. Wallace, and J. W. Bialek, “Optimal

power flow as a tool for fault level constrained network capacity analysis,” IEEE Trans. Power Syst., vol. 20,

no. 2, pp. 734–741, May 2005.

[19] Applying Time Series to Power Flow Analysis in Networks With High Wind Penetration

Boehme, T.; Wallace, A.R.; Harrison, G.P.;

Power Systems, IEEE Transactions on

Volume 22, Issue 3, Aug. 2007 Page(s):951 - 957

[20] A. Wallace and G. Harrison, “Planning for optimal accommodation of

dispersed generation in distribution networks,” in Proc. CIRED 17th Int.

Conf. Elect. Distrib., Barcelona, Spain, May 2003.

[21] Multi-Agent Systems for Power Engineering Applications—Part I: Concepts, Approaches, and Technical

Challenges

McArthur, S.D.J.; Davidson, E.M.; Catterson, V.M.; Dimeas, A.L.; Hatziargyriou, N.D.; Ponci, F.; Funabashi,

T.;

Power Systems, IEEE Transactions on

Volume 22, Issue 4, Nov. 2007 Page(s):1743 - 1752

[22] Exploiting Multi-agent System Technology within an Autonomous Regional Active Network Management

System; Davidson, E.M.; McArthur, S.D.J.;

Intelligent Systems Applications to Power Systems, 2007. ISAP 2007. International Conference on

5-8 Nov. 2007 Page(s):1 - 6

[23] Current technology Issues and Identification of technical opportunities for active network management

(ANM); BERR 2008; Sinclair Knight Merz

[24] GenAVC Active local distribution network management fo r embedded generation ; DTI ; Econnect 2005

[25] A technical review and assessment of active network management infrastructures and practices ; DTI ;

2006 EA Technology

[26] Electricity Networks Strategy Group Downloaded 26th

May 2008

http://www.ensg.gov.uk/assets/anm_deployment_register_-_january_2008_-_final.pdf

[27] T.Ackermann, Wind Power in Power Systems, John Wiley & Sons, 2005

[28] Taking an active approach

Djapic, P.; Ramsay, C.; Pudjianto, D.; Strbac, G.; Mutale, J.; Jenkins, N.; Allan, R.;

Power and Energy Magazine, IEEE

Volume 5, Issue 4, July-Aug. 2007 Page(s):68 - 77

[29] Integration of operation of embedded generation and distribution networks. MCEE 2002 ; Strbac et al

[30] http://www.powerworld.com/Document

%20Library/version.130/Simulator13_Help_Printed.pdf ; 21 August 2008 11:36AM

[31] http://www.nationalgrid.com/uk/Electricity/Data/Demand+Data/ ; 21 August 2008 11:45 AM

11 - Bibliography

Page 93: Connecting Renewable Generation with Active Network Management

[32] Penche Celso; Layman's guidebook on how to develop a small hydro site ; European Small Hydropower

Association ; 1998 2nd Edition;

[33] Solutions for the connection and operation of distributed generation ; DTI ; EA Technology Ltd ;2003

[34]�http://www.aura-nms.co.uk/

[35]�http://www.supergen-amperes.org/

�����http://www.fenix-project.org/

Front cover pictures

[Large] http://www.provenenergy.co.uk – Example of a 'Wind Croft'

[Small Left] Distributed Intelligent Load Controllers(DILCs);Technical Guide Version 2.0 2007 Econnect ; Fig1

[Small Top Right] Typical PLC Installation

11 - Bibliography

Page 94: Connecting Renewable Generation with Active Network Management

Appendix A1 – Extra Simulation Information

A1.1 Steady State Circuit Screen-shots

A1.1.1 Original Case

The four combinations of load and generation for the passive voltage control case

� Minimum Load & Minimum Generation

� Minimum Load & Maximum Generation

� Maximum Load & Minimum Generation

� Maximum Load & Maximum Generation

are shown in the following four figures

Appendix - 1

Page 95: Connecting Renewable Generation with Active Network Management

Fig. A1.1 – Original case at Min-Load Min-Gen

Page 96: Connecting Renewable Generation with Active Network Management

Fig. A1.2 – Original case at Min-Load Max-Gen

Page 97: Connecting Renewable Generation with Active Network Management

Fig. A1.3 – Original case at Max-Load Min-Gen

Page 98: Connecting Renewable Generation with Active Network Management

Fig. A1.4 – Original case at Max-Load Max-Gen

Page 99: Connecting Renewable Generation with Active Network Management

A1.1.2 PF 0.95

The four combinations of load and generation for the PF 0.95 case

� Minimum Load & Minimum Generation

� Minimum Load & Maximum Generation

� Maximum Load & Minimum Generation

� Maximum Load & Maximum Generation

are shown in the following four figures

Appendix - 1

Page 100: Connecting Renewable Generation with Active Network Management

Fig. A1.5 – PF 0.95 case at Min-Load Min-Gen

Page 101: Connecting Renewable Generation with Active Network Management

Fig. A1.6 – PF 0.95 case at Min-Load Max-Gen

Page 102: Connecting Renewable Generation with Active Network Management

Fig. A1.7 – PF 0.95 case at Max-Load Min-Gen

Page 103: Connecting Renewable Generation with Active Network Management

Fig. A1.8 – PF 0.95 case at Max-Load Max-Gen

Page 104: Connecting Renewable Generation with Active Network Management

A1.1.3 LDC

The four combinations of load and generation for the LDC case

� Minimum Load & Minimum Generation

� Minimum Load & Maximum Generation

� Maximum Load & Minimum Generation

� Maximum Load & Maximum Generation

are shown in the following four figures

Appendix - 1

Page 105: Connecting Renewable Generation with Active Network Management

Fig. A1.9 – LDC case at Min-Load Min-Gen

Page 106: Connecting Renewable Generation with Active Network Management

Fig. A1.10 – LDC case at Min-Load Max-Gen

Page 107: Connecting Renewable Generation with Active Network Management

Fig. A1.11 – LDC case at Max-Load Min-Gen

Page 108: Connecting Renewable Generation with Active Network Management

Fig. A1.12 – LDC case at Max-Load Max-Gen

Page 109: Connecting Renewable Generation with Active Network Management

A1.1.4 LDC & Voltage Booster

The four combinations of load and generation for the Voltage Booster with LDC case

� Minimum Load & Minimum Generation

� Minimum Load & Maximum Generation

� Maximum Load & Minimum Generation

� Maximum Load & Maximum Generation

are shown in the following four figures

Appendix - 1

Page 110: Connecting Renewable Generation with Active Network Management

Fig. A1.13 – Voltage Booster & LDC case at Min-Load Min-Gen

Page 111: Connecting Renewable Generation with Active Network Management

Fig. A1.14 – Voltage Booster & LDC case at Min-Load Max-Gen

Page 112: Connecting Renewable Generation with Active Network Management

Fig. A1.15 – Voltage Booster & LDC case at Max-Load Min-Gen

Page 113: Connecting Renewable Generation with Active Network Management

Fig. A1.16 – Voltage Booster & LDC case at Max-Load Max-Gen

Page 114: Connecting Renewable Generation with Active Network Management

A1.2 – Crib Sheet

Simulation procedure for a single 4wk simulation run for one ANM case scenario.

PW.. Three files exist.

One-line

Simple profile TSS setup file

4wk profile TSS setup file

Spreadsheet.. 2 Files exist.

Input & Output spreadsheet

Collated Results spreadsheet

-------------------------------------------------------------

PW..

Load PW project file.

Modify one-line for chosen ANM case

Check : PFC cap on bus6 is ON / Generator capability curve is correct / Bus Reg setting /

Post Flow actions ON

Open TSS and load 4wk profile

Adjust Dump Load Scale Factor appropriately on one-line

Change Dump Load MVAr to match using Model Explorer

Spreadsheet..

Open I/O Spreadsheet, Choose WF size, Copy Gen Profile Data as indicated to clipboard

PW.. Paste into Generator column in TSS inputs, update WF Size in Aggregated Results worksheet

Clear Results / Reset Run

Hit Single Step Full-Newton Solution (This seems to ensure that one-line file on screen is

always used in the TSS run)

Run Simulation

Results/Buses/Right click on Vbus column..Do a quick plot of Voltage to check operation as

expected

Export all results into R_name worksheets

Spreadsheet..

Check aggregated results worksheets / also Results worksheet if curtailment error high

Copy and paste indicated values into 'Collated Results' Spreadsheet

Save the I/O spreadsheet to capture the detailed simulation results. Close file and re-open generic

I/O file (always make general improvements in this file)

Go back to I/O spreadsheet and adjust WF size and repeat the procedure. End..

Appendix - 1

Page 115: Connecting Renewable Generation with Active Network Management

A1.3 – Individual Week Wind Profiles

Appendix - 1

Page 116: Connecting Renewable Generation with Active Network Management

A1.4 - Simplified Generation & Load Profile

Fig. A1.21 - Basic test profile used in development

These profiles were used for initial testing of the curtailment modelling and also for familiarisation

with the TSS capability of PW

� Generation 0-6-0 , 0-12-0

� Min Max loads for 48 steps

Appendix - 1

L o a d s a n d D G p r o f i l e

0

5

1 0

1 5

2 0

2 5

3 0

3 5

1 5 9

13

17

21

25

29

33

37

41

45

B u s 2 M W

B u s 3 M W

B u s 4 M W

B u s 5 M W

B u s 2 M V A r

B u s 3 M V A r

B u s 4 M V A r

B u s 5 M V A r

G e n 6 # 1 M W

Page 117: Connecting Renewable Generation with Active Network Management

A1.5 – PW setup screenshots

Fig.A1.22 – Model Explorer dialogue for globally modifying dump load reactive compensation

Fig. A1.23 – Postflow Power Solution actions for Curtailment

Appendix - 1

Page 118: Connecting Renewable Generation with Active Network Management

Fig. A1.24 – Transformer setup showing WF Voltage Booster settings

Fig. A1.25 – Bus 2 Transformer Control Setup Dialogue

Appendix - 1

Page 119: Connecting Renewable Generation with Active Network Management

A1.6 – Curtailment results for previous simulation

Differences between this simulation and that presented in the main report are for this case

� LDC regulation set at 1.015 Vpu on bus 2 not 1.005 Vpu

� X and R values for Bus 5-6 reversed in the passive voltage control cases

� +0.1 MVAr compensation capacitor not present

Fig. A1.26 – Previous Curtailment results for different LDC setting

Appendix - 1

Page 120: Connecting Renewable Generation with Active Network Management

A1.7 – Loss Results and one-line diagram using reversed X/R

Fig. A1.27 – Circuit losses with resistance and reactance swapped

Appendix - 1

L o s s e s A N M _ 0 4

0

1

2

3

4

5

6

0 1 : 0 0 : 0 0

0 3 : 0 0 : 00

0 5 : 0 0 : 00

0 7 : 0 0 : 0 0

0 9 : 0 0 : 00

1 1 : 0 0 : 00

1 3 : 0 0 : 0 0

1 5 : 0 0 : 0 0

1 7 : 0 0 : 00

1 9 : 0 0 : 0 0

2 1 : 0 0 : 00

2 3 : 0 0 : 00

0 1 : 0 0 : 0 0

0 3 : 0 0 : 00

0 5 : 0 0 : 00

0 7 : 0 0 : 0 0

0 9 : 0 0 : 00

1 1 : 0 0 : 00

1 3 : 0 0 : 0 0

1 5 : 0 0 : 00

1 7 : 0 0 : 00

1 9 : 0 0 : 0 0

2 1 : 0 0 : 00

2 3 : 0 0 : 00

1 T O 2 C K T 1

M v a r L o s s

1 T O 2 C K T 1

M W L o s s

2 T O 3 C K T 1

M v a r L o s s

2 T O 3 C K T 1

M W L o s s

2 T O 5 C K T 1

M v a r L o s s

2 T O 5 C K T 1

M W L o s s

3 T O 4 C K T 1

M v a r L o s s

3 T O 4 C K T 1

M W L o s s

5 T O 6 C K T 1

M v a r L o s s

5 T O 6 C K T 1

M W L o s s

1 L o s s M v a r

1 L o s s M W

1 T O 2 C K T 1

T a p R a t i o

Page 121: Connecting Renewable Generation with Active Network Management

Fig. A1.28 – Circuit with Reactances and Resistances reversed

Page 122: Connecting Renewable Generation with Active Network Management

Appendix A3 – Project Errata

A3.1 – CD File Listing

PowerWorld

ANM_2_6.pwd Project files for main simulation runs

ANM_2_6.pwd

ANM_2_6.tsb

orig_anm_06_curt.tsb TSS File for simple ramp profile ANM_Curt (Useful for forcing

simulation to max or min load figures)

Excel

I/O File ANM_2_6.xls This is the master version which all result 'snapshots' are saved from

ANM_4wk_Results_2_6.xls Aggregated 4wk Results file ANM

V5_PF098 Folders with the 'snapshots of results' for the 11 simulation runs

V5_PF095

V5_LDC

V5_LDC_VBoost

Dissertation Document (Open Office)

*.odt Individual Chapters

*.pdf

master.odt Complete document (Note landscape pages are incorrect)

master.pdf Complete document (Note landscape pages are incorrect)

Wind Chapter Book 21

*.pdf

Appendix 3

Page 123: Connecting Renewable Generation with Active Network Management

A3.2 - Project Gantt Chart

Fig. X - Project Gantt Chart

Appendix 3

Page 124: Connecting Renewable Generation with Active Network Management

A3.3 - Email Dialogue with PowerWorld Support

Ian Moore(Author) and Santiago Grijalda (PowerWorld Support)

�����

������� ������������������ �������������������� ���������������������

� ����������� ���������������������������������������������������������������������������������������� ��� �������������������������� ������������������� ������������������ � ������������������������������������������������������ �!∀�������������#��������������������������������������!∀������������������������������#�!∀���∃�������������� �������������#�!∀����������� ���������������������������������������������� ������������������������������ ����������% ��������� ���&����������������������������� �������������∋�� ��#�(��������� !∀������������� ��� ���� �� �������������%�������������� ��)�������������∗ �����������������������������+��,�� ��������� ���������� ���������������−��������������������������������������∀������ �������������������������������������������������������������� �������� ���������&����!∀������������������������������� ����������������)��������������������������������+��

���������������������� ��� ��

������

����� !����������.�/�����0�� 1������ �2�

�����!�������3 ���45��#446�6047��!����������8�%��������9 ��:����1������ �� ������;<0�=��������������������� ���������� �� ���������������

Appendix 3

Page 125: Connecting Renewable Generation with Active Network Management

Hello,

Yes the University has a copy of SimAuto which I can use.

I have approximately 4/5 weeks left on the practical side of the project before I begin my write-up

which will take 3 weeks.

So I think it would unfortunately be too late to start development of a remote interface for simulation.

Regarding the simulation objectives for the project I think I will have easily surpassed them using Powerworld.

The curtailment is actually the most complex thing I am trying to implement and it would be brilliant if

I could accomplish this using the basic Powerworld functionality that I am now familiar with.

I realise that you may not have time to help me further with this but here is my progress on the problem below :

I have tried further combinations of post-powerflow actions but have had no progress.

Using 4 different model criteria checks (of essentially the same voltage for bus 6) and 4 actions to

reduce the generation on bus6 by -1MW each time I can only get the curtailment to take 2 correct iterations.

Using different combinations of Check,Postcheck and inserting 'Solve PowerFlow before proceeding' actions in

between seems to make little difference.

Except for the case of a single 'check' followed by a single 'postcheck' the model criteria checks

appear to be using the original powerflow solution value for the bus voltage.

Hence when an overvoltage occurs all the actions get exercised and the curtailment is too coarse.

e.g

a)

IF bus 6 gt 1.03 THEN reduce gen by -1MW CHECK

IF bus 6 gt 1.0301 THEN reduce gen by -1MW POSTCHECK

This works fine giving a 2-step reduction for large overvoltages and a single step reduction for small overvoltages

b)

IF bus 6 gt 1.03 THEN reduce gen by -1MW CHECK

Solve PowerFlow before proceeding

IF bus 6 gt 1.0301 THEN reduce gen by -1MW CHECK

Solve PowerFlow before proceeding

IF bus 6 gt 1.0302 THEN reduce gen by -1MW CHECK

Solve PowerFlow before proceeding

IF bus 6 gt 1.0303 THEN reduce gen by -1MW CHECK

This always results in 4 step reductions even for low overvoltages. Hence it appears to be still

using the bus voltage value from the original powerflow solution

c) Various combinations of POSTCHECK, CHECK and Solve PowerFlow before proceeding

Results in similar to b)

I will try and get some assistance from Janus Bialek who also works at Edinburgh University.

Thanks

Ian F Moore

0794 229 1688

0131 478 0228

Appendix 3

Page 126: Connecting Renewable Generation with Active Network Management

-----Original Message-----

From: Santiago Grijalva [mailto:[email protected]]

Sent: Thu 03/07/2008 20:04

To: Moore, Ian F

Cc: [email protected]

Subject: RE: Problem with remote regulation of bus using transformer

Dear Ian,

You may be able to achieve what you need with post power flow solution

actions and model conditions, but it may become tricky. Does your university

have a license of Simulator with SimAuto. SimAuto is the automation tool

ideal for that type of experimental research. This Simulator allows you to

create your own application in Matlab, C++, VB, Delphi, etc. from which you

can call all the routines of Simulator. With SimAuto it would be easy to set

up and application that does the following:

Open Case,

Scale Load

Solve Power Flow

If Voltage x > VY, then do Curtail Generation

While reactive reserve < 10 do begin

Solve Power flow

If x and y thenDo something,

Etc.

End

Etc.

You have all the flexibility to code your data structures and call Simulator

routines when you need them. All the data in the case and routines are

exposed so it is extremely flexible. Let me know if you already have this

add-on.

Santiago

From: Moore, Ian F [mailto:[email protected]]

Sent: Thursday, July 03, 2008 1:48 PM

Appendix 3

Page 127: Connecting Renewable Generation with Active Network Management

To: Santiago Grijalva

Cc: [email protected]

Subject: RE: Problem with remote regulation of bus using transformer

Hello,

The 'upper limit' I'm trying to regulate to is the overall network voltage

limits for the network study scenario of +/-3%.

A small explanation of basic characteristics of the 6 bus network I am

using :

1) Bus 4 suffers from being the lowest voltage point especially with high

output from the Bus6 Wind Farm

2) Bus 6 suffers from being the highest voltage point again when high

output from the wind farm occurs

It exhibits a characteristic 'K' shaped voltage profile.

Hence my voltage management strategy is to :

* use the OLTC to remotely regulate bus 4 at 0.97pu

* use curtailment of the (wind)generator on bus 6 to prevent bus

6 going above 1.03pu

This regulation scenario allows maximum connection capacity for DG (goal

of ANM) and could be described as real power generation 'curtailment' with

effectively conventional 'LDC' for bus2

Hope the above explains my reasoning a bit better.

I have followed your advice and implemented 'post flow solution actions'

Using a

'check' action of reducing the generator output by -20% when bus6 is

above 1.03

and also a

'post check' action of reducing the generator output by another -20%

when the bus voltage is still above 1.031

With this I have managed to effect a crude 2 stage regulation on the

generator output.

This represents a very simple curtailment which currently works with my

simple 'pyramid shape' wind generation profile steps of 2 MW increases or

decreases of 0-6-0 0-12-0.

Is it possible to refine this and have the post powerflow action

re-iterate itself continually reducing the gen output by -20% each time

until the bus6 over voltage condition is no longer violated ?

I've attached the files so you can see what I'm doing.

Appendix 3

Page 128: Connecting Renewable Generation with Active Network Management

P.s Thanks for the help. My supervisor here at University is very happy

with the project progress so far!

Thanks

Ian F Moore

0794 229 1688

0131 478 0228

-----Original Message-----

From: Santiago Grijalva [mailto:[email protected]]

Sent: Wed 02/07/2008 20:15

To: Moore, Ian F

Cc: [email protected]

Subject: RE: Problem with remote regulation of bus using transformer

HI Ian,

Some ideas below

Santiago

From: Moore, Ian F [mailto:[email protected]]

Sent: Wednesday, July 02, 2008 10:14 AM

To: Santiago Grijalva

Cc: [email protected]

Subject: RE: Problem with remote regulation of bus using transformer

Hello,

I've been away for a long weekend and so am now back on the job.

Today I have managed to get bus 4 regulating remotely and also all of the

other buses.

I admit to still being slightly confused. The only way that I can get the

TSS to recognise and 'sync'

to the bus I have chosen to regulate as indicated in the one-line is to hit

the 'single solution full newton'

button on the main run bar.

After I do this the log window indicates a tap change and all the TSS runs

work fine.

One issue is that the TSS will not give you all the message logs that single

solution. That is just to make the TSS faster when there are many hours.

I shall now continue to analyse my network with regard to trying to

simulate 'Active Network Management'.

Am I right in thinking that Powerworld is targeted to Transmission rather

than Distribution network analysis.?

Definetely is mostly used in very large systems.

Appendix 3

Page 129: Connecting Renewable Generation with Active Network Management

One of my next cases is to try and model Wind Generation on bus 6. Ideally

I'd like to 'curtail' the quantity

of generation when the network voltage reaches it's upper limit. Have you

any ideas how to do this?

I don't understand exactly what you mean by curtail generation. Active or

reactive power? Why would the voltages increase? What upper limit, the

regulation limit?

The OPF seems only to include thermal constraints. So I couldn't quite

figure out how to do this curtailment either via participation factors or

economic cost curves. Use pre/post scripts maybe?

You can do it using post-Power Flow Solution Options in Simulator Options ->

Power Flow Solution Page, -> Advanced Options and defining and model

condition where the action will occur.

My existing method is to ignore the generation when the voltage limit is

reached when I export my data out to

excel and simplistically deal with it there.

Thanks

Ian F Moore

0794 229 1688

0131 478 0228

-----Original Message-----

From: Santiago Grijalva [mailto:[email protected]]

Sent: Wed 25/06/2008 16:15

To: Moore, Ian F

Subject: RE: Problem with remote regulation of bus using transformer

Hi Ian,

OK, bus 4 then. This is the result am getting in the TSS, which shows that

actually the voltage at bus 4 is correctly regulated within the limits.

I attached the new TSB file. The previous one I had saved in v14 beta, which

I used in development, sorry about that.

Let me know if you get the same results.

Santiago

From: Moore, Ian F [mailto:[email protected]]

Sent: Wednesday, June 25, 2008 4:22 AM

To: Santiago Grijalva

Subject: RE: Problem with remote regulation of bus using transformer

Hello,

Thanks for the quick reply. I have followed your advice on making

Appendix 3

Page 130: Connecting Renewable Generation with Active Network Management

modifications to the base case

before opening the TSS Dialog. This makes sense and I experimented by

changing the value of the shunt

on bus 6 after opening the TSS and before so now understand how the current

settings are referenced.

However I still cannot get the remote bus to regulate. Instead nothing gets

regulated as the attached bmp

shows.

I have not made any changes to the default simulation options in the

general simulation or TSS options.

When I try to load your TSB file I get an error message 'Information - : 0,

0, not found in case'

This happens when choosing either of the options 'delete' or 'append'

existing data.

I'm running XP Professional with Service Pack 2, Powerworld Version 13

Build Apr 16,2008 (Evaluation).

The TSS inputs are fairly simple (just load and generation values) so do

not include any changes to the AVR settings on the xfrmr.

I am currently working from home, but will try the case on the University's

full version of Powerworld

tomorrow.

P.s Sorry my mistake, the bus I am actually trying to regulate is load bus

4 not 6. However the problem

is still present. I can get the remote bus 6 or 4 to regulate in a normal

simulation run but the TSS

run ignores the xfrmr AVR 'bus to regulate setting'.

Ian

Ian F Moore

0794 229 1688

0131 478 0228

-----Original Message-----

From: Santiago Grijalva [mailto:[email protected]]

Sent: Tue 24/06/2008 22:13

To: Moore, Ian F

Subject: RE: Problem with remote regulation of bus using transformer

Ian,

When you OPEN the time step simulation (TSS) tool, a system reference is

created with the current settings of the power system model. When you reset

the simulation this reference is restored. If you change data outside the

TSS while the TSS Dialog is open those changes will take place only if you

don't reset the simulation AND if the TSS input data does not change the

same values. You probably opened GEN 6 or changed its AVR setting or modify

the transformer control regulation while the TSS was open. Then resetting

Appendix 3

Page 131: Connecting Renewable Generation with Active Network Management

the TSS will override those changes.

I normally suggest making all the changes to the BASE case before opening

the TSS tool. In this manner, I was able to simulate the regulation of

voltage at bus 6. These are some results. I also attached is the TSB file

with the full results.

The set volt field in the remote regulated bus is black because the

transformer specifies a voltage regulation range. You will have a value

there for buses remotely regulated by a generator or continuous shunt. If

you have additional questions, let us know.

Santiago

Santiago Grijalva, Ph.D.

Senior Consultant and SW Engineer

PowerWorld

-----Original Message-----

From: Moore, Ian F [mailto:[email protected]]

Sent: Tuesday, June 24, 2008 2:40 PM

To: [email protected]

Subject: Problem with remote regulation of bus using transformer

Dear Sirs,

I am undertaking an MSc Dissertation project using PowerWorld to simulate

connection of Distributed

Generation using Active Network Management methods. I am using a recently

downloaded evaluation version.

I have successfully run a timestep simulation for the circuit detailed in

the attached PDF.

However I am now trying to control the voltage on remote load bus (no.6)

using the grid supply transformer.

From the main simulation window the voltage regulation functions as

expected, the xfrmr tapping up and

down to control the remote bus as I vary the generator output using the

manual increment control.

However when running my timestep series the bus fails to regulate.

Any ideas? The transformer shows a sensitivity of - 1.1 or similar. Also

there is an entry in the simulation/remotely regulated bus display although

no value is present here for the 'set volt'.

Thanks

Ian F Moore

0794 229 1688

0131 478 0228

_____

Appendix 3

Page 132: Connecting Renewable Generation with Active Network Management

Appendix A4 – Miscellaneous

A4.1 – Typical G59 Protection

Fig. X- Typical G59 protection suite for MV DG connection [173]

Appendix 4