of 112

Frequency Response From Wind Turbines

Published on November 2016 | Categories: Documents | Downloads: 28 | Comments: 0
100 views

This is a 1st year transfer report for a PhD at Prifysgol Caerdydd. (The final Thesis is now published at http://orca.cf.ac.uk/42939) Capability of MW scale wind turbines to participate in power system frequency response is investigated. Although this report does not include a completed literature survey, construction of a wind turbine test rig is detailed which will later be used to explore the effects on individual wind turbines when providing frequency response.

Written in Word and then opened in OpenOffice and converted to pdf. No apologies for any strange formatting!

Comments

Content


1
st
year Transfer Report
Frequency Response from Wind Turbines
Institute of Energy - Cardiff University
PhD Title : System Support from Demand and Generation
Flexnet Workstream – Smart, Flexible Controls
Supervisors : Dr. Janaka Ekanayake, Prof. Nick Jenkins
Student : Ian Moore
Date : 19
th
November 2009
i
Synopsis
Background
As of October 2009 the UK reached 4 GW of installed wind capacity. By 2012 it is expected
that there will be “12GW of wind schemes either operational, being built or already with
planning permission” [
1
].
Any future large installed capacity of wind power will need to provide essential system
ancillary services of which frequency response is one such service. Unlike conventional
synchronous based generating plant, modern wind turbines due to the nature of their power
electronic based interface do not, by default, provide inertial support during the initial stage of
a system frequency event.
Research Question
This PhD project focuses on the following two areas :
“What is the best method of implementing primary frequency response from the
planned future capacity of wind power ?”
“What are the effects on the wind turbine when providing this frequency response ?”
Planned Work
To address this question the planned work encompasses :
• Model based simulation of wind turbines at the power system level
• Construction of a wind turbine test rig for machine level investigation
• Experimental testing of system response and machine interaction
This report provides an overview of the initial work conducted in the first year which includes
a literature review, basic simulation and test rig construction.
Project Methodology
Desired system response
In order to ascertain what response is required from wind turbines it is necessary to review the
response which is provided by existing synchronous based generation plant. It seems reason-
able to assume that wind turbines will be required to emulate a response similar to the existing
response characteristic of synchronous plant. The current method of how frequency response
performance is maintained was reviewed.
A review of the UK grid connection code was undertaken. Existing intermittent generation
such as wind farms, classified as ‘Power Park Modules’ by National Grid Company, are not
required to provide frequency response although the capability must be present. It is noted
that sustained secondary frequency response would only be possible by operating a power
park at a sustained lower output during normal operation with an associated lower energy rev-
enue generation.
ii
Implementation of response
The report includes a review of the basic functioning of power electronic converter schemes
and common control methods for FPC and DFIG machines. These were found to be based
around vector and/or load angle control of generator-side and grid-side Voltage Source
Converters.
A simple scheme for primary frequency response by restoration of inertial response was
simulated. This was achieved by adding the negative rate of change of system frequency
processed via a 1
st
order delay, to the turbine torque setpoint. This method showed a
successful contribution to primary response by extraction of turbine kinetic energy.
Paper review discovered another author presenting an alternative method of implementing a
modified inertial response which maximised extracted kinetic energy.
The simulation undertaken used a high wind penetration scenario of 20GW of wind, operating
alongside 40GW of synchronous plant. All loads and generating plant used single lumped
equivalent models. Exploration of how a dispersed resource of wind farm capacity can best
provide primary frequency response will be investigated to include partitioning the wind tur-
bine capacity into a multi-machine model.
Effects on wind turbine machine
In addition to providing a response suitable for the power system at the wind farm point of
connection, the effects possible on the turbine, of implementing the frequency response must
be investigated. Any demanded increase in torque and hence rotor or stator currents, outside
of a turbines normal designed operating mode could result in unwanted mechanical stresses
and oscillations in the drive train or tower. The following areas will be investigated :
• Torque effects/oscillation
• Aerodynamic effects – may explore the possibility of using ‘GH BLADED’
• Converter ratings
Test Rig Development
This is now nearing final completion with voltage and current sensors at the assembly stage.
Basic inverter operation producing a sinusoidal wave by PWM switching at a low current and
voltage was achieved. Further refinements or problems which need resolving are :
Over-current shutdown - Suspected noise causing over-current shutdown is expected to be
rectified in a second revision of the gate driver board which is under construction for the gen-
erator side VSC.
Pendulum machine disturbance - Initial problems with unexplained shutdown of the DC Pen-
dulum machine controller were seen to disappear when the mains supply was fitted with a
smoothing reactor. However a small disturbance did once again occur (Nov 09) even with the
reactor fitted although this was not as severe are previously (machine continued to run).
AVR slew rate - The AVR component of the controller appears to have a slew rate limited in-
put which may cause problems in providing accurate voltage regulation of a synchronous gen-
erator when connected.
iii
Acknowledgements
I would like to thank my supervisors Professor Nick Jenkins and Dr. Janaka Ekanayake for
their assistance, direction and encouragement.
Also I would like to acknowledge Dr.Nolan Caliao and Ajith Tennakoon for their work in
initial testing and commissioning of the wind turbine test rig.
Additionally thanks are due to Denley Slade and the staff in the Electronics Workshop for
their skills and efforts constructing the test rig and assistance in design.
Finally thanks of course to the sponsors FLEXNET
iv
Contents
1 Introduction .............................................................................................................................. 2
1.1 Project Overview ............................................................................................................... 2
1.2 Motivation ......................................................................................................................... 2
1.2.1 Frequency Response Services .................................................................................... 3
1.2.2 Frequency Response from Wind Turbines ................................................................. 3
1.3 Objectives .......................................................................................................................... 4
2 Wind Turbine Generator Systems ............................................................................................ 5
2.1 Energy Extraction ............................................................................................................. 5
2.2 Modes of Operation .......................................................................................................... 7
2.3 WT design features .......................................................................................................... 8
2.3.1 Fixed speed versus variable speed ............................................................................. 8
2.3.2 Power Control ............................................................................................................ 8
2.3.3 Electro-mechanical coupling ...................................................................................... 9
2.3.4 Mechanical Coupling ................................................................................................. 9
2.3.5 Other design features ................................................................................................. 9
2.4 Generator Types ................................................................................................................ 9
2.4.1 Induction Generator ................................................................................................... 9
2.4.2 Doubly Fed Induction Generator ............................................................................. 10
2.4.3 Full Power Converter ............................................................................................... 11
2.5 Power Electronic Converter Fundamentals ..................................................................... 12
2.5.1 Converter Classification ........................................................................................... 12
2.5.2 Bridge Configurations .............................................................................................. 12
2.5.3 Square-wave Switching Scheme .............................................................................. 13
2.5.4 Pulse Width Modulation Switching Schemes .......................................................... 16
2.5.5 Back to Back Frequency Converter ......................................................................... 16
2.6 Electrical Machine Control ............................................................................................ 18
2.6.1 Load Angle Control Theory ..................................................................................... 18
2.6.2 Vector Control .......................................................................................................... 19
2.7 Wind Turbine Generator Control Schemes ..................................................................... 22
2.7.1 Full Power Converters ............................................................................................. 22
2.7.1.1 Generator Side Control ................................................................................... 22
2.7.1.2 Grid Side Control ............................................................................................ 25
2.7.2 Doubly Fed Induction Machines .............................................................................. 27
2.7.2.1 DFIG Control Scheme ................................................................................... 27
2.7.2.2 Other DFIG Control Schemes ......................................................................... 28
3 Connection Requirements & Response Capability ................................................................ 29
3.1 UK Requirements for grid connection ............................................................................ 29
3.1.1 General ..................................................................................................................... 29
3.1.2 Steady State Reactive Power and Voltage Control .................................................. 30
3.1.3 Fault Ride Through Capability ................................................................................ 31
3.1.4 Power System Stabiliser and Black Start Capability ............................................... 32
3.1.5 Frequency Response ................................................................................................ 32
3.1.6 Reserve ..................................................................................................................... 34
3.2 Desired Response ............................................................................................................ 35
3.2.1 NGC Benchmarking of Plant FR capability ............................................................ 35
3.2.2 Performance of Synchronous Plant .......................................................................... 35
3.2.3 System Requirements for FR ................................................................................... 36
3.3 Primary Response Capability from WTs ........................................................................ 37
v
3.3.1 Primary Frequency Response Schemes ................................................................... 37
3.3.2 Machine Effects & Converter Current Limits .......................................................... 41
3.4 Secondary Response Capability from WTs .................................................................... 43
3.4.1 Wind Farm Control Scheme .................................................................................... 43
3.4.2 Frequency Controller ............................................................................................... 44
4 Modelling ............................................................................................................................... 45
4.1 Introduction ..................................................................................................................... 45
4.1.1 Equations of Motion ................................................................................................. 45
4.2 System Model ................................................................................................................. 47
4.2.1 Synchronous Plant Response ................................................................................... 48
4.2.2 Reduced Order Machine Model ............................................................................... 48
4.3 Control Scheme ............................................................................................................... 49
4.4 Simulink Model ............................................................................................................... 49
4.5 Setup ................................................................................................................................ 50
4.6 Results ............................................................................................................................. 50
4.6.1 Open Loop Wind Turbine Response ........................................................................ 50
4.6.2 Closed Loop Wind Turbine Response ..................................................................... 52
4.6.3 Wind Turbine and Synchronous Response .............................................................. 52
4.7 Simulation setup and results summary table ................................................................... 53
4.8 Discussion of results ....................................................................................................... 55
5 Experimental Wind Turbine Test Rig .................................................................................... 56
5.1 Overview ......................................................................................................................... 56
5.2 Design ............................................................................................................................. 58
5.2.1 DC Motor/ Pendulum Motor – Block A .................................................................. 58
5.2.2 Generator Machine – Block B ................................................................................. 60
5.2.3 Back to Back PWM converters – Block C & D ....................................................... 62
5.2.4 Generator and Grid Side Controller – Block E & F ................................................. 66
5.2.5 Power System – Block G ......................................................................................... 68
5.2.6 General Assembly and Connection .......................................................................... 68
5.2.7 Setup ......................................................................................................................... 69
5.3 Results ............................................................................................................................. 69
5.3.1 Measurement and Open Loop Control Test ............................................................. 69
5.3.2 Bridge Inverter Test ................................................................................................. 72
6 Further Work .......................................................................................................................... 75
6.1 Risks ................................................................................................................................ 76
6.2 Gantt Chart ...................................................................................................................... 76
7 Appendices ............................................................................................................................. 78
7.1 Simulations ...................................................................................................................... 78
7.1.1 Simulation Baseline Record ..................................................................................... 78
7.1.2 Setup ......................................................................................................................... 79
7.1.3 M-Files ..................................................................................................................... 80
7.1.4 Model Parameters .................................................................................................... 87
7.2 Laplace Transformation .................................................................................................. 89
7.3 Experimental Test Rig .................................................................................................... 91
7.3.1 Procedure for Use ..................................................................................................... 91
7.3.2 Equipment Specifications ........................................................................................ 92
.......................................................................................................................................... 97
7.3.3 Hardware Design ...................................................................................................... 98
8 References ............................................................................................................................ 104
vi
List of Abbreviations
ADC – Analogue to Digital Converter
ASIC – Application Specific Integrated Circuit
AVR – Automatic Voltage Regulator
DAC – Digital to Analogue Converter
DFIG – Doubly Fed Induction Generator
DSP – Digital Signal Processor
FPC – Full Power Converter
FR – Frequency Response
K.E – Kinetic Energy
IG – Induction Generator
IM – Induction Machine
MOSFET – Metal-oxide-semiconductor Field Effect Transistor
MPPT – Maximum Power Point Tracking
PM - Permanent magnet
ROCOF – Rate of Change of Frequency
RSC – Rotor Side Converter
SG – Synchronous Generator
SVM – Space Vector Modulation
Sync – Synchronous
uP - Microprocessor
VSC – Voltage Source Converter
WT – Wind Turbine
1
1 Introduction
1.1 Project Overview
This report concentrates on the theory, simulation and planned experimental implementation
of Frequency Response (FR) from Power Converter based WTs which are currently the
favoured choice of technology for wind power plant.
This PhD forms part of the work of the FLEXNET project which is investigating the design,
operation and optimisation of the future electricity system for the UK.
For experimental work a 1kW laboratory based wind turbine test rig is being constructed at
Cardiff University for practical implementation of wind turbine power electronic controls and
evaluation of other novel machine/plant control schemes.
1.2 Motivation
Increasing worldwide demand for Energy, worries about depletion of existing fossil fuel
reserves, nuclear proliferation, economic security and, by no means least Climate Change, are
some of the main reasons for the planned expansion of the wind power sector in the UK and
worldwide.
For the UK which has a privatised electricity system, whether market based methods are used
or central planning, as is the case in some nationalised industries, uncertainty still exists in
regard to the future make-up of the electrical energy supply.
Some features of a future UK energy sector are likely to be :
• Variability of output from the expansion of new Renewable sources such as wind
• Inflexibility of any future baseload Nuclear plant
• Declining indigenous Gas reserves
• Impetus to reduce CO
2
emissions particularly from coal
As with some other forms of Renewables such as Solar, Wave and Tidal these forms of
Energy ‘harvesting’ are stochastic in nature and cannot be dispatched with the same ease as
conventional generating plant such as Coal, Gas and Hydro. Also of note is that much of these
Renewables are of a distributed nature and thus are of a smaller unit size and are
geographically dispersed.
Thus in general there is an impetus in the industry for requirements for more flexibility from
plant in terms of participating in provision of energy output, reactive power and additional
ancillary services due to this new power system topology.
For the UK in particular, the planned large scale expansion of wind power is necessitating
research into how this relatively new type of plant can be successfully integrated into the
electrical power system. Displacement of conventional plant under a large penetration of wind
will require some ancillary services to be provided from WTs, of which FR is one such
service.
2
1.2.1 Frequency Response Services
Figure 1.1 demonstrates what happens when a Frequency ‘event’ occurs on the electricity
network. These rare events when there is a sudden unexpected large deficit between load and
generation which can result in an unwanted drop in system frequency (potentially causing
system collapse). Immediately after the step imbalance, stored inertial energy in spinning
machinery begins to be consumed by the generators helping to maintain their electrical
output. After a short delay in response, conventional synchronous plant contracted to have
spare reserve power begin to open steam governors and thus increase the generator power
output. This then restores the frequency on a temporary basis until ‘secondary’ slower
response plant can come ‘on-line’.
Figure 1.1 – Modes of operation of frequency control on the UK grid [
2
]
For maximisation of energy capture and thus increased energy revenues, WT plant in the UK
would participate in only the ‘primary’ response shown in the figure above. This short term
response would only affect the energy output of the machine for up to 30 seconds.
1.2.2 Frequency Response from Wind Turbines
Key technical aspects to investigate in the research area can be identified as below :
• Converter based turbines have no natural inertia response, thus access to this
stored K.E will need to be enabled
• The quantity of WTs on the system will vary at different times, thus management
and dispatch of this response will be likely different to that from conventional
Synchronous plant
• The positive power primary response currently given by Sync plant will again
need to be replicated in some form by WTs
3
50.2
49.8
49.5
49.2
10 s 30 s 60 s
F
r
e
q
u
e
n
c
y

(
H
z
)
time
continuous service
primary
response
secondary
response
to 30 min
occasional services
event
O
X
50.2
49.8
49.5
49.2
10 s 30 s 60 s
F
r
e
q
u
e
n
c
y

(
H
z
)
time
continuous service
primary
response
secondary
response
to 30 min
occasional services
event
O
X
• Capability of WTs to provide this service, dynamic effects on their machines and
loads on the turbine from providing this response.
1.3 Objectives
This project aims to contribute to the subject of investigation of provision of FR from power
converter based wind turbines by :
a) Modelling the GB system FR and obtaining/developing/refining a suitable
control method capable of an appropriate response on the Power System.
b) Undertaking experimental testing of such a control scheme in order to verify
its functioning and discover some of the practical implementation
issues/characteristics.
c) Explore further Turbine system issues and machine effects such as
aerodynamic effects, possibly using proprietary WT software simulation tools.
These objectives are further elaborated in the chapter ‘Further Work’.
4
2 Wind Turbine Generator Systems
The majority of WT capacity in existence today consists of machines in the range of 500kW
to 2MW, horizontal axis machines having two or three blades. These types of WTs are the
most common and are still the favoured choice for development with the trend being towards
even larger machines [
3
].
2.1 Energy Extraction
The theoretical maximum quantity of energy that can be extracted from a horizontal flow of
wind can be calculated as below :
The available energy in a wind stream is given by
1
2
3
P AU
air
ρ ·
where ρ is the air density, U is the wind speed, and A is the area swept by the wind turbine
blades.
However, the energy which can be extracted by the wind turbine is less than the available
energy in the wind and is given by
air
P
p
C
m
P ·
Where C
p
is called the power coefficient and depends on the tip speed ratio ρ which is the
ratio between the velocity of the rotor tip and wind speed defined by

U
R
r
ω
λ ·
where ω
r
is the aerodynamic rotor speed and R is the radius of the rotor.
C
p
has a theoretical maximum of 0.59 (known as the Betz limit) but will be typically up to 0.4
for a commercial 3-bladed turbine in operation.
For a given wind turbine design there is an optimum value of tip speed ratio λ
opt
, which gives
the maximum power extraction. Figure 2.1 shows a typical performance curve for a modern
high-speed wind turbine, the maximum efficiency of the turbine of approximately 0.45
occurring at a tip speed ratio of 7.
5
Figure 2.1 – Variation of coefficient of power with tip speed ratio [
4
]
A principle cause of the variation of the efficiency of power extraction of the blade assembly
is due to the variation of the angle of attack α of a fixed blade with the incident wind. These
components are shown in Figure 2.2 where U is the wind perpendicular to the turbine axis, W
is the apparent wind relative to the rotating blade and β is the blade pitch angle. At a low tip
speed ratios the blade is in a stall condition, at higher tip speed ratio the blade has a low angle
of attack and drag effects predominate, both of these effects thus causing less than optimum
power extraction [4] .
Figure 2.2 – Velocity components in the plane of a blade cross-section [4]
If the operational speed of a turbine rotor is allowed to change in order to keep the tip speed
ratio more or less constant an increased quantity of energy can be captured compared to a
fixed speed machine. shows the set of torque speed curves which define the performance
characteristics for a wind turbine machine for different wind speeds. As expected higher wind
speeds result in a higher combination of torque and rotor speed hence giving a higher power
output. This product of torque and rotor speed gives rise to the maximum power curve shown.
6
U

0
2000
4000
6000
8000
10000
12000
14000
0 500 1000 1500 2000 2500 3000 3500
Rotor speed [rpm]
G
e
n
e
r
a
t
o
r

t
o
r
q
u
e
[
N
m
]
6 m/s
7 m/s
8 m/s
9 m/s
10 m/s
11 m/s
12 m/s Curve for maximum
power
Speed limit
A
B
C
Aerodynamic torque
Figure 2.3 –Torque speed characteristics of a variable speed wind turbine [
5
]
2.2 Modes of Operation
Practical design constraints mean that the WT must be constrained from operating above a
maximum torque and maximum rotor speed. Both of these limits, and also the electrical
machine rating, coincide approximately with the rated wind speed of the WT which normally
is around 12m/s. Additionally there is a minimum wind speed below which operation of the
WT provides no benefit. Referring to Figure 2.4, these operating modes for a variable speed
type WT can be summarized as follows :
Below cut-in : Machine is not rotating and produces no power output.
Max Power Tracking : B to C - Electrical torque is controlled such that the machine speed
tracks the maximum power curve.
Constant Speed : A to B and C to D – Due to power converter constraints at low speed and
aerodynamic noise constraints at high speed, torque is allowed to vary but speed is held
almost constant [
6
]
Pitch Control or Stall : D to E - Speed and torque are both limited to their maximum values
by adjustment of the blade pitch angle. This stalling or feathering of the blade reduces the
torque produced. If no pitch mechanism is provided, stalling of the blades at high wind speeds
can be achieved through suitable blade design.
Shutdown : Above a certain speed the WT is brought to a halt to avoid damage.
7
Generator Speed
Rated Torque
D
C
E
B
A
Shutdown Speed
Speed Limit
Generator
Torque
Cut-in Speed
Figure 2.4 – Operating modes for a WT
For a fixed speed WT the modes would be the same with the exception of the maximum
power tracking mode which would instead be replaced with a constant speed operational
mode.
2.3 WT design features
2.3.1 Fixed speed versus variable speed
Fixed speed – This type of turbine is designed to operate within a relatively narrow shaft
speed range which is determined by the variation in slip of the induction generator connected
to it. This might be up to 1-2% at rated output [
7
]. With changing wind speed and an
essentially constant shaft speed, the tip speed ratio will change and thus the turbine will not
operate at its optimum power production point except for a single wind speed.
Variable speed – By allowing a more constant tip-speed ratio the turbine can be operated to
extract more power from the wind. There is a trade-off between complexity and this extra
performance. This design requires a either a full power converter design to allow the shaft
speed to vary independently of grid frequency, or alternatively a partial power converter
design (DFIG) which allows a greater range of slip than a conventional IG machine. Typical
slip ranges achievable with this configuration are -40 to +30% [
8
]
A slight variation of the variable speed design is a wound rotor induction machine with a
variable resistance rotor. A commercial implementation of this is known as Opti-slip
manufactured by Vestas. This design can allow variations in speed of typically 0-10% above
synchronous speed [8]. Notably this design avoids the necessity for slip rings and their
associated maintenance by incorporating all of the switching components on the rotor.
2.3.2 Power Control
Fixed blades with appropriately designed aerodynamic profiles or ‘pitching’ blades are used
to reduce the input mechanical shaft torque under excessively high wind speed conditions.
8
2.3.3 Electro-mechanical coupling
Direct grid connection of a synchronous machine driven by a wind turbine rotor is not
possible as the relatively rigid electrical coupling would give rise to high mechanical stresses
and also unwanted variations in electrical output during aerodynamic disturbances. These
might include tower shadow effect. Hence IG based machines which have inherent damping
due to their slip operation, or alternatively full converter interfaced machine connection
topologies are used which are fully decoupled via a dc-link.
2.3.4 Mechanical Coupling
Geared drives – In order to use industry standard 4-pole electrical machinery or similar a step-
up gearbox is necessary. A typical 500kW 40 meter diameter turbine has a rotor speed of 33
r.p.m [4]
Direct drive – By using a large diameter machine with multiple poles a ‘gearless’ drive
arrangement is possible. This has reliability advantages as gearboxes have proved to be a
common failure in early turbines. These down-time and maintenance issues are especially
relevant for offshore turbines.
2.3.5 Other design features
Output voltage ratings are typically 690V with a step-up transformer located close-by. Newer
designs are beginning to use higher generator output voltages.
2.4 Generator Types
The below topologies represent conventional designs typically implemented in horizontal 2 or
3 bladed versions common today from 250kW to 5MW.
2.4.1 Induction Generator
An induction Generator (IG) based WT is shown in Figure 2.5. This turbine operates in a
similar fashion to large scale induction motors which are commonly found in industry.
Similarly these WTs are equipped with power factor correction through capacitor banks to
compensate for reactive power consumption and a soft-starter to reduce in-rush currents on
start-up.
Benefits – Low cost, low maintenance squirrel cage rotor construction
Disadvantages – Reactive power consumption during faults, possibility of overvoltage in
islanding condition, less than optimum power extraction capability (varying tip-speed ratio),
essentially uncontrolled i.e no control of real or reactive power.
9
Squirrel-cage
induction generator
Soft-
starter
Capacitor
bank
Figure 2.5 - Induction Generator based turbine [
9
]
2.4.2 Doubly Fed Induction Generator
A Doubly Fed Induction Generator (DFIG) based WT is shown in Figure 2.6. By using a
wound rotor and a bi-directional part-scale converter this machine is able to operate over a
range of shaft speeds. In addition to real power control, terminal voltage can be controlled
through reactive power import and export capability. Control of this reactive power via the
rotor is effectively amplified and it is possible to use only a partial scale converter located
here to achieve the same reactive power control as a full scale one located on the stator [7].
Wound rotor
induction generator
Power Converter
Crowbar
Figure 2.6 – Doubly Fed Induction Generator [9]
10
2.4.3 Full Power Converter
A full power converter (FPC) based WT is shown in. By inserting a full bridge converter and
inverter between the electrical generator and the grid, complete rotational decoupling of the
turbine and grid system is accomplished. There is freedom to use an IG or SG machine. Also
if the generator requires no magnetizing current as in the case of a permanent magnet based
SG, the converter can be a simple diode bridge rectifier.
Power converter
Induction/Synchronous
generator
Figure 2.7 – Full Power Converter based WT [9]
Benefits – Complete control of output to grid side and hence ease of compatibility with grid
connection requirements. Dc-link provides damping for torque oscillations caused by varying
windspeed.
Disadvantages – Additional cost of power electronics

11
2.5 Power Electronic Converter Fundamentals
Both the DFIG and FPC type of WT employ a power electronic assembly to convert from ac
to dc and back to ac. For the DFIG this synthesised ac waveform is used principally to enable
control of the rotor magnetisation current and hence overall control of the machine and power
export through the directly connected stator. For the FPC all of the power generated must pass
through the power electronic assembly. To obtain the complete ac to ac conversion two
individual converters are connected ‘back to back’ via a dc-link.
2.5.1 Converter Classification
The general name ‘Converter’ is given to a circuit which can operate as both an inverter and a
rectifier although the term inverter is commonly used in its place [
10
]. Two basic types of
converter exist namely the current source converter and the voltage source converter.
Current Source Converter
This requires a stiff dc current source ideally with infinite Thevenin impedance at the input. It
can be constructed from a variable voltage source with feedback current control and a series
inductor. Asymmetric voltage blocking devices such as MOSFETs are not suitable for use in
a CSC, instead devices such as thyristors must be used [10].
Voltage Source Converter or Voltage Fed Converter (VSC)
This converter configuration uses a stiff dc voltage source at one side and converts it to ac at
the other, the Thevenin impedance of the source ideally being zero, a capacitor can be added
to assist in this [10]. This type of converter is commonly used for WT power electronics
applications.
2.5.2 Bridge Configurations
Depending on harmonic noise requirements, component cost considerations etc it is possible
to choose between a number of bridge configurations for the synthesis of 3-phase ac.
Two-level inverter
This is the simplest topology and is shown in Figure 2.8. It consists of three pairs of half
bridges a,b and c with a total of 6 semiconductor devices which are indicated in the figure as
simple switches.
Multi-stepped inverters
These use a greater number of switching devices to obtain a more accurate sinusoidal output,
an advantage being the reduced size of ac and dc filtering.
12
a
b
c
Upper
Lower
DC - Link
Vdc
Va
Figure 2.8 – 6-Pulse three phase inverter bridge
2.5.3 Square-wave Switching Scheme
In its simplest method of operation, fabrication of a three phase output to the 3-phase
balanced load shown in Figure 2.8 is done as follows. This method is called square-wave
switching :
General Rules
• Switches are either fully ON or OFF (this reduces switching losses)
• No two switches in the same half bridge can conduct at the same time. This would
be a short circuit of the DC supply with destructive currents and is known as
‘shoot-through’. For prevention of this scenario it is common practice for
switching logic to incorporate a ‘deadband’ period which enforces a delay period
between alternation of switching of the upper and lower bridge devices.
• When switching OFF a device, a path must be made for the conduction of the
inductive current in order to prevent overvoltage across the device and hence
destruction of the semiconductor switch. ‘Freewheeling’ diodes provide the path
for this current decay, ready for the reversal of direction of the current into the
opposite bridge leg.
Simple switching scheme
• Identical drive signals are sent to all of the bridges but are phase shifted by 2π/3
(this creates the 3 separate phase voltages).
13
• Since we are generating a.c signals current is expected to be always non zero so
the legs will switch alternately between upper ON and lower OFF and lower ON
and upper OFF.
Hence this scheme is implemented using the following sequence indicated in Table
2.1. This pattern alternately applies either +V
dc
or –V
dc
to a single phase in series with
the other two phases in parallel. Thus allowing conduction of current through the
upper device or the lower device and producing an alternating current. Voltages
developed across the phases are :
• legs in parallel will have 1/3 of Vdc
• single conducting leg 2/3
• upper conducting phases have positive voltage
• lower conducting phases negative voltages.
Table 2.1- Switching states and applied voltages for square-wave scheme
Half Bridge States
Ua Ub Uc
La Lb Lc
Va
(of Vdc)
Vb
(of Vdc)
Vc
(of Vdc)
Comment
0 0 0
1 1 1
OR
1 1 1
0 0 0
0
0
0
0
0
0
0
0
0
0
0
0
Three phases at same
potential. No freewheel
path.
Invalid states.
All ON 0 0 0 Destructive short circuit
current. Invalid state.
All OFF Bridge idle
1 0 1
0 1 0
1/3 - 2/3 1/3 Valid State
1 0 0
0 1 1
2/3 - 1/3 - 1/3 Valid State
1 1 0
0 0 1
1/3 1/3 - 2/3 Valid State
0 1 0
1 0 1
-1/3 2/3 - 1/3 Valid Stat
0 1 1
1 0 0
-2/3 1/3 1/3 Valid State
0 0 1
1 1 0
-1/3 - 1/3 2/3 Valid State
14
The resulting applied voltage waveforms from this scheme are shown below in Figure 2.9.
Note that the upper gate drive switching signals are the same wave form shape as the voltages
developed Va0, Vb0 and Vc0. The lower gate drive signals are the inverse of these.
Va0 is the voltage of the mid point bridge of leg a with respect to a fictitious midpoint on the
DC supply voltage. Va is the phase voltage of phase ‘a’.
2/3 Vdc
1/3 Vdc
PI
2 PI /3
Phase a
Phase b
Phase c
Va0
Vc0
Vb0
+0.5
Vdc
-0.5
Vdc
+0.5
Vdc
-0.5
Vdc
+0.5
Vdc
-0.5
Vdc
Va
Vab
Vdc
Vbc
Vdc
Vca
Vdc
Figure 2.9 – Switching waveforms and applied voltage for square-wave scheme
15
Observations that can be made on this switching sequence are :
• Either the upper or lower switch must be ON and the other OFF as conduction in
one direction or the other is always required to fabricate the alternating current.
• Since we have 3 phases the resultant overlap means either 2 upper devices and 1
lower or vice-versa 1 top device and 2 lower devices must always be active. This
gives the resultant applied voltage to the phases as shown in the table.
2.5.4 Pulse Width Modulation Switching Schemes
An improvement on the square-wave switching method is to use Pulse Width Modulation.
This technique modulates the length of the ON pulse and thus provides voltage control of the
synthesized output. In its most basic format a switching duty cycle is output depending on the
desired ON-OFF time ratio. The frequency of repetition of this duty cycle is determined by
the ‘carrier wave’ frequency.
More advanced PWM based schemes can produce a greater accuracy in sinusoidal output
with reduced harmonic content. Two popular methods are listed below :
SPWM – Sinusoidal Pulse Width Modulation is a common technique which varies the output
pulse width in proportion to the magnitude of an internally generated sine-wave reference
signal. This scheme will synthesize a good approximation of a sine-wave when an inductive
load is connected to an inverter bridge. The switching signals and output waveform of such a
SPWM scheme are demonstrated in Figure 5.69.
SVM – Space Vector Modulation utilises the concept of a rotating space vector. Although
computationally intensive one of its benefits is that it can ‘optimise the harmonic content’ of
an isolated 3-ph neutral load. This is of relevance to machine loads as these are often do not
have the neutral connected [10].
2.5.5 Back to Back Frequency Converter
Connection together of the dc sides of two 3-phase VSCs completes a circuit known as the
back-to-back frequency converter shown in Figure 2.10. This circuit enables frequency to
vary independently on both sides of the inverters and also power can flow in either direction.
In addition to its application in variable speed WT systems it is widely applied in industrial
drive applications and is known in its most flexible configuration as a 4-quadrant variable
speed ac drive.
16
Figure 2.10 – Back to back frequency converter
An observation of this circuit is that when equipped with MOSFET or IGBT devices, due to
the ‘body diode’ on each transistor the generator side (left hand side of Figure 2.10) switches
are able to freewheel and thus act as a diode bridge when power is required to flow from the
3-phase to the dc-link. The grid side switches of course require appropriately modulated gate
drive pulses in order to transfer current to the 3-phase grid side connection (right hand side of
Figure 2.10).
17
2.6 Electrical Machine Control
Two control methods which commonly appear in inverter bridge control schemes for use with
electrical machines are load angle control and vector control. For vector control use of the d-q
reference frame is necessary. These two control methods and the d-q reference frame are
discussed below.
2.6.1 Load Angle Control Theory
This method seeks to control the flow of power through an inductive element by manipulating
the angle between the voltage at the sending and receiving end sources. The principle is the
same as controlling the flow of power through a predominantly inductive power system
element such as a transmission line. In the case of a VSC, a coupling reactor provides the
inductive element across which the phase angle is measured. The real power flow is
determined by the angular difference between the voltages. The reactive power flow is
determined by the difference in magnitude of the voltages. For the circuit (a) and phasor (b)
shown in Figure 2.11
V
R
Sending Source
i
jX V
S
Receiving Source
V
S
jXI
V
R
I
δ
φ
B)
A)
A)
Figure 2.11 – Load angle power transfer (a) circuit and (b) phasor diagram
The relationship between power flows, angle and voltages can be derived as below :
Complex power S
S
= V
S
I
S
* = V
S

*


,
`


.
| −
jX
R S
V V
(2.1)
18
= V
S


,
`


.
|


jX
R S
* *
V V
= j
X
j
X
V
R S S
* 2
V V

(2.2)
Since V
S
= V
S
e

and
*
R
V = V
R
then S
S
= P
S
+ jQ
S
= j
,
`

.
| +

X
V jV V V
j
X
V
R S R S S
δ δ sin cos
2
(2.3)
Therefore
P
S
= δ sin
X
V V
R S
(2.4)
Q
S
= δ cos
2
X
V V
X
V
R S S
− (2.5)
where δ is the load angle, φ is the power factor angle, V
S
is the sending end voltage, V
R
is the
receiving end voltage and X is the inductive reactance between them.
The steady state active and reactive power flow equations (2.4) and (2.5) form the basis for
this control method whereby the voltage magnitude seen at the receiving end V
R
and the load
angle δ are controlled to provide the required real and reactive power flows.
For a VSC placed at the receiving end, the PWM reference sine wave and duty cycle can be
adjusted to give the required load angle and magnitude independently of the generator at the
sending end.
2.6.2 Vector Control
In order to implement a vector based control scheme it is necessary to convert the 3-phase
rotating quantities of a machine into a stationary reference frame in two dimensions. A brief
explanation of this is given below for an IG although the basic principles are equally applied
to a SG machine also.
The per-phase equivalent machine circuit shown in Figure 2.12 for an IG is only valid for
steady-state conditions. For high performance control of a machine a controller based on
equations derived from this model is not sufficient.
19
Figure 2.12 – Equivalent circuit model of induction generator [
11
]
An IM can be considered to be a ‘transformer with a moving secondary, where the coupling
coefficients between the stator and rotor phases change continuously with the change of rotor
position [10]’. It is possible to accurately model the machine to include the varying
inductances but the differential equations involved would make it too complex to implement
as a control system. Hence some form of simplification is required.
A 3-phase machine can be represented as an equivalent 2-phase machine with d
s
, q
s
being
direct and quadrature stator components and d
r
, q
r
being direct and quadrature rotor
components. These however still have time varying components.
Parks Transformation - R.H.Park in the 1920’s solved this problem by replacing the stator
voltages, currents and flux linkages instead with variables rotating at a synchronous speed in a
fictitious rotor winding. This is known as ‘Parks Transformation’. This transformation
eliminates time varying inductances. Later H.C Stanley and then Krause and Thomas ‘showed
that time varying inductances could be eliminated by referring the stator and rotor variables
to a common reference frame which may rotate at any speed (arbitrary reference frame)’ [10].
d-q reference frame – This final transformation leads to all the machine variables being
described in the d-q reference frame as shown in Figure 2.13 . The derivation of these
equations is too complicated for this report however the reader is referred to [
12
] and [
13
].
Figure 2.13 – dq representation of an induction machine [9]
20
Vector Control (Field Oriented Control) - The basic idea behind vector control (which is also
known as field oriented control) is to force an IM to behave as a separately excited dc brushed
machine. By doing this electromagnetic torque can be made nearly instantaneously equal to a
demanded torque. Permutations of this control include stator flux-oriented, rotor flux-oriented
and air gap flux-oriented methods, of which all of these can be indirect or direct methods.
Direct and indirect methods refer to the necessity or not, for use of sensors to measure air gap
flux linkages [12].
Rotor Flux Oriented Control – This is one example of the possible vector control methods
and its application to an IM. In this control scheme the rotor flux linkage vector is kept
perpendicular to the rotor current vector. The vector diagram for this scheme is shown in .
Figure 2.14 – dq representation of stator and rotor currents with rotor flux orientation [7]
This alignment is accomplished by setting the q-axis rotor flux ψ
qr
to zero and secondly by
setting the d-axis rotor current i
dr
to zero. In terms of practical implementation, a result of this
control method is that :
• ‘torque control’ of the machine is determined by the q-axis stator current i
qs

• ‘flux control’ (for control of voltage) is determined by i
ds
.
Some coupling does exist between these d and q control terms which can be compensated for
in a practical control scheme.
21
2.7 Wind Turbine Generator Control Schemes
Examples of vector control schemes and load angle control schemes for WTs are shown in
this section. In all of these schemes the dc-link voltage is assumed to be regulated at a
constant voltage by managing the flow of real power from the generator to the grid by an
appropriate means. If this real power is not transferred there is a danger of the WT over-
speeding or the dc-link voltage rising to an excessively high value. This is a particular
problem during grid fault conditions when low grid-side terminal voltage can prevent export
of real power. Control of back-to-back converters can be split into grid-side and generator-
side control functions.
2.7.1 Full Power Converters
2.7.1.1 Generator Side Control
For an IG or a Synchronous machine the generator side converter control can be realised
either as load angle control or vector control. Control schemes presented here export power to
the dc-link to ensure the turbine tracks its maximum power curve. Generators which have no
reactive power transfer capability such as PM synchronous machine can use a simple
uncontrolled diode bridge rectifier for the generator side converter.
Load Angle Control – Figure 2.15 shows the implementation of this control strategy [
14
]. The
desired power P
ref
is determined by the turbine maximum power point look-up table. The
internal generator voltage e
G
is calculated from the rotor base speed ω
b
and the rated voltage
E
G
. Manipulation of these values gives inputs to the PWM control of angle control α
G
and
terminal voltage control v
G
.
Vector Control for Synchronous Generator – A generator side vector control strategy is
shown for a synchronous based WT in Figure 2.16. This uses a d-axis defined along the flux
linkage vector ψ
m
. In practical terms this means that if the flux linkage vector ψ
m
is known,
the torque of the machine can be controlled by the q component of the stator current i
qs
.
Correspondingly control of the d component of the stator current i
ds
exerts control of the
reactive power production of the generator. The two PI controllers scale the current demands
into suitable d and q components of voltage, for the VSC PWM control signals.
22

ω
r
P

ref G
G
G G
p x
e v
α ·
ref
p
PWM
control
2 sin
2 cos
ds G G
qs G G
v v
v v
α
α
·
·

ω
r

r
G G
b
e E
ω
ω
·
2
G ref G
G
G
e q x
v
e

·
G
α
G
v
Figure 2.15 – Generator side load angle control strategy for FPC
Flux control in d-axis
Torque control in q-axis
SYN
Generator
ds
v
qs
v
Swing equation
in e m
d
T T J
dt
ω − ·
e
T
in
T
VSC
GE
p
r
ω
e ref
T

qs ref
i

qs
i
+

+
+
1
.
e ref
T
k

m
ψ
qs
v
G
e
G d ds
b
e X i
ω
ω

ds
i
ds ref
i

ds
i
+

I
P
K
K
s
+
+
+ ds
v
qs
i
q qs
b
X i
ω
ω
m
ψ
qs
i
G
e
ds
i
m
ψ
m
ω
2 P
r
ω
I
P
K
K
s
+
Figure 2.16 – Generator side vector control strategy for a synchronous machine [14]
Vector Control for PM Synchronous Generator - For a permanent magnet machine the direct
stator current reference i
ds
in Figure 2.16 can be set to zero as no reactive power is transferred.
Vector Control for Induction Generator – The generator side vector control strategy for an
induction generator based WT can be implemented by selecting the d-axis to align with the
rotor flux. Then by regulating the d axis stator current i
ds
air gap flux is controlled and by
regulating the q axis stator current i
qs
torque is controlled. Such a control scheme is shown in
Figure 2.17. Since this is an asynchronous machine the VSC carrier frequency must be set
according to the desired slip s and actual rotor speed ω
r
as calculated.
23
Figure 2.17 – Generator side vector control for induction machine [9]
24
2.7.1.2 Grid Side Control
The grid side control scheme is required to transfer the power incoming from the generator
side VSC, out to the grid side. This ensures a constant dc-link voltage. Secondly the grid side
VSC must provide reactive power transfer as appropriate to its agreed grid connection. An
FPC connected to the grid is shown in Figure 2.18. Real power to transfer is P
gr
and reactive
power to transfer is q
GR
. An inductive reactor X
gr
is connected between the converter and the
WT terminal outputs, the terminal output voltage being v
1
Figure 2.18 –FPC connected to grid [9]
Load Angle Control – A control scheme to maintain correct real power export (and hence
regulate dc-link voltage) and also terminal voltage control is shown in Figure 2.19.

DCref
v

dc
v
GRref
p


+
I
P
K
K
s
+
sref
v

GRref
q

+

I
P
K
K
s
+
s
v
θ
GR
v
( )
, ,
GR ref s
f p v θ


( )
, ,
GR ref s
g q v θ


Figure 2.19 – FPC grid side load angle control [14]
25
Dc-link error and terminal voltage error are passed through PI controllers to produce
demanded values for real and reactive power transfer. These commanded values pass through
appropriate function to give the load angle setpoint θ and the converter terminal voltage v
GR
.
As the grid connection is an infinite bus v
s
cannot be changed.
Vector Control – This method seeks to control the grid side PWM converter by manipulating
the VSC output V
GR
and the terminal voltage V
S
in the dq reference frame. The scheme is
shown in Figure 2.20. Control of real power and reactive power are implemented by q axis
and d axis currents respectively.
DC ref
v

dc
v
q ref
i
− −
+
I
P
K
K
s
+
s ref
v

d ref
i

+

I
P
K
K
s
+
s
v
R-F
Trns
R-F
Trns

+
+

I
P
K
K
s
+
I
P
K
K
s
+

+
+
+
s
X
s
X
+
+
q
i
d
i
qs
v
ds
v
qGR
v
dGR
v
Figure 2.20 – FPC grid side vector control [
15
]
Vector Control with AVR controlled dc-link - A slight variation on the basic vector control
method presented is when a wound rotor synchronous machine uses its AVR to maintain the
dc-link voltage. In this case the dc-link error summing point with PI control in the upper loop
of Figure 2.20 is replaced by a maximum power point reference and power summing point
instead [
16
].
26
2.7.2 Doubly Fed Induction Machines
A property of DFIG machines is that transfer of reactive power to the network is possible via
the rotor side converter or the stator side converter. The advantage of reactive power transfer
via the rotor is that it is magnified by a factor of 1/s. The converter is also capable of
transferring real power to the rotor from the network when in sub-synchronous operation and
from the rotor to the network in super-synchronous operation [7]. The schematic for a typical
DFIG and its overall control scheme is shown in Figure 2.21.
Figure 2.21 - Schematic of DFIG WT [11]
The crowbar protection shown operates by shorting the rotor through a resistance to protect
the rotor from over-current in a fault condition.
2.7.2.1 DFIG Control Scheme
An example of a commonly implemented scheme is detailed as below.
Generator Side Control – The generator side controller provides control of stator power p
s
though torque control and additionally controls stator reactive power q
s
. Similar to vector
control implemented for FPC machines, by splitting the controlled current into the two
orthogonal components in the dq reference frame, individual control of torque and terminal
voltage is facilitated. In the case of the DFIG, the rotor currents i
dr
and i
qr
are controlled as
opposed to the stator currents i
ds
and i
qs
of a synchronous or singly fed induction machine.
This type of controller is also known as PVdq or current-mode control [7].
Implementation of the torque control loop is shown in Figure 2.22. Torque demand T
c
from
the maximum power point curve (not shown) is summed with an optional term which
provides synthesis of the inertial action found in synchronous machines. The resulting torque
demand is converted to a rotor quadrature reference value whereby it is then summed with the
actual current value. This is then output through a PI control, summed with a decoupling term
and finally passed to rotor side VSC (not shown) as a q-axis voltage setpoint. Note that in this
figure, ς is the Laplace s constant.
27
Figure 2.22 – DFIG rotor side torque control loop [11]
Implementation of the voltage control loop is shown in Figure 2.23. The upper part of the
input to the first summing block represents the generator magnetising component of the i
dr
current. The lower part represents the component controlling reactive power flow with the
network [Cartwright]. After passing through a delay term, then summation with measured i
dr
a
PI control converts the demanded current into a d-axis voltage term. After adding a
decoupling term the demanded V
dr
is passed onto the VSC PWM input (not shown).
Figure 2.23 – DFIG rotor side voltage control loop [11]
Grid Side Control - The grid-side controller maintains the dc-link voltage by import or export
of real power.
2.7.2.2 Other DFIG Control Schemes
Rotor Flux Magnitude and Angle Control (FMAC) – This scheme adjusts the magnitude and
angle of the rotor flux vector. It has an advantage of low interaction between the voltage and
power loops and has good system damping and voltage recovery after faults [7]
28
3 Connection Requirements & Response Capability
3.1 UK Requirements for grid connection
In order for plant operators to connect their generators to the electricity network whether at
distribution or transmission level the plant must comply with what are known as ‘Grid
Codes’. These specifications which are issued by the respective power system operator are
necessary so that plant behaves appropriately under all known operating conditions and also
abnormal conditions. For wind farms it is noted that the requirements apply at what is known
as the ‘Point of Connection’ where the aggregated power enters the network rather than the
individual WT terminals themselves. The following sub-sections concerning the UK grid-
codes are taken directly from [
17
]
3.1.1 General
The connection of new generation in Great Britain is governed by the Grid Codes of National
Grid plc [17]. A collection of non-synchronous generating units that are powered by an
intermittent source, joined together by a system with a single electrical point of connection
(may include a DC Converter) to the GB transmission system is categorised as a “Power Park
Module”. The Grid Code only applies to a Power Park Module such as a wind farm, not
individually to power park units (i.e. individual wind turbines). Almost all the performance
requirements that are mandatory for the power park module are applicable to modules
installed in England and Wales with a completion date on or after 1 January 2006. However,
the performance requirements applicable to power park modules in Scotland vary and the
Codes are under continual review. Therefore it is recommended to refer to the most up-to-date
Grid Codes.
For a Generating Unit or Power Park Module using an intermittent power source, the
requirement is that the active power output shall be independent of system frequency for
system frequency changes within the range 50.5 to 49.5 Hz and should not drop with system
frequency by greater than the amount specified in Figure 3.24 for system frequency changes
within the range 49.5 to 47 Hz.

49.5 47.0 50.5
95% of Active
power output
100% of Active
power output
Frequency
Figure 3.24 - Requirement placed on the output power of a generating plant in terms of
frequency [17]
29
At the point of connection the active power output under the steady state conditions of any
Generating Unit, DC Converter or Power Park Module directly connected to the GB
Transmission System should not be affected by voltage changes in the normal operating
range, that is t5% continuously or t10% for 15 minutes for 400 kV, t10% continuously
for 275 or 132 kV and t6% continuously for less than 132 kV.
3.1.2 Steady State Reactive Power and Voltage Control
Conventional synchronous plant is required to control the voltage and also absorb or generate
reactive power, in accordance with the needs of the power system. Normally the transmission
system operator determines the operating settings of these generators.
All Power Park Modules (excluding those connected to the total system by a current source dc
converter and those connected at 33kV or below) must be capable of supplying rated MW
output at any point between the limits 0.95 power factor lagging and 0.95 power factor
leading. With all plant in service, the reactive power limits defined at lagging and leading
power factor as a function of the active power output are defined in Figure 3.25. These
reactive power limits will be reduced pro rata to the amount of plant in service.
The Power Park Modules are also required to provide continuously acting automatic voltage
control system to provide control of the voltage and operation of the plant without instability
over the entire operating range of the plant. The automatic control system shall be designed to
ensure smooth transition between the shaded area bounded by CD and the non-shaded area
bounded by AB in Figure 3.25.
Point A is equivalent (in MVAr) to: 0.95 leading Power Factor at Rated MW output
Point B is equivalent (in MVAr) to: 0.95 lagging Power Factor at Rated MW output
Point C is equivalent (in MVAr) to: -5% of Rated MW output
Point D is equivalent (in MVAr) to: +5% of Rated MW output
Point E is equivalent (in MVAr) to: -12% of Rated MW output
Figure 3.25 - NGC plc reactive power requirement [17]
30
3.1.3 Fault Ride Through Capability
Each power park module and any constituent power park unit shall remain transiently stable
and connected to the system without tripping for a close-up solid three-phase short circuit
fault or any unbalanced short circuit fault on the GB transmission system operating at
voltages of 200 kV or above for a total fault clearance time of up to 140 ms. In this case,
(a) during the period of the fault each power park module shall generate maximum
reactive current without exceeding the transient rating limit of the generating unit
or power park module and/or any constituent power park unit.
(b) each power park module shall be designed such that upon both clearance of the
fault on the GB transmission system and within 0.5 seconds of the restoration of
the voltage at the grid entry point active power output shall be restored to at least
90% of the level available immediately before the fault.
shows the typical fault recovery for cases with two circuit breakers and three circuit breakers
(see top right hand corner of each diagram for the configuration).
Figure 3.26 - Typical fault recovery for two-ended and three-ended circuits [17]
For voltage dips of greater than 140 ms, each power park module and any constituent power
park unit shall remain transiently stable and connected to the system without tripping for
balanced voltage dips and associated durations any where on or above the solid line of Figure
3.27. In this case,
31
(a) provide active power output, during voltage dips at least in proportion to the
retained balanced voltage at the grid entry point except in the case of a
asynchronous generating unit or power park module where there has been a
reduction in the intermittent power source in the time range in Figure 3.27 that
restricts the active power output below this level and shall generate maximum
reactive current without exceeding the transient rating limits of t the generating
unit or power park module and/or any constituent power park unit.
(b) restore active power output within 1 second of restoration of the voltage to at least
90% of the level available immediately before the fault.
Figure 3.27 - Minimum voltage dips above which generators should be stable and connected
[17]
3.1.4 Power System Stabiliser and Black Start Capability
The requirements for excitation control facilities, including PSS can be agreed when signing
the bilateral agreement.
The GB Grid Code states that black start capability is agreed at a number of strategically
located power stations.
3.1.5 Frequency Response
According to the Grid Codes each power park module must be capable of operating in a
manner to provide frequency response at least to the solid boundaries of Figure 3.28. Each
power park module must be capable of providing some response, in keeping with its specific
operational characteristics, when operating between 95% to 100% of registered capacity as
illustrated by the dotted lines in Figure 3.28. If the frequency response capability falls within
32
Voltage as a %
of nominal
Voltage duration
the solid boundaries, the power park module is providing response below the minimum
requirement which is not acceptable.
The capability profile specifies the minimum required level of primary, secondary and high
frequency response. The definitions of these responses are based on the curves shown in
Figure 3.29. The phrase “Minimum Generation (MG)” applies to the entire power park
module operating with all generating units synchronised to the system. The Designed
Minimum Operating Level (DMOL) is the output at which a power park module has no high
frequency response capability. It must be less than or equal to 55% of the registered capacity.
Figure 3.28 – Minimum frequency response profile for a t0.5 Hz frequency change
[17]

33

Figure 3.29 - Interpretation of Primary, Secondary and High Frequency response [17]
3.1.6 Reserve
The power park modules are not obliged to provide reserve. However, the power park
modules can participate for providing fast reserve or short term reserve under an ancillary
services agreement or under a bilateral agreement. Definition of the two terms fast reserve
and short term reserve are as follows :
(a) Fast Reserve - provides the rapid and reliable delivery of active power through an
increased output from generation or a reduction in consumption from demand
sources, following receipt of an electronic dispatch instruction from National Grid.
(b) Short Term Operating Reserve - Short Term Operating Reserve (STOR) is a
service for the provision of additional active power from generation and/or
demand reduction. The STOR service could be a committed service or a flexible
service.
34
3.2 Desired Response
In order for continued operation of a future power system with a varying mix of wind power
and conventional synchronous plant, an assumption is made that any future wind capacity will
need to provide a primary response of a format similar to that required for existing
synchronous based generating plant. An important characteristic of primary response is that it
should be “released increasingly with time, through automatic governor action, in the period
10-30 seconds after the incident and sustained for a further 20 seconds.” [2].
3.2.1 NGC Benchmarking of Plant FR capability
For a generating plant participating in FR services a typical response characteristic to a
change in frequency is shown in Figure 3.30
Figure 3.30 – Typical plant frequency response characteristics [2]
The upper graph shows a negative system frequency deviation of 0.8Hz after 10 seconds and
0.5Hz after 60 seconds. The sample at 10 seconds being representative of a point in the
primary response period (0 to 30 seconds) and the 60 second point being representative of a
point in the secondary response period (30 seconds to 30 minutes).
The process of benchmarking a generating plant is outlined in [2]. The middle graph shows a
suitable frequency test waveform which might be input to a generator system. The lower
graph shows the positive power response which would be expected from a generating plant.
3.2.2 Performance of Synchronous Plant
By taking the response of plant to differing frequency deviations a response profile can be
constructed as shown in Figure 3.31. The Minimum Stable Generation (MSG) and Generator
Registered Capacity (GRC) represent the limits of operation of the generator. As can be seen
part-loaded plant has a proportionally greater capability for an increase in plant output when
system frequency drops. For an increase in system frequency, highly loaded plant has a
greater capability for reduction in plant output. Larger deviations in frequency show an
increased quantity of response as expected.
35
Figure 3.31- Typical Genset Frequency response profile [2]
These response profiles are of a similar format to minimum frequency response profiles as
specified in grid code requirements e.g Figure 3.28.
It is noted in [2] that due to the nature of ‘air breathing engines’ such as gas turbines, a drop
in frequency (i.e synchronous shaft speed) may effect such a plant’s capability to output a
positive power response.
3.2.3 System Requirements for FR
The quantity of frequency responsive generation on the system will need to vary depending
on the system demand. As demand drops a greater quantity of FR plant will need to be on-
line. For the primary response phase for the UK system this is shown in Figure 3.32.
If the ‘largest credible generation loss risk’ occurs, which is 1320MW, a similar quantity of
response to cover this loss is still needed to control the frequency deviation. Additionally
since there are less loads on the system any characteristic contribution from load reduction
due to frequency drop will be less. Other factors which contribute to this include low
frequency load tripping and speed of response of contributing generators. A similar need for
increasing response for lower system demand also occurs for the secondary response service
[2].
Figure 3.32 – System frequency response requirements [2]
36
3.3 Primary Response Capability from WTs
3.3.1 Primary Frequency Response Schemes
Individual WT modelling
Figure 3.33 shows an overview of an example WT model used in frequency response
modelling. Included is a wind model, a two-mass model of the rotating components, and a
high level turbine controller principally for limitation of power capture at above rated speeds
consisting of a blade pitch angle control block.
Figure 3.33 – Wind turbine model for frequency modelling [
18
]
Auxiliary Signal for Inertial Response
One simple method used by [
19
,
20
] for FR is shown in . By introducing a proportional and 1
st
order delay term acting on the system frequency signal, a modified Torque/Power signal is
produced. This signal is combined with the existing torque setpoint of the machine which
finally results in a modified current setpoint.
Figure 3.34 – Block diagram of auxiliary torque/power signal for DFIG inertial response [20]
37
The differential gain term provides a torque in proportion to the rate of change of frequency
similar to the torque occurring in a synchronous machine due to inertial energy release. The
delay term provides ‘shaping’ of the response.
The performance of a 1.5 MW DFIG turbine with this FR loop, connected onto an 8MVA
stand-alone diesel system is shown in the graphs in Figure 3.35 for a disturbance of -0.15p.u
[20]. The solid line indicates the response with the WT FR loop included, the dashed line is
with no response from the WT. With the FR loop included the lower left graph shows a
constant slip angle during the frequency disturbance thus indicating that the turbine rotor
speed is now ‘coupled’ back to the system frequency. In the right-hand middle graph, the
improvement in performance of the system is evident from the reduction in ROCOF and also
the reduction in frequency excursion magnitude.
Figure 3.35- DFIG (at 100% power) response to a frequency disturbance with (solid) and
without (dashed) supplementary inertia effect [20]
38
Identical Inertial Restoration
A further refinement of the ‘delayed df/dt’ method is possible, in order to obtain an identical
SG type response, by dynamic modification of the gain value used for K. If slip is forced to
stay constant, because this is a variable speed turbine the K.E extracted for a particular change
in frequency will vary depending on the operating speed of the turbine. By suitably scaling
gain value K to depend on speed, the response loop can be shown to give a constant FR
contribution regardless of variations in operating speed [20].
Maximum Energy Extraction Algorithm
A new ‘stepwise’ algorithm is presented in [20] which combines a stepwise increase in
electromagnetic torque followed by an adjustable ramp-down period. The profile for this
torque response is shown in Figure 3.36(b). Its operation can be described as follows :
• Onset of frequency event – After detection of a drop in system frequency at time
t
0
, an increase in torque is commanded ∆T. This results in an increased power
output for the machine hence a reduction in system ROCOF. The increase in
electrical torque above the aerodynamic torque results in turbine deceleration at a
rate proportional to the difference in torque and the machine inertia value.
• Gentle rampdown and avoidance of stall condition – Stall is the operation of a
blade when a low lift to drag ratio occurs [4]. To avoid operation of the turbine in
such a low efficiency condition during the primary response phase a lower limit of
operational efficiency is set. In this algorithm it has been chosen to be at point ω
crit
shown in Figure 3.36(a). Time t
1
and t
2
are chosen such that time t
2,
the end of the
wind turbine contribution to the frequency response phase, will coincide with the
chosen ω
crit
. A method is shown in the paper to perform this calculation.
The size of ∆T, on set of the ramping down, end of the response time t
2
and quantity of K.E
extracted could be adjusted to suit the response required. In the paper ∆T was set at 20% and
the ramp down length (t
2
– t
1
) to 10 seconds.
Figure 3.36- (a) Variation of aerodynamic torque with generator speed for wind speed = 11.5
m/s (rated power). (b) Stepwise torque method for FR. [20]
Results of this algorithm are shown in Figure 3.37. A sustained positive power response of
approximately 13 seconds is produced. At 15.6 seconds the electro-mechanical torque and the
aerodynamic torque become equal and the deceleration of the machine ends.
39
Figure 3.37- Performance of stepwise method for maximum K.E extraction [20]
Comparison of Schemes
The performance of three methods of inertial response from DFIG WTs is undertaken in [20]
and is presented in Figure 3.38.
• Power Control - dotted line. ‘Power control’ which limits the power increase to
10% and the net exchange of active power to zero. This is thought to be a scheme
which would be attractive from a power system operator’s perspective, although it
was noted that this scheme might cause turbine stall.
• Inertial response – Dash dot. This is the simple ‘delayed df/dt’ scheme using a
proportional gain and delay on a frequency input to produce a modified torque
demand.
• ‘Step ramp’ algorithm – solid line.
• No response from WT – dashed line.
As can be seen in the right hand graphs the step-ramp algorithm as configured shows the
greatest speed decrease out of all of the methods and provides a significant reduction in
frequency excursion.
40
Figure 3.38- Comparison of primary FR methods. Dashed, no support. Solid, Step-wise 20%
torque increase. Dash dot, Inertial response (K=39, T = 0.1). Dotted, Power control (10% step
and net exchange is zero. [20]
3.3.2 Machine Effects & Converter Current Limits
Regulation Parameter Choice
The effects of using different machine parameters for inertial regulation purposes are
highlighted in [20]. Differences between using P
stator
, T
el
, P
total
and the effect on rotor power
and possible stalling of the turbine are discussed. Additional effects of introducing MPT to
the overall control and its anti-stalling effect is shown.
Converter Over-currents
Reference [20] draws attention to the over-current limitation of the DFIG’s RSC and hence
the FR capability of the machine at different operating points. This is demonstrated in Figure
3.39. The upper left graph shows the limited capability to provide a positive power response
when the turbine is operating at 80% of rated power. A 0.25 MVA RSC is used on the
1.5MVA turbine in the simulation.
41
Figure 3.39 - Influence of DFIG initial loading to the delivery of inertia effect. Dash-dot,
rated power. Dashed, 80% loading. Solid, 40% loading. [20]
42
3.4 Secondary Response Capability from WTs
Although the primary objective of this work is provision of primary response, it is of interest
to consider wider wind farm control and provision of reserve as when implemented they may
have closely related control and communication aspects.
3.4.1 Wind Farm Control Scheme
A wind farm controller which includes FR functionality is presented in [18]. This is shown in
Figure 3.40. Individual WTs communicate their available power with the central controller
and receive dispatched power and voltage setpoints according to the Voltage Monitor and
Frequency Controller.
Figure 3.40 – Wind Farm Control Strategy [18]
‘Five principle tools’ to deal with FR/reserve are identified in [18] as :
• Absolute Limit – Capping of maximum power output.
• Ramp Limit – Restriction in rate of rise or fall of power.
• Balance Control – Similar to conventional generator control in providing a known
ramp up or ramp down capability.
• Delta Control – Tracking of maximum power output capability but at a constant
percentage below e.g actual production could be set at a constant 5% below
available production to provide spinning reserve.
• Maximum Export Limit (MEL) – This is an upper limit on power export defined
for UK generators. This limit can be changed by the transmission operator to
ensure appropriate levels of FR are in place.
43
3.4.2 Frequency Controller
A WT controller is presented in [18] which has capability of implementing Delta, MEL or
Balance control respectively. Depending on grid frequency and power available from the
turbine an appropriate reference power for the turbine converter is produced.
The results for a multi-megawatt FPC equipped induction machine WT for a negative and a
positive frequency disturbance using MEL control are shown in Figure 3.41. Wind speed
input, turbine speed, pitch angle, available power output and actual power output are shown
for the three control methods possible. Wind speed of 12 m/s is used along with a turbulence
intensity of 18%.
Figure 3.41 – MEL Control [18]
[B-H] concludes that rapid power response for participation in FR is possible from a large
FPC induction machine based WT and aggregation of outputs from a dispersed resource of
WTs would be expected to smooth the outputs seen in the results.
Note that the model used does not feedback into a system frequency model.
44
Available
Power
Output
Power
4 Modelling
4.1 Introduction
Method - The modelling here is undertaken by transferring all of the mathematical
relationships which describe the system into a set of s-domain transfer functions. This
common procedure enables the solution of differential equations by standard algebraic
methods [
21
]. Matlab-Simulink is used to solve these equations and also provides a useful
graphical interface to describe the model and then view the results in the time-domain.
Assumptions – For simplification and speed of simulation all of the loads are lumped together
as one single mass. This amalgamated representation is also the case with the response from
the sync plant and the wind turbine plant. Note that future simulation work is planned with a
‘partitioned’ response.
Simulation Objectives – The primary objectives of the simulations presented here in this
chapter are to :
• demonstrate relationship between frequency and power on a simplified UK system
• show how this frequency is regulated by sync plant and additionally WT plant
• verify that the model produces the same results as given in [19]
• present a baseline model to be used for investigation of more advanced WT
response
Note - As in common with most power systems modelling all units are expressed in p.u
quantities unless otherwise stated
4.1.1 Equations of Motion
The imbalance between electrical torque T
e
and mechanical torque T
m
and the relationship
between these and rotational inertia are the fundamentals of load-frequency studies. These are
derived below for a single machine [13]
During an imbalance between power and load the net accelerating torque T
a
is :
T
a
= T
m
- T
e
(4.1) where T
m
, T
e
are positive for a generator (N.m)
The combined inertia is accelerated accordingly and gives rise to the ‘Swing equation’:
J
dt
d
m
ω
= T
a
(4.2) J is moment of inertia (kg.m
2
), ω
m
angular velocity (rad/s)
For power system studies inertia is normally given in terms of the per unit inertia constant H
which is defined as :
H =
base
VA
E K.
=
base
m
VA
J
2
0
5 . 0 ω
(4.3) where ω
0m
is rated angular velocity (rad/s)
45
Substituting for J in the swing equation :
2H
m base
e m
m
m
VA
T T
dt
d
0 0
/ ω ω
ω −
·


,
`


.
|
(4.4)
Because T
base
= VA
base
/ ω
m
the p.u equation of motion can be expressed as :
2H
dt
d
r
ϖ
=
e m T T −
(4.5) where we define
r ω
as
m
m
0
ω
ω

Dispensing with the ‘m’ subscript we can define ω
r
and ω
0
as the angular velocity and rated
angular velocity respectively in rad/s.
Including a component of damping torque proportional to speed deviation we get :
2H
dt
d r ω
=
e m T T −
- K
D

r ω
(4.6)
Rearranging to obtain the acceleration :
dt
d r ω
=
H 2
1
(
e m T T −
- K
D

r ω
) (4.7)
As ∆
r ω =
r ω -
0 ω then
dt
d r ω
= dt
d r ) ( ω ∆
(4.8)
By substituting for
r ω
we can now express our swing equation in final form as :
dt
d r ) ( ω ∆
=
H 2
1
(
e m T T −
- K
D

r ω
) (4.9)
For load-frequency studies the preferable quantities to analyse are Power and frequency as
opposed to Torque and frequency [13].
For a small deviation (delta) from initial values (subscript 0) and all values in p.u :
P = ω
r
T
P = P
0
+∆P
T = T
0
+ ∆T
ω
r
= ω
0
+ ∆ω
r
46
P
0
+ ∆P = (ω
0
+ ∆ω
r
) (T
0
+ ∆T) (4.10)
Assuming that the product of ∆ω
r
and

∆T is comparatively small then :
∆P = ω
0
∆T + T
0
∆ω
r
(4.11)
Also since ω
0
= 1 and if ∆ω
r
is small then ∆P = ∆T
Hence for small speed deviations, generator power P
gen
and load power P
load
we can express
the swing equation as :
dt
d r ) ( ω ∆
=
H 2
1
(
load gen P P −
- K
D

r ω
) (4.12)
Note that because of p.u quantities ∆
r ω is of course identical to ∆ f on a synchronous
system.
4.2 System Model
Figure 4.42 shows equation (4.12) implemented as ∆P model in the s-domain along with the
plant response which forms a closed loop feedback system. Transfer of the equation on to the
s-domain is shown in the appendices. From an initial steady state operating point,
disturbances are applied to the system and the effects observed with differing plant frequency
response controls.
Sensitivity of loads with respect to a change in frequency is given by the system damping D.
Stored inertia and therefore initial rate of change of frequency for a power disturbance is
dictated by H
eq
, the equivalent combined inertia of the system (not including any additional
synthesised wind turbine inertia). Synchronous governor response is determined by the
combined transfer functions of the Droop, Governor and Turbine transfer functions which are
shown in the next section. Wind turbine response is given by inserting the model of Figure
4.44 in to the system model. Both the wind turbine response and synchronous response
represent aggregated models of the individual plant on the system.
∆f setpoint
= 0
1
D + 2H
eq
s
∆P
Sync Gov. Response
Wind Turb. Response
+
+
Disturbance
∆f
-
+
Figure 4.42 - GB delta power inertial model
47
4.2.1 Synchronous Plant Response
This response consists of the expected ∆P increase from all of the synchronous plant on the
system. The quantity of this plant and its scheduled level of response (i.e it’s droop) is
determined through the setting the gain in the ‘Governor Droop’ block. The below figure is
taken from the Simulink model and is of the form of a generating unit with a reheat steam
turbine [13]. The simulation model uses a time constant of 12 seconds for the Turbine Re-heat
[19].
- 1 1
G o v e r n o r
D r o o p G o v e r n o r T u r b i n e
T u r b i n e R e - h e a t
O u t 1
1
1
0 . 3 s + 1
2 s + 1
1 2 s + 1
1
0 . 2 s + 1
- K -
I n 1
1
w r _ d P d _ s y n c
Figure 4.43 – Steam Turbine Transfer function
4.2.2 Reduced Order Machine Model
Manageable simulation of the power system for frequency studies requires use of a reduced
order aggregated wind turbine model. For the wind turbine plant the linearised ∆T model
used is shown below in Figure 4.44[19]. This model is accurate for small changes in torque
about an initial operating point. Parameters used for this model are calculated using the data
for a 2MW IG machine given in appendices Table 7.3. Note this machine model can also be
used to represent a DFIG with appropriate parameter selection.
-
+
Kp + Ki/s
X2
[1 + sT
1
]
X
3
X
1
1 . ∫
J




-
+
0.6
p.u
T
shaft ∆T
x
ω
r
∆P
(2MW)
ω
r
T
dem
∆f ∆T
f_ctrl
+
+
T
setp




-
+
Max power
lookup curve
0.6
p.u
Initial
Aero Torque
Initial
Torque
G(s)
Frequency Response
Control Block
η
Figure 4.44 - Linearised ∆T wind turbine model
48
4.3 Control Scheme
By increasing the Torque set point of the wind turbine in response to a frequency deviation a
short term increase in electrical power output of the machine can be achieved. Supplementary
control to enable this inertial/fast primary response action is shown in Figure 4.45. The
combination of the transfer functions below provides an increase in commanded power output
in response to the change of frequency.
The synthesis of the inertia is provided by the gain k
1
acting on the rate of change of
frequency deviation signal. Appropriate adjustment of this gain can give replacement of the
inertia. Depending on the gain chosen the turbine can be made to done one of the following
[19] :
• make the rotor speed track the system frequency
• provide a set quantity of inertia independent of rotor operating speed (this is the
ideal ‘like for like’ replacement of Synchronous plant inertia
• customised synthesis of inertial response as chosen by the system operator
Additional shaping of the response is provided by the ‘washout’ filter in combination with
gain k
2
and also the first order delay block using T
f
in combination with gain k
f.
These additional first order delay terms extend the overall response towards the end of the
primary response phase.

1 + s T
s T
w
w

dt
d

1 + s T
k
f
f

k
1
k
2
T
f_ctrl ∆ f
Figure 4.45 - Wind turbine frequency response control block
4.4 Simulink Model
The model used to produce the results in this report is shown in Figure 4.46 and is based on
the Matlab mdl file used in [19]. The complete model showing subsystem components is
shown in the appendices. The only difference between the models is that the ‘No of m/c’ gain
block uses a different value and to compensate for this, a gain block needed to be added
before the ‘freq sig lag’ switch. More detailed comments about the differences between these
models are available in the appendices.
49
G B _ f a s t _ P R _ 1 . m d l
I a n M o o r e - 1 4 A p r i l 2 0 0 9
G B S y s w i t h F a s t P R f r o m W i n d
G B _ f a s t _ P R . d o c
R e v i s i o n 1
1 6 t h O c t o b e r 2 0 0 9
I a n M o o r e
s y n c _ g o v _ r e s p o n s e
I n 1 O u t 1
W T _ r e s p
I n 1 O u t 1
f 1 . m a t
D O C
T e x t
1
5 0
2 H t e r m
1
6 . 2 1 s + 1
w _ d
w _ d
P d _ d i s t u r b
f w _ r
P d _ s y n c
P d _ s y n c
Figure 4.46 - Simulink model (top level only shown) for investigation of primary response
from WTs
4.5 Setup
The 2020 high wind penetration scenario given in appendices Table 7.4 was selected for
simulation. This scenario assuming that all plant on the system is at full output uses the
following parameters :
Total Capacity 63.54 GW
Synchronous Capacity 40.78 GW
Wind Turbine Capacity 19.4 GW
Total equivalent inertia (H) 3.11
The wind capacity was operating at an initial output of 0.6pu before the load disturbance.
4.6 Results
Simulations were conducted for various control parameters, load disturbances and
participating frequency response combinations.
4.6.1 Open Loop Wind Turbine Response
For a system disturbance of +0.05pu which equates to an increase in load (or loss of
generation) of 317.5 MW, with no response from wind turbine capacity and only (the
inherent) inertial contribution from spinning synchronous plant, the system frequency is
shown in Figure 4.47. The system comes to equilibrium due to the decrease in power taken
from frequency dependant loads. The initial rate of change of frequency is determined by the
inertia of the spinning synchronous plant.
50
For the same disturbance the open loop response of the wind turbine capacity is shown in
Figure 4.48. A peak positive power increase of approximately 10% is shown for the first 15
seconds followed by a net reduction in power output for the period 15 to 30 seconds after the
disturbance; in the second phase, the turbine operating at a reduced power output, due to
being driven off its optimum operating point, on the maximum power point curve. Variation
of the supplementary controller parameter k
2
shows the effect on gain of the response during
its positive and negative output.
7 0 7 5 8 0 8 5 9 0 9 5 1 0 0 1 0 5 1 1 0
4 9 . 7 5
4 9 . 8
4 9 . 8 5
4 9 . 9
4 9 . 9 5
5 0
T i m e [ s ]
F
r
e
q
u
e
n
c
y

[
H
z
]
Figure 4.47 – System frequency for a generation disconnection of 0.05 p.u (no
response from generation)
7 0 7 5 8 0 8 5 9 0 9 5 1 0 0 1 0 5 1 1 0
0 . 5 4
0 . 5 6
0 . 5 8
0 . 6
0 . 6 2
0 . 6 4
0 . 6 6
0 . 6 8
0 . 7
0 . 7 2
0 . 7 4
T i m e [ s ]
W
i
n
d

T
u
r
b
i
n
e

T
o
r
q
u
e

[
p
.
u
]


K 2 = 0 . 5 , K 1 = 3 , T w = 1
K 2 = 1 . 0 , K 1 = 3 , T w = 1
Figure 4.48 – Variation in Wind Turbine Power in response to frequency deviation shown in
figure for varying k
2
51
4.6.2 Closed Loop Wind Turbine Response
For a larger disturbance of +0.0189 pu (1320 MW), Figure 4.49 shows the no-response and
the wind turbine only responses effect on system frequency. This shows a reduction of
frequency excursion measured at the beginning of the secondary response phase from
approximately 49.05 Hz to 49.35 Hz. Additionally the initial rate of frequency change is
reduced by a factor of five from – 0.1 Hz/sec to -0.02 Hz/sec.
7 0 7 5 8 0 8 5 9 0 9 5 1 0 0 1 0 5 1 1 0
4 9
4 9 . 2
4 9 . 4
4 9 . 6
4 9 . 8
5 0
5 0 . 2
T i m e [ s ]
F
r
e
q
u
e
n
c
y

[
H
z
]


W i t h o u t W T f r e q u e n c y r e s p o n s e
W i t h W T f r e q u e n c y r e s p o n s e
Figure 4.49 – System frequency after 1320 MW generation disconnection, no plant
response and Wind Turbine only response
4.6.3 Wind Turbine and Synchronous Response
For the combined response from synchronous plant and wind turbine plant, Figure 4.50 shows
the results obtained for the same disturbance as in the previous case. The addition of the wind
turbine response clearly shows a reduction in the initial frequency excursion from
approximately -0.25 Hz to -0.125Hz.
7 0 8 0 9 0 1 0 0 1 1 0 1 2 0 1 3 0 1 4 0 1 5 0
4 9 . 7 5
4 9 . 8
4 9 . 8 5
4 9 . 9
4 9 . 9 5
5 0
T i m e [ s ]
F
r
e
q
u
e
n
c
y

[
H
z
]


W i t h o u t W T f r e q u e n c y r e s p o n s e
W i t h W T f r e q u e n c y r e s p o n s e
Figure 4.50 – System frequency after 1320MW generation disconnection, synchronous
governor response and wind turbine plus synchronous governor only response
52
4.7 Simulation setup and results summary table
Figure of
[19]
Scenario Disturb
pu
Freq
dev
Hz
(pu)
K1 K2 Tw Freq
Signal
Gain
Freq
Signal
Lag
Block
Sync
Droop
setting
Wind
Cap.
(no of
m/c)
Results
(Te_WT), f
Comments
On match to paper
Unless
specified
otherwise
-3 1 1 k1, X1, X3 are negative
3a 0.25Hz
disturb
No response
0.005pu 0.25
(-0.005)
na na 0 0 Time constant 5sec not
11?
3b FPC
Open Loop
0.005pu 0.25
(-0.005)
50 OFF 0 0 0.6 to 0.72 Good
3c DFIG OFF 0 0 Not simulated here
4a 0.25Hz
disturb same
3a
O.L
Various Various OFF 0 0 Did previously simulate
4b 0.5Hz
disturb
O.L
0.01pu -0.5
(-0.01)
50
100
OFF 0 0 0.6 to 0.85
0.6 to 1.15
Less than Paper 0.6 to 1.2
0.59 to 1.18
6.
2020
High Wind
0.9Hz
disturb
Wind + No
Sync
0.0189pu -0.9
(-0.019)
500 ON
(1/20s+1)
-11 19.4/63.54 -0.65Hz Freq at
30sec
(-0.013pu)
Good
Paper is -0.55Hz at 30 sec
(0.6to0.7pu Te at Wind
Turbine)
8.
2020
High
Wind
Sync Only
Sync +
Wind
0.0189pu -0.9
(-0.019)
500 ON
(1/20s+1)
-11 0
19.4/63.54
2.5sec to initial spike
of 0.25Hz (-0.005pu)
Settles -0.08Hz
(-0.0016pu)
Similar
Similar
Page 53 of 112
Table Key : O.L – Open Loop, Assumed : Output Power (pu) in paper = Te_WT (pu) in this model
Page 54 of 112
4.8 Discussion of results
The simulations undertaken clearly demonstrate the capability of DFIG and FPC based wind
turbines in provision of frequency control in the primary response phase. The desired
response from conventional synchronous generation is for a gradual release in the period 0
to10 seconds, followed by a sustained response in the subsequent 10 to 30 second period [2].
It is logical to assume that future power system requirements will require this form of
response from wind turbine plant.
Figure 4.49 shows a good contribution to control of frequency when there is no synchronous
response on the system. Figure 4.50 also shows good response when a combination of
synchronous and wind turbine response operates. However for this latter case the overshoot in
frequency correction (at 93 seconds) is unwanted and optimisation of controller parameters
and design would be warranted.
In addition to the key research objective of investigating an optimum response from the UK
‘fleet’ of WTs further evaluation of the following points is suggested :
• Overload capability of existing marketplace WT Converter units
• Effect of response on WT rotor speed and onset of stall conditions
These would enable increased refinement of parameters for optimum response and further
ascertain the practicality of implementation of the FR scheme.
Page 55 of 112
5 Experimental Wind Turbine Test Rig
5.1 Overview
A wind turbine test rig consisting of a 1kW motor generator set with a 3-phase full
inverter/converter bridge is being developed in the Institute of Energy here at Cardiff
University. This is to be controlled via a dSPACE® ‘rapid control prototyping’ embedded
system.
Important capabilities of the Test Rig are :
• accepts various driven machines i.e Synchronous, IG or DFIG generator
• auto-generates downloadable controller code from Matlab-Simulink models
• easily configurable inverter/converter bridges
• allows practical evaluation of different turbine control-schemes and machine
topologies
Photographs of the test rig are shown below in Figure 5.51 to Figure 5.53.
Figure 5.51 - Early development of rig showing dc pendulum machine left and 3-ph sync
generator to right
Page 56 of 112
Figure 5.52 – Complete wind turbine test rig
Figure 5.53 – From left to right : Transformer and line module, full bridge back to back
converters and resistive load bank
Page 57 of 112
5.2 Design
The functional blocks shown in Figure 5.54 are described in this section. Extra specifications
and details of custom made circuitry relating to these hardware components can be found in
the appendices.
Figure 5.54 – Wind Turbine Test Rig functional blocks
5.2.1 DC Motor/ Pendulum Motor – Block A
Since the purpose of this Rig is primarily to investigate wind turbine electrical machines and
their associated control aspects, instead of a real wind tunnel and blade assembly a DC motor
is used to supply the mechanical power to the wind generator electrical machine. This is a 4-
quadrant machine capable of acting as a motor or a generator in both rotational directions.
The controller for this machine, which is provided by the manufacturer, is capable of closed
loop control of speed and torque both via an internal setpoint or through an external input
signal. It also has a mode for direct control of armature current. Its functioning is indicated in
Figure 5.55 below and also by graphics on the controller front panel. A transfer function for
such an ‘armature’ controlled machine can be found in appendices Figure 7.72.
Since this is a shunt DC machine with constant field excitation the current of the machine and
hence also its torque is controlled directly by the turn-on duty cycle of the armature thyristors
(contained within the Power Electronics block). The operation of a dc shunt machine means
that the terminal voltage applied to the machine is countered by a back-emf generated by the
machine itself which is proportional to the armature speed.
The proportional elements of the control loop shown provide appropriate gain to increase or
decrease machine applied voltage (and hence current and therefore torque) reducing the error
in the loop and finally by the integrator action reducing the error effectively to zero.
Page 58 of 112
DC Sep Exc
Shunt Machine
Power
Electronics
T
n
Armature
Current
n setpoint
Thyristor
ON - Time +
-
I set
I set
PI
PI
+
-
Constant
Voltage
Field
PI
T setpoint +
-
Figure 5.55 – DC pendulum controller function
Proposed Control of Aerodynamic Torque
For simple replication of the aerodynamic operation of the turbine an appropriate control loop
is proposed which will determine the torque setpoint for the driving DC pendulum machine
according to a basic wind speed versus torque output look up table. This look-up table
corresponding to the basic turbine aerodynamic performance for different wind speeds. This
will be implemented via the main embedded controller. The control action is indicated in
Figure 5.56. Note that only a single Torque Speed curve characteristic is shown which
corresponds to a single wind speed.
[
22
]
Page 59 of 112
Pendulum
motor
Generator
Excitation
Speed Torque
Characteristic
Speed/V
ADC
DAC
Torque
set point
Controller
ADC
Torque/V
PI
Load
Figure 5.56 – Proposed implementation of aerodynamic torque characteristic
5.2.2 Generator Machine – Block B
Via a flexible shaft coupling various machines can be directly connected to the DC motor.
For the synchronous wound rotor machine (4-pole non-salient) an open loop ‘excitation
voltage controller’ provides a current in proportion to the mark-to-space ratio switching
control input. Specifications for these machines can be found in the appendices.
Interconnection of the blocks A and B along with relevant signal scaling and unit part number
identifying the components are shown in Figure 5.57.
Page 60 of 112
Figure 5.57 – Pendulum motor (Block A) to Generator machine (Block C) connection
Page 61 of 112
5.2.3 Back to Back PWM converters – Block C & D
Two identical ‘6-pulse’ bridges are arranged in a back to back configuration, via a dc-link, to
form a complete a.c to a.c convertor. This circuit configuration is very flexible and enables
independent control of the grid and generator. The bridge arrangement along with placement
of necessary voltage and current sensors is shown in Figure 5.60
MOSFETS
BUZ 384 N-channel devices are employed which are nominally rated to 10.5A. These were
part of a pre-manufactured assembly which include extra components for suppression of
noise. These extra circuit components are shown in the appendices but essentially consist of a
series resistor and capacitor connected across the drain and source. Such an ‘R-C turn-off
snubber’ prevents voltage spikes and oscillations across the MOSFET during device turn-off
[
23
].
Gate Drives
To convert the logic switching output signals from the DSpace controller to an appropriate
level suitable for operating the MOSFETs an integrated 3-phase bridge driver chip,
International Rectifier IR2133 was used. This IC provides essential features of :
• Level shifting via external bootstrap capacitors to provide high turn on voltage
required for the floating upper bridge gate drive outputs
• Electrical isolation to protect the embedded controller from high voltages in power
side component failure scenarios
• Additional protection logic including ‘deadband’
Operation of Bootstrap circuitry
In order to drive an IGBT or MOSFET with the lowest ‘on-state’ voltage drop across the
device (and hence lowest power loss) the gate voltage must be 10 to 15V above the source
voltage [
24
]. This means the gate drive supply would need to be in excess of the DC rail
voltage. The option used to provide this supply for their IR2133 integrated circuit is a
‘bootstrap’ arrangement which is indicated in Figure 5.58.
Page 62 of 112
Figure 5.58 – Bootstrap supply schematic [
25
]
When the lower leg power transistor is conducting the bootstrap capacitor charges up to Vcc
via the bootstrap diode. This lower device then terminates conduction and the upper device
begins conduction. At this stage Vs rises to the dc-link voltage Vdc whereby the bootstrap
capacitor voltage V
B
is now lifted to Vdc + Vcc. This process repeats, as every time the
transistor is turned on, the bootstrap capacitor discharges to provide the power supply for the
amplifier driving the upper gate drive.
Figure 5.59 shows the final drive stages for the gate outputs which consist of an upper and
lower pair of transistors arranged to drive each gate in a push-pull configuration.
Figure 5.59 – Final output stage in a typical monolithic gate drive [24]
Page 63 of 112
These driver circuits are connected to the embedded controller via opto-isolating transistors
for protection.
Switching Scheme
The dSPACE controller comes equipped with pre-configured PWM outputs as below :
• 3-phase PWM or SVM (1 set of 3-phases)
• 1-phase PWM (4 individual channels)
Thus for the test rig implementation one converter will need to be driven using the 3-phase
PWM outputs, the other will need fabricating from 3 single PWM channel outputs. The fourth
PWM signal is to be used for the DC link regulation duties. In non- SVM mode all of the
PWM signals can be modulated by an appropriate sine wave as demonstrated later in this
chapter.
Note that the generator side MOSFETS can provide rectification without being switched by
virtue of their reverse bias rectifying property (body diodes) and thus will allow power flow
from the ac generator side to the dc-link and hence produce a voltage on the bus.
Page 64 of 112
Page 65 of 112
M
Ph 2
Ph 3
Ph 3
Ph 2
Ph 1
Ph 1
IR 2133
Gate Drive
Ph 2
Ph 3
Ph 3
Ph 2
Ph 1
Ph 1
IR 2133
Gate Drive
dSPACE DS1103 PPC
LEM
LV 25-P
LEM
LTS 25-NP
AVR
Red
Yell
Blue
VLL_gen
LEM
LV 25-P
LEM
LTS 25-NP
VLL_grid
LEM
LV 25-P
Current
Sensor
Figure 5.60 – Wind Turbine Test Rig back to back converter bridge configuration
5.2.4 Generator and Grid Side Controller – Block E & F
These are the two main control duties to implement. These will regulate the appropriate
voltages, currents and power flows on the generator, DC Link and grid side to implement the
required wind turbine control scheme and correct synthesis of the grid side a.c power output.
A typical DFIG control scheme is shown in Figure 5.61 below. The scheme uses vector
control of the generator via the rotor current to give reactive power control using i
d
and real
power control using i
q
. The grid side converter maintains real power flow to or from the grid
in order to regulate the dc link voltage.
Appropriate load angle control and variations on basic vector control are planned to be
implemented for both FPC and DFIG machines.
In addition to speed, position information from a tachometer on the machine shaft is available
to the embedded controller. This will be used in the d-q transformation within the generator
and grid side PWM control schemes.
Figure 5.61 – Block diagram for the control of a DFIG [
26
]
Page 66 of 112
Embedded Control System
Block E & F are both implemented on a dedicated control system. The embedded controller,
software, instrumentation and development interface is capable of automatically generating
downloadable code from Matlab-Simulink blocks. Additionally modification of parameters
and viewing of variables and data is possible in real time including construction of custom
display instrumentation. Notable features of the embedded controller platform are :
• fast main CPU (1 GHz)
• slave motor control DSP
• range of peripheral communication interfaces
Figure 5.62 shows a system diagram for dSPACE hardware platform.
Figure 5.62 – DS1103PPC embedded controller architecture [
27
]
Page 67 of 112
Sensors/Instrumentation
Industry standard closed loop voltage and current transducers are used :
• Voltage : LEM LV-25P, galvanically isolated, closed loop compensated transducer
using hall effect
• Current : LEM LTSR 15-NP, galvanically isolated, closed loop hall effect, ASIC
based
Circuit configuration for these sensors can be found in the appendices
5.2.5 Power System – Block G
This consists of basic configurable power system load elements including a 3-phase
transformer, line module with series resistive inductive and shunt capacitive elements and a
resistive load bank. This will enable testing of the wind turbine machine for various real and
reactive power loadings including step changes in output. In the later stage of the project it is
planned to connect the test rig to another motor/generator load system in the laboratory to see
how the wind turbine machine responds to step changes in system frequency. More
information on this power trainer can be found from relevant equipment manual [
28
].
5.2.6 General Assembly and Connection
Cabling
Twisted pair cabling is used for the gate drive to MOSFET gate connection in keeping with
recommended practices for minimisation of radiated noise and to reduce the possibility of
spurious switching signals [
29
,
30
] .
Earthing
To ensure noise free operation, isolation for protection of equipment, and personal safety
from electric shock, please note the overall earthing arrangements highlighted below and in
Table 5.2
• dSPACE CP1103 connector panel has all BNC bodies tied to mains earth.
• Voltage and current sensors are provided with insulated BNC connectors
• Gate drive BNC’s are insulated
The driven generator, both converters and the dc-link are floating with respect to mains earth.
Table 5.2 – Earthing scheme for wind turbine test rig
Item Earthing Voltage Comment
AC driven machine Isolated from mains 0-400 Vac
Gen – Side Inverter Upper gate drive BNCs at Vdc-link
DC-Link Referenced to AC
driven machine
0-600 Vdc dc voltage above 50V
Grid Side Inverter Upper gate drive BNCs at Vdc-link
AC – Grid Side Isolated from mains
Voltage Sensors Floating BNCs 0-500V in /
+/- 5V out
Connection of these using metal BNC
may compromise operator safety
Current Sensors Floating BNCs +/- 5V out Connection of these using metal BNC
may compromise operator safety
DSpace Break Out box BNCs Earthed +/- 30V dc Logic 0V tied to mains earth internally
Page 68 of 112
Note connection of any neutral points on the driven machine, DC-Link or bridge generated a.c
may impact/necessitate careful reconsideration of the above and review of any safety
measures taken.
Protection
A slow blow fuse rated to 1.6A is fitted to the synchronous rotor excitation winding in order
to protect the rotor against over-current.
A slow blow fuse rated to 1.6A is fitted to the dc-link to prevent excess current in the dc-link
5.2.7 Setup
The basic procedure to create, implement and run a control scheme for the test rig is detailed
in the appendices along with comments on specific use of the relevant software packages.
5.3 Results
The initial testing of the rig was undertaken in open-loop configuration using a low level of
excitation for the a.c generator and hence a low value of d.c link (approx 30 V) and a
correspondingly low voltage fabricated a.c wave output.
5.3.1 Measurement and Open Loop Control Test
Open Loop Control - This test demonstrates the open loop control capability of the control
system by outputting a control voltage determined by the centre left ‘slider control’ shown in
Figure 5.64. This signal is output from the 6
th
DAC channel (shown in Figure 5.65) on the
controller and is connected to the external speed input signal of the pendulum machine
controller. The simulink model for this is shown in Figure 5.65 and the subsystem block in
Figure 5.63.
Instrumentation Capability –Basic measurements of currents, voltages, shaft speed and torque
are input to the relevant ADC channels and then displayed on the PC instrument layout.
Additionally appropriate blocks were used to calculate shaft power and also the real and
reactive power components of the measured currents and voltage to give the outputs P_2 and
Q_2 shown.
Closed Loop Control – Relevant simulink blocks to implement closed control of voltage via
the synchronous machine AVR control input are shown at the bottom of Figure 5.65. This
functionality was not tested however as no over-voltage protection scheme was present on the
AVR input.
- K -
v o l t p e r r p m e x t
1
r p m _ s e t _ g a i n r p m _ l i m i t e r
0 . 1
d s 5
D A C
D S 1 1 0 3 D A C _ C 6
1
C o n s t a n t
r p m _ s e t _ r a w r p m _ s e t
Figure 5.63 – Open loop speed setpoint control

Page 69 of 112
Figure 5.64 – Instrumentation layout for basic measurement experiment
Page 70 of 112
+ / - 1 i s + / - 1 0 V e q u i v
C o n v e n t i o n :
P o s i t i v e T o r q u e
f o r m o t o r - g e n - l o a d
d u t y c y c l e :
0 - 1 0 V c t r l v o l t a g e
g i v e s 0 - 1 0 0 %
1 0 0 V e r r o r
g e t 1 0 0 % d u t y
1
v o l t _ s e t _ g a i n
- K -
v o l t p e r r p m
- K -
v o l t p e r d u t y
- K -
v o l t p e r V d i f r
- K -
v o l t p e r N m
0 . 1
v o l t p e r A m p
f ( u )
s q r t _ p s q r _ q _ s q r
s p e e d _ s e t _ D A C 6
- K -
r p m p e r v o l t
- K -
r a d _ s p e r r p m
0 . 1
d s 6
0 . 1
d s 4
0 . 1
d s 3
0 . 1
d s 2
0 . 1
d s 1
1 0
a s 4
1 0
a s 3
1 0
a s 2
1 0
a s 1
- K -
V d i f r p e r v o l t
T e r m i n a t o r 3
T e r m i n a t o r 2
T e r m i n a t o r 1
T e r m i n a t o r
R a t e L i m i t e r
R T I D a t a
si gnal rm s
RMS1
si gnal rm s
RMS
P r o d u c t 1
P r o d u c t
P I D
P I D C o n t r o l l e r
- K -
N m p e r v o l t
M a x _ d u t y
D i v i d e
M U X A D C
D S 1 1 0 3 M U X _ A D C _ C O N 1
D A C
D S 1 1 0 3 D A C _ C 8
D A C
D S 1 1 0 3 D A C _ C 7
D A C
D S 1 1 0 3 D A C _ C 5
D A C
D S 1 1 0 3 D A C _ C 4
D A C
D S 1 1 0 3 D A C _ C 3
D A C
D S 1 1 0 3 D A C _ C 2
D A C
D S 1 1 0 3 D A C _ C 1
A D C
D S 1 1 0 3 A D C _ C 2 0
A D C
D S 1 1 0 3 A D C _ C 1 9
1
C o n s t a n t 1
1 0
A m p s p e r v o l t
V
I
PQ
Active & Reactive
Power
3
3 p h a s e s
r p m
N m
N m
A m p s
V d i f r
r a d _ s
p o w e r _ s h a f t
A m p s _ r m s
V d i f f _ r m s
p o w e r _ l o a d _ p e r _ p h a s e p o w e r _ l o a d
Q _ 2 Q _ 2
P _ 2
S _ 2
P F
d u t y _ s e t
v o l t _ s e t _ r a w
v _ e r r o r
Figure 5.65 – Matlab-Simulink model for basic testing of instrumentation capabilities
Page 71 of 112
5.3.2 Bridge Inverter Test
Square Wave switching - This test demonstrates the simple square wave switching scheme as
detailed in chapter X. By setting the ‘sin_true’ constant to zero as shown in Figure 5.66 a
50:50 duty cycle square pulse is sent to each PWM output channel. Channel b and c are
delayed by the appropriate 2Pi/3 phase angle. Note that by sending a pulse with a low value of
0 and a high value of 1 means that no modulation is present on the output signal from block
‘DS1103SL_DSP_PWM3’.
For experimentation some variation of the frequency and pulse widths can be achieved by
varying the input parameters shown on the instrument layout of Figure 5.67. Additionally
there is a ‘Plotter’ component in the top right hand corner to verify that the modulating signals
chan_a, chan_b and chan_c are correct.
f _ r a d s = 5 0 * 2 * p i
n e e d t o c r e a t e i n W o r k s p a c e
p e r i o d = 2 * p i / f _ r a d s
d o u b l e ( ( i n t 1 6 ( ( 2 * p i / f _ r a d s ) * 1 0 0 0 ) ) ) / 1 0 0 0
c o n v e r t s t o 3 d p p r e c i s i o n
N O T E S :
t e s t 1 . m d l
E x e r c i s e s b a s i c 3 - p h P W M s y n t h e s i s
I a n M o o r e O c t 2 0 0 9
1
s i n _ t r u e
1
_ O N _ O F F
T r a n s p o r t
D e l a y 1
T r a n s p o r t
D e l a y
S w i t c h
S i n e W a v e 3
S c o p e 3
S c o p e 2
S c o p e
R T I D a t a
P u l s e
G e n e r a t o r
D u t y c y c l e a
D u t y c y c l e b
D u t y c y c l e c
P W M S t o p
D S 1 1 0 3 S L _ D S P _ P W M 3
c h a n _ b
c h a n _ a
c h a n _ a
p u l s e _ a
c h a n _ c
Figure 5.66 – Simulink model for bridge inverter testing
Page 72 of 112
Figure 5.67 – Instrumentation layout for bridge inverter testing
Page 73 of 112
Figure 5.68 shows the phase a gate drive signal and the voltage developed across phase a
(resistive) load for the square wave scheme driving a load with resistive and inductive
elements. For this RL load, as expected there is a delay in the change in current and hence
delay in voltage rise as the respective magnetic fields change in the inductive components.
Figure 5.68 – Fabrication of a.c waveform in RL load using square wave switching
(CH1 50.0V/div, CH2 10.0V/div, time base 10.0ms/div)
Sinusoidally Modulated PWM -
Sinusoidal modulated PWM signals are created by inputting a sinusoidal signal which varies
between zero and one into the DS1103SL_DSP_PWM3 block. This block outputs a centrally
aligned PWM signal with a duty cycle in proportion to the block input. For example a zero
input gives a 0% duty cycle output, an input of 0.5 gives a 50% duty cycle output, an input of
one gives a 100% duty cycle output.
The instrument layout shown in Figure 5.67 includes the facility for frequency adjustment of
the generated sinusoidal wave which is fed into the DS1103SL_DSP_PWM3 block.
Figure 5.69 shows the results for this switching scheme with a series resistive inductive load.
This waveform shows a considerable improvement over the previous switching schemes and
bears close resemblance to the ideal sinusoidal shape required of a grid side inverter. Some
noise is present which is to be expected on a basic Sinusoidal modulated PWM scheme.
Figure 5.69 – Fabrication of a.c waveform in RL load using Sinusoidally modulated PWM
(CH1 50.0V/div, CH2 2.00V/div, time base 10.0ms/div)
Page 74 of 112
Gate drive logic
signal
Voltage across load
Gate drive logic
signal
Voltage across load
6 Further Work
The contribution made by the PhD is planned in three main areas of work and are planned as
below
1 - Modelling of appropriate Response from Wind Turbines :
This work involves using Matlab-Simulink to further investigate and optimise a suitable
control algorithm for provision of FR from WTs. The key thrust of this work is to use multi-
machine representation of WTs and varying their response either individually or collectively
to provide the optimum response.
2 - Experimental testing :
Testing of the FR control scheme will be undertaken in order to verify its functioning and
discover some of the practical implementation issues/characteristics.
This will firstly involve stand-alone operation of the WT rig and its response to a change in
load.
A second more realistic stage of testing is planned where the WT will be connected onto a
larger Power system whereby a change of frequency will be instigated on the larger system by
again a step change in load or step adjustment of the main power system generation output.
Observation of the effects of the implemented FR on the test rig DC driving machine may
feed into the 3
rd
work topic detailed below.
3 - Exploration of further Turbine system issues :
Whilst providing a positive power response to the grid, a major concern with any WT system,
is the effect that this transient change in power will have on the mechanical and aerodynamic
assembly of the turbine. Excessive stresses may lead to premature failure or maintenance and
downtime issues for the WT. This work topic will explore these areas from the point of view
of the effect on FR on the WT assembly.
Note this work may involve the use of proprietary WT software simulation tools such as those
from Garrad Hussan in Bristol.
Page 75 of 112
6.1 Risks
Some potential problems identifiable in the project are –
WT Rig
• General noise from switching causing malfunctioning/spurious operation of the
WT rig
• Harmonic noise from WT rig interfering with operation of Power System Trainer
• Learning curve associated with developing a complex embedded controller
Simulation
• Balance between accuracy and simulation run time of models
• Access to accurate data for Turbine machine and Power system modelling
Turbine System issues
• Learning curve for aerodynamic theory
• Obtaining use of 3
rd
party software
6.2 Gantt Chart
The sequence of tasks and approximate time allocations are indicated in the Figure 6.70.
Development of the WT test rig is described in more detail than the simulation work as this
practical element of the project involves less uncertainty in terms of task sequence.
Page 76 of 112
Figure 6.70– Gantt Chart for PhD
Page 77 of 112
7 Appendices
7.1 Simulations
7.1.1 Simulation Baseline Record
Model Name : GB Sys with Fast PR from Wind File : GB_fast_PR.mdl
Date : 13 April 2009 Author : Ian Moore
References
Paper : J.Ekanayake. N.Jenkins, “Frequency Response from Wind Turbines”, Wind
Engineering Vol 32 No 6 2008
Change History
Previous Version : Fig7.mdl, Janaka, March 2009
Janakas Mods (compared to Paper) : Added a first order lag block for smoothing, which in
fact provided the extended power response for figs 6 & 8
Ians Mods : Addition of Freq Signal Gain (needed because of pu WT confusion at system
summing point) / Addition of switch to remove lag block so to enable sharp response of fig 3
and 4 results
Ians Minor mods : Signal names added for scope identity /Converted pu for output to freq for
format like in paper
General Settings
Machine Type IG, parameters as appendix X
kp,ki = 10,1 although paper states 0.5,0.5 Note : doesn’t seem to make a difference
Summary Comments :
Fig 3a – Can’t compare full model as don’t have it, and unsure about how to verify effect on
shaft speed 0.06
Fig 3 and 4 are done without the Freq Signal delay block. Is this of importance?
4b shows variation in gain
General comment on paper : Fig3,4 work doesn’t quite join up to Fig 6,8
Extra Questions :
(Possibly Minor)
Sync Droop appears set at 11%, 1/11 not 4 % ?
Procedure : Run the setup file then the simulation file.
Page 78 of 112
7.1.2 Setup
The Simulink Solver settings are as shown in the dialogue below
Page 79 of 112
7.1.3 M-Files
Inertia_setup.m
%Frequency response from wind turbines - JE NJ GS 2008
%Constants for simplified WT inertia model
%12 March 2009
%Ian Moore

%This file works out the parameters to manually insert in the main model
%The values are taken from Janakas paper
%Later models have their own parameters input so some of the control
%parameters below are not used. e.g K1 controller gain etc

% GB sys
Heq = 4.55
D = 1

%2MW IG/DFIG Wind Turbine in p.u
S=2e6 % VA
Rs=0.00491 % Stator Resistance
Rr=0.00552 % Rotor Resistance
Xls=0.09273 % Stator Reactance
Xlr=0.1 % Rotor Reactance
Xm=3.96545 % Magnetising Reactance
H=4.5 % Lumped Inertia Constant

Lm=Xm % Because X=wL at 1pu w
Xr=Xlr % Xlr
Lr=Xr % Rotor Leakage Inductance

%Control Model Parameters, Voltage path v_qr / v_qs
kp=0.5 % All in pu
ki=0.5

% Supplementary Control Gains in pu
K1=3.0
K2=1
Tw=1

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%Work out the Inertia from H given
w_e=1500*2*pi*2/60 % Rotational speed electrical, 2 pairs of poles on
this machine
w_m=(1500*2*pi)/60
J=(2*H*S)/(w_m*w_m) % kg.m sq
%Dont actually use this, see notes
%Initial power output for integrator 2MW
%LUT

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Lss=Xm+Xls
Lrr=Lr+Lm
L0=(Lrr-((Lm*Lm)/Lss))

%%%%%%%%%%%%%%%%% Choose IG or DFIG %%%%%%%%%%%%%%%%
%X1=Lss/Lm %DFIG
%X2=1/Rr
%X3=Lm/Lss
Page 80 of 112
%T1=L0/(w_e*Rr)

X1=Lrr/Lm %IG
X2=1/Rs
X3=Lm/Lrr
T1=Lss/(w_e*Rs) %Note its w_s in paper

%%%%%%%%%%%%% Convert X2/[1+sT1] to 1/(as+1) format
s_term = T1/X2
non_s_term = 1/X2
%%%%%%%%%%%%%%% END %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%


Page 81 of 112
GB_Fast_PR_setup.m
%Set some defaults so we can run a sim directly from the GUI
%without mfile having to run an mfile first
%Note these are global so we can use the workspace editor to change
%paramaters quickly rather than changing an m-file

global Ld_Db; Ld_Db = 0.0189;
global sync_droop; sync_droop = -11;
global GW_WT; GW_WT = 19.4;

global K1; K1 = -3;
global K2; K2 = 1;
global Tw; Tw = 1;
global Kf; Kf = 500;
global Kf_Lag; Kf_Lag = 20;
global f_Lag_ON; f_Lag_ON = 0;

global B_K1; B_K1 = -3;
global B_K2; B_K2 = 1;
global B_Tw; B_Tw = 1;
global B_Kf; B_Kf = 500;
global B_Kf_Lag; B_Kf_Lag = 20;
global B_f_Lag_ON; B_f_Lag_ON = 0;
Page 82 of 112
GB_Fast_PR_sim.mdl
% File name - GB_fast_PR_sim.m
% Ian Moore - 20/4/09
% Requires - GB_fast_PR_setup.m Run once first
% Does - Runs the mdl sim and plot results
function GB_fast_PR_sim
clear
mdl_file = 'GB_fast_PR_3'
%global sigsOut
global tout;
global f;
global Te_WT;

global param;

global Ld_Db;
global sync_droop;
global GW_WT;

global K1;
global K2;
global Tw;
global Kf;
global Kf_Lag;
global f_Lag_ON;

global B_K1;
global B_K2;
global B_Tw;
global B_Kf;
global B_Kf_Lag;
global B_f_Lag_ON;

% Set the plots to cycle through linestyle by default
set(0,'DefaultAxesColorOrder',[0 0
0],'DefaultAxesLineStyleOrder','-|--|-.|:')

%%%%%%%%%%%%%%%%%%%%%%%%%% 3a OL plot of f %%%%%%%%%%%%%%%%%%%%%%%%%%%%
figure(1); clf; %Clear the figure
% [ Ld_Db, sync_droop, GW_WT, k1, k2, Tw, Kf, Kf_Lag, f_Lag_ON]
param = [ 0.005, 0, 0 , -3, 1, 1, 500, 20, 1 ];
update(); sim(mdl_file);
plot1 = plot(f.time,f.signals.values,'b');xlabel('Time
[s]');ylabel('Frequency [Hz]');
xlim([70 110]);ylim([49.75 50]);
%title('Power System Frequency Deviation - Fig3a');
hold all; grid on; %Hold plot and cycle line colours
%legend('Frequency (Hz)');
print -f1 -r600 -djpeg fig3a;hgsave('fig3a');
%%%%%%%%%%%%%%%%%%%%%%%%%% 4a Plot %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%

figure(2); clf; %Clear the figure

%param = [ 0.005, 0, 0, -3, 0, 1, 500, 20, 1 ]; update(); sim(mdl_file);
param = [ 0.005, 0, 0, -3, 0.5, 1, 500, 20, 1 ]; update(); sim(mdl_file);
plot1 = plot(Te_WT.time,Te_WT.signals.values,'b');xlabel('Time
[s]');ylabel('Wind Turbine Torque [p.u]');
xlim([70 110]);ylim([0.54 0.74]);
Page 83 of 112
%title ('Open Loop Turbine Torque for 0.25hz Freq Deviation- Fig 4a');
hold all; grid on; %Hold plot and cycle line colours

param = [ 0.005, 0, 0, -3, 1, 1, 500, 20, 1 ]; update();
sim(mdl_file);plot_Te_WT('r');
%param = [ 0.005, 0, 0, -3, 2, 1, 500, 20, 1 ]; update();
sim(mdl_file);plot_Te_WT('m');

%legend('K2 = 0, K1 = 3, Tw = 1','K2 = 0.5','K2 = 1.0','K2 = 2.0');
legend('K2 = 0.5, K1 = 3, Tw = 1','K2 = 1.0, K1 = 3, Tw = 1');
print -f2 -r600 -djpeg fig4a;hgsave('fig4a');
%%%%%%%%%%%%%%%%%%%%%%%%%% 4b Plot %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%
%Bigger Disturbance
figure(3); clf; %Clear the figure
param = [ 0.01, 0, 0, -3, 0, 1, 500, 20, 1 ]; update(); sim(mdl_file);
plot1 = plot(Te_WT.time,Te_WT.signals.values,'b');ylabel('Wind Turbine
Torque (p.u)');
xlim([70 110]);ylim([0.4 1.2]);
title ('Open Loop Turbine Torque for 0.5hz Freq Deviation - Fig 4b');
hold all; grid on; %Hold plot and cycle line colours

param = [ 0.01, 0, 0, -3, 0.5, 1, 500, 20, 1 ]; update();
sim(mdl_file);plot_Te_WT('r');
param = [ 0.01, 0, 0, -3, 1, 1, 500, 20, 1 ]; update();
sim(mdl_file);plot_Te_WT('g');
param = [ 0.01, 0, 0, -3, 2, 1, 500, 20, 1 ]; update();
sim(mdl_file);plot_Te_WT('m');
legend('K2 = 0, K1 = 3, Tw = 1','K2 = 0.5','K2 = 1.0','K2 = 2.0');
print -f3 -r600 -djpeg fig4b;hgsave('fig4b');
%%%%%%%%%%%%%%%%%%%%%%%%%% Fig 6 Plot %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%

figure(4); clf; %Clear the figure
param = [ 0.0189, 0, 0, -3, 1, 1, 500, 20, 1 ]; update(); sim(mdl_file);
plot1 = plot(f.time,f.signals.values,'b');xlabel('Time
[s]');ylabel('Frequency [Hz]');
xlim([70 110]);ylim([49 50.2]);
%title ('2020 High Wind System frequency Drop, No Synchronous Response,
with and without Wind Response for 1320GW loss');
hold all; grid on; %Hold plot and cycle line colours
param = [ 0.0189, 0, 19.4, -3, 1, 1, 500, 20, 1 ]; update();
sim(mdl_file);plot_f('r');
legend('Without WT frequency response','With WT frequency response');
print -f4 -r600 -djpeg fig6;hgsave('fig6');
%%%%%%%%%%%%%%%%%%%%%%%%%% Fig 8 Plot %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%

figure(5); clf; %Clear the figure
param = [ 0.0189, -11, 0, -3, 1, 1, 500, 20, 1 ]; update(); sim(mdl_file);
plot1 = plot(f.time,f.signals.values,'b');xlabel('Time
[s]');ylabel('Frequency [Hz]');
xlim([70 150]);ylim([49.75 50]);
%title ('2020 High Wind System frequency Drop, Including Synchronous
Response, with and without Wind Response for 1320GW loss');
hold all; grid on; %Hold plot and cycle line colours
param = [ 0.0189, -11, 19.4 , -3, 1, 1, 500, 20, 1 ]; update();
sim(mdl_file);plot_f('r');
legend('Without WT frequency response','With WT frequency response');
print -f5 -r600 -djpeg fig8;hgsave('fig8');

Page 84 of 112
%{
%%%%%%%%%%%%%%%%%%%%%%%%%% Fig FR Control %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%
Add workspace outs
figure(6); clf; %Clear the figure
param = [ 0.0189, -11, 19.4, -3, 1, 1, 500, 20, 1 ]; update();
sim(mdl_file);
plot1 = plot(f.time,f.signals.values,'b');ylabel('Frequency (Hz)');
%set(plot1,'XLim',[70 130],'YLim',[48.8 50.2])
title ('2020 High Wind System frequency Drop, Including Synchronous
Response, FR Control');
hold all; grid on; %Hold plot and cycle line colours
legend('Without WT frequency response','With WT frequency response');
%}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function plot_Te_WT(colour)
plot(Te_WT.time,Te_WT.signals.values,colour)
end

function plot_f(colour)
plot(f.time,f.signals.values,colour)
end

function update
Ld_Db = param(1);
sync_droop = param(2);
GW_WT = param(3);
K1 = param(4);
K2 = param(5);
Tw = param(6);
Kf = param(7);
Kf_Lag = param(8);
f_Lag_ON = param(9);
end

end
%miscellaneous commands
%get(plot1);xlim;
Page 85 of 112
G B _ f a s t _ P R _ 1 . m d l
- 1 1
1 9 . 4 / 6 3 . 5 4
k 1
k 2
I a n M o o r e - 1 4 A p r i l 2 0 0 9
G B S y s w i t h F a s t P R f r o m W i n d
G B _ f a s t _ P R . d o c
f r e q s i g l a g
1 / 3 . 5
1
0 . 3 s + 1
2 s + 1
1 2 s + 1
1
0 . 0 1 2 9 s + 0 . 0 0 4 9 1
1
0 . 2 s + 1
1 0 s + 1
s
f 1 . m a t
S w i t c h
N o o f m / c
- K -
L o o k - U p
T a b l e
1
s
- K -
- K -
- K -
d u / d t
0
1
0 . 6
0 . 6
0 . 9 5
C l o c k
5 0
- K -
1
- 3
1
2 0 s + 1
1
s + 1
s
s + 1
2 H t e r m
1
6 . 2 1 s + 1
w _ d
P d _ s y n c
P d _ W i n d
P d _ W i n d
T e _ W T T e _ W T T e _ W T
T e _ W T
P d _ W T P d _ W T
P d _ W T
T _ d e m
T _ d e m
f w _ r
Figure 7.71 – Complete frequency response model in Matlab-Simulink
Page 86 of 112
7.1.4 Model Parameters
Synchronous plant Parameters taken from [3]
Governor =
]
]
]



+ 1 2 . 0
1
s
Turbine =
]
]
]



+
]
]
]



+
+
1 3 . 0
1
1 12
1 2
s s
s
Droop =
]
]
]



11
1
2MW induction wind turbine model parameters
Stator resistance (R
s
) : 0.00491 pu
Rotor resistance (R
r
) : 0.00552 pu
Stator reactance (X
ls
) : 0.09273 pu
Rotor reactance (X
lr
) : 0.1 pu
Magnetising reactance (X
m
) : 3.96545
Lumped inertia constant (H) : 4.5 sec
Controller Parameters
k
p
= 0.5, k
i
= 0.5
Closed Loop Simulation Parameters
T
w
= 1, k
1
= 3, k
2
= 1, k
f
= 500, T
f
= 20
Calculation of Inertia
H
eq
= ∑
· ,.... ,
*
gas coal i
sys
i
i
S
S
H
Inertia constant H is the kinetic energy in watt-seconds divided by the VA base where ω
0m
is the rated angular
velocity in rad/s.
H =
base
m
VA
J
2
0
2
1 ω
TABLE 7.3 - PARAMETERS FOR SIMPLIFIED WIND TURBINE MODEL
Turbine
Type
X1 X2 X3 T1 or T2
DFIG

m
ss
L
L
r
R
1
ss
m
L
L
r s
R
L
ω
0
Page 87 of 112
FPC (IG
based)

m
rr
L
L
s
R
1
rr
m
L
L
s s
R
L
ω
0
TABLE 7.4 - PLANT MARGIN AND OPERATING CAPACITY
Generator Type Scenario - High Wind 2020
Installed Capacity
(GW)
Plant Margin
(GW)
Operating Capacity
(GW)
New Coal 3.7 1.3 2.41 FR
Coal 16.9 7.61 9.3 FR
Gas 27.3 12.29 15.02 FR
Nuclear 6 0 6 FR
Interconnector 3.3 0 3.3 FR
Other 6.8 2.04 4.76 FR
FR Sync Cap = 40.78
Onshore Wind 14.3 5.72 FR
Offshore Wind 34.2 13.68 FR
Other 5.6 3.36 No FR
Total Capacity 118.1 63.54
FR PEI based = 19.4
TABLE 3
H
EQ
ON SYSTEM BASE 63.5MVA
Generator Type Scenario - High Wind 2020
Capacity(GW) Hi Heq
New Coal 2.41 4.5 0.17
Coal 9.3 4.50 0.66
Gas 15.02 6.00 1.42
Nuclear 6.00 3 0.28
Interconnector 3.30 0 0.0
Other 4.76 4.5 0.34
Onshore Wind 5.72 0 0
Offshore Wind 13.68 0 0
Other 3.36 4.5 0.24
Total 63.54 3.11
Page 88 of 112
7.2 Laplace Transformation
For a rotating machine we know per unit torque is synonymous with per unit power if there is
no change in speed. Although real plant steam turbines or diesel generating plant will have a
particular Torque Speed characteristic we assume here that the plant will adjust its torque
upwards slightly to maintain its setpoint Power output when a system frequency drop occurs.
This enables us to present the modified swing equation X describing the relationship between
power and frequency on the system.
dt
d r ) ( ω ∆
=
H 2
1
(
load gen P P −
- K
D

r ω
)
In terms of change in frequency and ∆P we can replace to give
dt
f d ) (∆
=
H 2
1
(
P ∆
- K
D
∆ f )
This is converted into the S-Domain as below :
Re-arranging to separate input and output
2H
dt
f d ) (∆
+ K
D
∆ f =
P ∆
For simplification of presentation (as is common with most authors) from now on we dispense
with superbars as Power System load studies normally use per unit quantities.
Taking the Laplace transform with all initial conditions as zero gives
2Hs ∆F(s) + K
D
∆F(s) = ∆P(s)
where for example the notation ∆F(s) indicates the term is the Laplace transform of the time
domain function ∆f(t) and where s is a constant with the unit of 1/t
Collecting the terms gives
Page 89 of 112
∆F(s) ( 2Hs + K
D
) = ∆P(s)
The transfer function of the system is
G(s) = ∆P(s) / ∆F(s)
= 1 / 2Hs + K
D

This equation can be seen implemented in Figure 4.42
Page 90 of 112
7.3 Experimental Test Rig
7.3.1 Procedure for Use
Setup
The basic procedure to create, implement and run a control scheme for the test rig is as
follows:
• Create Matlab-Simulink model to include appropriate dSPACE blocks
• Create object code using Real-Time Workshop from within Matlab-Simulink
• Download code (.sdf file) to the target system (DS 1103PPC)
• Create appropriate instrumentation layout from within dSPACE ‘control desk’ for
viewing and/or modification of system data/parameters.
• Start execution of embedded controller code and start ‘animation’ of instrument
panel
Please take of note the items in the following subheadings regarding the overall process of
development.
Equipment Initialisation
Make sure Ethernet link cable from controller to PC is connected before booting the PC.
Matlab Simulink Model
Consider using ‘output’ connectors on sub system models and top-level blocks.
Visibility of parameters to other blocks and also modification of these.
Consider use of workspace parameters with respect to ability to modify parameters ‘in-line’.
Organise a method for version control of Matlab models and corresponding instrument layout.
dSPACE Instrument Layout
Consider re-usability of instrument layouts and appropriate naming.
Object Code
No comments.
Control System Operation
Observe appropriate operation for each item of equipment as below :
Pendulum Controller
This machine has no emergency stop. WARNING - Use of the rocker switch 13 [LD Manual]
will result in the machine only temporarily returning to a zero setpoint if the external control
input is used. Return switch 11 to the ‘internal’ setting position to avoid this immediately after
performing an emergency shutdown of the pendulum machine with this switch.
Page 91 of 112
Table 7.5 – Convention taken for rotation for test rig
Reference : Looking at rear housing of the Pendulum machine case, output shaft
facing away
Quadrant Pendulum Rotation
1 motor Clockwise
2 Generator Counter clockwise
3 Motor Clockwise
4 generator Counter clockwise
7.3.2 Equipment Specifications
Interconnection of the LD equipment uses the signal scaling indicated in the table below
Table 7.4 – Signal scaling and naming for dc pendulum controller
Page 92 of 112
Signal Name Description Type Design Scaling Measured Direction Comments
M in Torque
transducer
Sensor Output 1 V / 3 Nm Taking U1
as + ?
Q1 op gives ?
EXTERN Speed/Torque
setpoint
Control Input 10 V / 3500
rpm / 22 Nm /
20A
+ 4.32 V for
1495 rpm
Speed + CW Rate limited
n out Speed sensor
output
Conditioned
sensor output
+/- 14 V
max/min
+11.7 V @
1500
(table 4)
+ CW Gain
Adjustable
Mout Torque sensor
output
Conditioned
sensor output
+/- 14 V
max/min
Q1 op gives -
V
Note U1,U2
is +
Gain
Adjustable
Synchronous Machine
4- Pole non salient with damper winding, Class 1
Table 7.6- Nameplate data
V Y/Delta 400/230
A 1.52 / 2.66
kW 0.8 kVA / 0.8
Cos phi 0.8 – 1 – 0.8
U/min 1500 50Hz
Uerr = 220V Max 1.6A
Is B/F IP20
Thermal CB included
Terminals :
Stator - U1 U2, V1 V2, W1 W2
Excitation - F1, F2
Excitation Voltage Controller
LD 745 021
Rectified AC by PWM of GTO/Thyristors
Table 7.7 - Nameplate data
Output [+, -] DC 0 - 200V 1A max
Input [0..10V, 0V] Manual step buttons, TTL/24V edge
triggered steps
Protection S/C and O/L
Page 93 of 112
DC Pendulum machine
Part no. PM 732 68
DC Shunt wound
Table 7.8- Nameplate data
150 – 300V Arm max 8.5A
1.0 – 2.0 kW
1500 – 3000 min-1
Err 200V
0.7A max
I KL B/F IP20
Thermal CB included
Torque Output [U1, U2] 1 Vdc / 3 Nm
Terminals
Armature (rotor) A1, A2
Field (stator) E1, E2
Figure 7.72 – Shunt Field dc motor equation [21]
Page 94 of 112
Pendulum machine control unit
LD 732 695
uP controlled
Table 7.9– Control Modes
modes
TORQUE CONTROL Closed Loop Torque
SPEED CONTROL Closed Loop Speed
UNCONTROLLED Uncontrolled (actually
armature current ctrl)
When sync m/c is on mains
grid?
Load char + Run-up char Auto record of run-up and
load
Figure 7.73 - Pendulum Controller Front Panel [
31
]
Page 95 of 112
dSPACE Controller Board
Table 7.10– dSPACE Controller Board Specifications [27]
Page 96 of 112
Page 97 of 112
7.3.3 Hardware Design
Additional MOSFET Circuitry
As supplied by LD-Didactic the MOSEFET modules come are configured as below (note
these are the original BUZ 73 versions)
Figure 7.74 Internal wiring for LD Mosfet Units – Drawn by Paul Farrugia
Note diagram is for the version supplied with 7A transistors and may have changed.
Gate Drive Boards
A typical connection of the IR2133 device is shown below
Figure 7.75 – Typical connection for IR2133 [32]
Page 98 of 112
A functional block diagram for this device is shown below :
Figure 7.76 – Internal functioning of IR2133 [
32
]
Overview of internal functioning :
Schmitt Triggers : Removes noise from logic input to signal to give a clean rising edge or
falling edge.
Level translator and PW discriminator : Couples input logic to internal level with necessary
noise immunity. e.g must tolerate Vss dropping below Vcom which can happen in practical
gate drive layout implementations. To reduce noise a Pulse Width (PW) discriminator filters
the gate switching signal.
Pulse Generator : For the top gate drive the signal is changed into a pulse format in order to
reduce power consumption used in the level translator stage.
Delay : For the lower gate drive the signal passes via a delay to ensure a minimum deadtime
(hence protecting against ‘shoot-through’) and then directly on to the ‘totem pole’ output
stage.
Pulse Discriminator : Changes the pulse format back into a square wave format.
Vdd/Vbs Level Translator : Raises the internal logic level suitable to drive the upper gate
drive transistors which are sitting at Vs (i.e same as the dc link voltage).
The following circuit was designed in accordance with recommendations given in [24]
Page 99 of 112
1
2
J5
1
2
J6
1
2
J7
1
2
J8
1
2
J9
1
2
J10
C14
1uF
C12
1uF
C15
1uF
R9
680R
R10
680R
R11
680R
R12
680R
R13
680R
R14
680R
C5
150pF
C7
150pF
C6
150pF
C10
150pF
C8
150pF
C9
150pF
1
2
4
3 5
6
7
8
HCPL2531
U6
1
2
4
3 5
6
7
8
HCPL2531
U7
1
2
4
3 5
6
7
8
HCPL2531
U8
Isolation Barrier
<--- Signal Power -->
GND
Place Bootstrap Capacitors at ICPins
VSS
RV1
20R
U
s
e

T
w
i
s
t
e
d

P
a
i
r

C
a
b
l
e
VCC
21
HIN1
22
HIN2
23
HIN3
24
LIN1
25
LIN2
26
LIN3
27
FAULT
28
ITRIP
1
CAO
3
CA-
4
VSS
7
LO2
10
LO1
11
CA+
5
VS3
12
HO3
13
VB3
14
VS2
15
HO2
16
VB2
17
Fault CLR
2
VS1
18
HO1
19
LO3
9
VB1
20
SD
6
COM
8
U9
IR2133
LIN3
LIN2
HIN3
HIN2
LIN1
HIN1
HIN1
HIN2
HIN3
LIN1
LIN2
LIN3
VS2 VS3
VB2
VS1
VB1 VB3
HO1
VS1
VB2
HO2
VS2
VB1
VB3
HO3
VS3
LO1
LO2
LO3
HO1
HO2
HO3
LO1
LO2
LO3
VS1
VS2
VS3
COM
COM
COM
C11
47uF50V
VCC
VSS
COM
ITRIP
VCC
Fclr
SD
VSS
SDI/P
Fclr I/P
ITRIP
VSS
VSS COM
COM VSS
Csen
1
2
J12
Csen
C13
0.1uF50V
D7
YellowLED
VSS S1
Standby
1
2
4
3 5
6
7
8
HCPL2531
U4
SD
VCC
GND
Fclr
R8
680R
R6
680R
C4
150pF
C3
150pF
HIN, LIN &FLTCLRall have100K internal pullups
SD will require a pullup
1
2
4
3 5
6
7
8
HCPL2531
U5
FAULT
FAULT
1 2 3 4 5 6 7 8 9
1
0
P1
S2
SD
S3
Fclr
1K
R19
Q2
BC108
1K
R20
SDI/P
Fclr I/P
1K
R15
Q4
BC117
1K
R32
1.5K
R25
LED
D9
RED
VSS
VSS
Q3
BC117
1K
R22
1K
R23
F
A
U
L
T
VCC
10K
R21
15K
R24
D3
1N4001
LED
D8
Green
POWERI/P
33 R26
33 R27
33 R28
27 R29
27 R30
27 R31
0.33
Farnell: 1174239
RShunt
1K
R18
D4
11DF4
D5
11DF4
D6
11DF4
9K R17
1K
R16
1
1
2
2
3
3
4
4
RAC05 PSU
RS: 193-301
PSU
1
2
J11
VCC
Page 100 of 112
Gate Driver Board Design Version 1 - Denley Slade 2009
Current and Voltage Sensors
Voltage sensor circuit using LEM LV25-P
Figure 7.77 - [
33
]
Current Sensing circuit using LEM - LTS 15 NP
Figure 7.78 - [
34
]
Page 101 of 112
Cabling
Table 7.11 – Slave I/O connections [DS1103 Hardware Installation and Configuration Nov
2007, dSPACE ControlDesk Help File]
Page 102 of 112
Cable Colour coding for the 1
st
gate drive is as follows :
5 Yellow ST2 PWM
7 Red HIN 1 26 Orange LIN 1
8 Green HIN 2 27 White LIN 2
9 Blue HIN 3 28 Grey LIN 3
19 Purple VCC(+5V) 37 Black GND
Page 103 of 112
8 References
Page 104 of 112
1
[
[
] – BWEA, “Wind hits 4GW barrier - now powers 2.3 million homes in UK”,
http://www.bwea.com/media/news/articles/pr20091020.html , accessed 19 November 2009

2
[
[
] - Erinmez, I.A., Bickers, D.O., Wood, G.F., Hung, W.W., “NGC Experience with frequency
control in England and Wales – Provision of frequency response by generators”, IEEE PES Winter
meeting, 31 January – 4 February 1999, New York USA.
3
[
[
] – Manwell J.F, McGowan J.G, Rogers A.L, “Wind Energy Explained – Theory, Design and
Application”, John Wiley & Sons, 2002.
4
[
[
] –Burton T, Sharpe D, Jenkins N, Bossanyi E, “Wind Energy Handbook”, John Wiley & Sons,
2001
5
[
[
] - Ramtharan, G.; Ekanayake, J.B.; Jenkins, N., “Frequency support from doubly fed induction
generator wind turbines”, IET Renewable Power Generation, Volume 1, Issue 1, March 2007, pp.
3-9.
6
[
[
] – Fox B, et al, “Wind Power Integration – Connection and system operational aspects”, IET,
2007
7
[
[
] – Anaya-Lara O, et al, “Wind Energy Generation – Modelling and Control”, John Wiley & Sons,
2009
8
[
[
] – Ackermann T, et al, “Wind Power in Power Systems”, John Wiley & Sons, 2005
9
[
[
] – Caliao D, “Modelling and Control of a Fully Rated Converter Wind Turbine” Thesis submitted
to University of Manchester, 2008
10
[
[
] – Bose B, “Modern Power Electronics and AC Drives”, Pearson Education, 2002
11
[
[
] – Bossanyi E, “GH Bladed - Theory Manual”, Issue no.17, Garrad Hussan Partners Ltd, Bristol,
2007
12
[
[
] – Krause P, et al, “Analysis of Electric Machinery and Drive Systems”, 2
nd
Edition, IEEE, 2002
13
[
[
] – Kundur P, “Power System Stability and Control”, EPRI, 1994
14
[
[
] - Ramtharan, G., Jenkins, N., Anaya-Lara, O., “Modelling and Control of Synchronous
Generators for Wide-range Variable-speed Wind Turbines”, Wind Energy, Vol 10, 2007, pp 231–
246
15
[
[
] - G Ramtharan, “Control of variable speed wind turbine generators”, PhD Thesis, 2008.

16
[
[
] - Seul-Ki Kim, Eung-Sang Kim, Jae-Young Yoon and Ho-Yong Kim, “PSCAD/EMTDC Based
Dynamic Modeling and Analysis of a Variable Speed Wind Turbine”, IEEE Power Engineering
Society General Meeting, 2004. Volume , Issue , 6-10 June 2004 Page(s): 1735 - 1741 Vol.2

17
[
[
] - National Grid Company plc, “The Grid Code,” Issue 3, Revision 25, 1 February 2008.
18
[
[
] – Banham-Hall D, et al, “Grid Connection Oriented Modelling of Wind Turbines with Full
Converters”, UPEC 2009
19
[
[
] - J.B.Ekanayake N.Jenkins G.Strbac, “Frequency response from wind turbines”, Wind
Engineering, Vol. 32, pp. 573-586. 2008
20
[
[
] – Milanovic J, Kayikci M, “Dynamic Contribution of DFIG-Based Wind Plants to System
Frequency Disturbance”, IEEE Transactions on Power Systems, Vol 24, no.2, May 2009
21
[
[
] – Bolton W, “Control Engineering”, 2
nd
edition, Longman, 1998
22
[
[
] – Caliao N, Tennakoon A, “Control Systems Development and Implementation with dSPACE
and Wind turbine test rig, Internal report, Cardiff University
23
[
[
] – Mohan N, et al, “Power Electronics: Converters, applications and design”, John Wiley and
Sons, 1989
24
[
[
] – International Rectifier, “Application Note AN-978 – HV Floating MOS-Gate Driver ICs”, IR,
RevD
25
[
[
] - Merello A, et al, “Design Tips – Using monolithic high voltage gate drivers DT04-4revA”,
International Rectifier
26
[
[
] – Heier S, “Grid Integration of Wind Energy Conversion Systems”, John Wiley and Sons, 2
nd
Edition, 2006
27
[
[
] - dSPACE, “DS1103PPC Controller Board”, dspace_2008_ds1103_en_pi777.pdf, downloaded
Dec 2008, http://www.dspaceinc.com
28
[
[
] - TecQuipment, “PSS1 – Power System Simulator - Operator Manual”, TecQuipment,
Nottingham UK
29
[
[
] – Valentine R, “Motor Control Electronics Handbook”, McGraw-Hill, 1998
30
[
[
] – Williams T, “The Circuit Designers Companion”, Newnes, 1991
31
[
[
] – Leybold Didactic Gmbh, “Instruction Sheet – Control Unit for Pendulum Machine 732 695”,
version 04/95 –Li

32
[
[
] – International Rectifier, “IR2133 Datasheet PD60107 revX”, downloaded March 2009,
http://www.irf.com/
33
[
[
] – LEM, “LTS 15-NP Datasheet”, downloaded 20 Nov 2009, http://www.lem.com
34
[
[
] - LEM, “LV-25P Datasheet”, downloaded 20 Nov 2009, http://www.lem.com

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close