power control of voltage source converter for distributed generationconverter control renewable...
DESCRIPTION
Este trabalho trata do controle de potência ativa e reativa dos conversores de tensão para geração distribuída..Apresenta uma simulação para Matlab.TRANSCRIPT
-
PHYSCON 2011, Leon, Spain, September, 5September, 8 2011
POWER CONTROL OF VOLTAGE SOURCE CONVERTERFOR DISTRIBUTED GENERATION
Jose Luis Domnguez-GarcaElectrical Engineering Area
IRECSpain
Oriol Gomis-BellmuntCITCEA-UPC
Fernando BianchiElectrical Engineering Area
IRECSpain
Antoni Sudria`-AndreuCITCEA-UPC
AbstractThe paper deals with the active and reactive power
control of voltage source converters (VSC) for dis-tributed generation. The control method explained isbased on the instantaneous power theory. The powerconverter regulates the DC-Bus voltage and also con-trols the reactive power injected to the grid. The pre-sented control strategy is evaluated by simulation inMatlab-Simulink.
Key wordsDistributed Generation, Converter control, PLL, Cur-
rent loop, qd0 Transform
1 IntroductionThe rapid increment of renewable power generation
is changing the power system shape, and creating newpower generation concepts as distributed generation.Distributed generation can be defined, according to[Ackermann, Andersson and Soder, 2001], as an elec-tric power source connected directly to the distribu-tion network or on the consumer side of the meter.On the other hand, CIGRE working group 37-23 de-fines it as low rating generation that is neither plannednor dispatched centrally and is usually connected tothe distribution network [CIGRE working group 37-23,1999]. Therefore, from the main ideas of both defini-tions, distributed generation can be understood as lowpower generation that is connected close to the distri-bution network. The power generation technologies in-volved in distributed generation include, for example,wind turbines, small hydro turbines, combined heat andpower (CHP) units, also known as cogeneration, fuelcells and photovoltaics (PV) cells.In order to enhance the quality of the power injected tothe grid by the units of distributed generation, a powerconverter is used between the grid and the generator en-abling the system to be regulated.Different control systems have been studied in order to
provide different grid supports or ride through faultscapability, e.g., frequency support [Hirodontis, Anaya-Lara, Burt and McDonald, 2009][Karlsson, Bjornstedtand Strom, 2005], to guarantee power quality [Hornikand Zhong, 2009], voltage dips [Wang, Duarte andHendrix, 2009], etc. The present paper studies andpresents an active and reactive power control of a con-verter to enable the system to improve the quality of thepower delivered.This paper is organized as follows. In Section II, thesystem under study is described. The Clarke and ParkTransforms are presented in Section III. In section IV,the instantaneous power theory is briefly introduced.The control scheme is developed in section V. In sec-tion VI, the proposed control is validated by means ofsimulation and discussed. Finally, the conclusions aresummarized in Section VII.
2 System descriptionThe system under study is shown in the Figure 1. It is
a power converter transforming the active power deliv-ered by a dc-source, into the active and reactive powersent to the network. The dc-source can be either a pho-tovoltaic cell, a wind turbine connected to a rectifier ora fuel cell.To transform from DC to three-phase AC, it is neces-sary a three-phase inverter. The design of this powerconverter can be done by means of different technolo-gies as diode inverter, thyristor inverter or IGBT in-verter. Because of the ability to regulate active and re-active power, the IGBT inverters are commonly chosento regulate the power delivery.
3 Transformation from abc reference frame toqd0
Current and voltage are usually described as threesinusoidal waves representing the three AC phases (a,b and c). Calculations in abc-frame are complex due
-
iDCs iDCl
iDCE
+ CE
LlD
C SO
URC
Evla
vlb
vlc
Ll
Ll
ila
ilb
ilc
DC
AC
Figure 1. Scheme of the system
to the sinusoidal components and its variability in thetime-frame.Therefore, in order to facilitate the control designof three-phase inverters the qd0 reference frame isused. In the case of balanced three-phase circuits, theapplication of the Park Transform to three AC variablesreduces to two DC variables, due to the 0-componentin balanced conditions is zero. Thus, simplified cal-culations can be carried out on these transformed DCvariables. Three-phase AC components are recoveredwith the inverse transform.
3.1 Clarke TransformationThe Clarke Transformation allows to change the
three dimension vector into two dimension vector(abc-frame to .frame).
(xx
)=
2
3
(1 12032
32
)xaxbxc
(1)This transformation can be understood as the pro-jection of the three axis onto a stationary two-axisreference frame, as it is shown in the figure 2. It isimportant to remark that it is possible to expand thevariable change matrix 1 to a 3x3 matrix 2, includinga 0 component (called homopolar component), whichis orthogonal to (, ) axis and usually a constantvalue (zero if the system is balanced). In this way thetransformation is bijective.
xxx0
= 23
1 12 120 32 3212
12
12
xaxbxc
(2)3.2 Park TransformationThe Park Transformation describes a rotation of an
orthogonal system ((, ) to (q, d)). The rotation isrepresented in the figure 3.
b
c
a
xabc
x
x
Figure 2. Representation of the Clarke Transformation
(xqxd
)=
(sin() cos()cos() sin()
)(xx
)(3)
q
d
xabc
xd xq
Figure 3. Rotation of the reference (, ) an angle
Again, the rotation can be expressed as a 3x3 matrixincluding the rotation orthogonal axis, which is thesame axis in (, , 0) and (q, d, 0) references, althougha vector rotation on a constant plain is expressed as a2x2 matrix.
xqxdx0
=sin() cos() 0cos() sin() 0
0 0 1
xxx0
(4)
-
3.3 qd0 TransformationThe qd0 Transformation includes the Clark and Park
Transformations, combining both transformations inone matrix.Hence, the Park Transformation with an angle of avector xabc R3) in the abc frame is given by
xqd0 = T ()xabc (5)
with
T () =2
3
cos() cos( 2pi3 ) cos( + 2pi3 )sin() sin( 2pi3 ) sin( + 2pi3 )12
12
12
(6)Then it can be said that the vector xqd0 is the vectorxabc in the dq0 frame on the angle reference.Since the matrix T () is invertible; the inverse trans-formation can be obtained as
xabc = T1()xqd0 (7)
with
T1() =
cos() sin() 1cos( 2pi3 ) sin( 2pi3 ) 1cos( + 2pi3 ) sin( +
2pi3 ) 1
(8)In conclusion, the qd0 transform can be interpreted asthe projection of the abc-frame variables onto a rotatingtwo-axis reference frame.
4 Instantaneous power theoryThroughout the paper, the instantaneous power the-
ory for balanced three-phase systems under sinusoidalconditions developed by Akagi [Akagi, Watanabe andAredes, 2007] is considered. A balanced three-phasesystem under sinusoidal conditions is defined as athree-phase system with equal amplitudes and 2pi/3 rad(120o) the angle between all the phases.Therefore, the instantaneous voltages and currents inabc frame can be written as:
xa =2Xcos(t+ )
xb =2Xcos(t+ 2pi/3)
xc =2Xcos(t+ + 2pi/3)
(9)
Electrical theory defines 3 different types of electricalpower: Apparent (S), Active (P) and Reactive (Q). TheApparent power in abc frame is expressed as: S = V I = P + jQWhere V and I are the phasor (phase vector) represen-tation of the voltage and the current.
Applying the Clarke and Park transformations de-scribed previously, the active and reactive power canbe expressed as:
P = 32 (vi + vi)Q = 32 (vi vi)
(10)
in the (,) frame and
P = 32 (vqiq + vdid)Q = 32 (vqid vdiq)
(11)
in the (q,d) frame.
5 Converter ControlThe purpose of the control of the power converter
is to regulate the reactive power and DC bus voltagewhich regulates active power. The DC bus voltage andthe reactive power references determine the currentreferences which determine the voltages to be appliedon the grid side.Figure 4 depicts the block diagram of grid sidecontrol. The reference currents are calculated bymeans of independent variables as active and reactivepower set-points, that is explained in the currentloops subsection. Furthermore, by means of the qd0Transformation, explained above, the control changeabc-frame currents to qd-currents, to compare themwith the reference currents. The angle necessary, tocalculate the Park Transformation and to synchronizethe system with the grid, is obtained by means thePLL, that is explained later, from the grid voltage.
DC
AC
Decoupling
Vlabc
ilabc
Q l*E*
Grid
Vzabc
+
-EPI
PE*
ild* ilq*
Vzq
PLL
Vzq
Vzd=0
grid
Vzq
x
/
x
/2/3 2/3
T()
ilqd
Ilqd*
+- PI
++
T()-1
Vlqd
Figure 4. Grid Control Scheme
5.1 PLLThe system control is designed in the qd frame. There-
fore, it is necessary to know which angle will be used
-
to obtain the desired reference. In this case, it is inter-esting to set the reference where voltage d-componentis zero.To find this angle it is usually used a phase locked loop(PLL). A basic scheme of a three-phase PLL, as it isshown in 5, consists in a comparison of the voltage d-component with a constant zero, the obtained error isregulated by a PI controller. The output of the con-troller is the grid frequency. The grid angle is obtainedby integration of the frequency. The grid angle is intro-duced in the Park Transform to calculate the qd voltagecomponents.The design of the controller parameters is detailed byS.K. Chung [Chung, 2000]
ParkTransformation
PI1s
+
Vd
Vabc
0
-
Vq
Figure 5. Scheme of Phase Locked Loop
5.2 Reference calculationThe q-axis may be aligned to the grid voltage, allow-
ing active and reactive decoupled control. To controlthe reactive power, a ild reference is computed as:
ild =2 Ql3 vzq (12)
The active power, which is responsible for the evolu-tion of the dc bus voltage, is controlled by the ilq. Alinear controller is usually designed to control the dcbus voltage and to keep it constant.
ilq =2 P E3 vzq (13)
The DC link voltage behavior is defined (in Laplacedomain) by the following equation
EDC = EDC0 +1
s C (iDCl iDCp) (14)
where the current values iDCp and iDCl can be
calculated as
iDCp =iabcp vabcp
E=PDCpE
(15a)
iDCl =iabcl vabcl
E(15b)
The control loop scheme is shown in Figure 6; The de-sign of both proportional and integral parameters of thePI controller is explained in [Junyent-Ferre`, 2007].
EPI
* IDCl*
E
+
-
IDCs
+
+
Figure 6. DC Bus Control Scheme
5.3 Current LoopsThe current control is implemented by the following
state linearization feedback:
(vlqvld
)=
(vlq + vzq Ll g ildvld + Ll g ilq
)(16)
where vlq and vld are the voltage control values. Thedecoupling can be described as:
d
dt
(ilqild
)=
( rlLl 00 rlLl
)(ilqild
)+
( 1Ll
0
0 1Ll
)(vlqvld
)(17)
5.4 Current Controller TuningThe controller is designed by means of the so-called
internal model control (IMC) theory [Harneforns andNee, 1998]. The parameters of a PI controller, with thegoal of obtain a desired bandwidth , which appear dueto a low-pass filter included by the IMC, are:
Kp = Ll Ki = rl (18)
6 Simulation ResultsThe system, previously described, is simulated with
Matlab-Simulink.The control of the converter is evaluated in twodifferent scenarios: a change in the active power
-
reference and a change in the reactive power reference.Also both PLL and qd0 Transformation response areevaluated to show how properly they work.
6.1 PLL & qd0 TransformIn the Figure 7, it can be observed that the Phase
Locked Loop (PLL) presented above has a good be-havior, reaching quickly the desired value. It is alsointeresting to note that the PLL controller behaves asa first order system. So, it satisfies the plant definitionthat appears in [Chung, 2000], where is related the PIcontroller parameters with the period of the system.
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1310
311
312
313
314
315
316
317
318
319
320Phase Locked Loop
time (s)
frequ
ency
(Hz)
Frequency CalculatedGrid Frequency
Figure 7. Dynamical response of the Phase Locked Loop system
In the Figure 8, can be observed an example of three-phase voltages in abc and what shape they take afterthe Park Transform. It is important to be noted thatthe chosen angle, which is given by the PLL controller,is the right one because the voltages in qd frame areconstant values.
1.5 1.55 1.6 1.65 1.7 1.751000
800
600
400
200
0
200
400
600
800
1000Park Transformation
time (s)
Volta
ge (V
)
Voltage abc frameVoltage qd frame
Figure 8. Application of Qd0 Transform to a sinusoidal balancedthree-phase voltage
6.2 Variation of Active PowerThe first scenario presented is a change of the active
power reference. Ensuring the capability of delivervariable active power, it is required due to the variabil-ity of the power demand from the consumers. In thisscenario the reactive power reference is set in zero.
In the Figure 9 can be seen the active power referencevarying with steps, and how the active power deliveredto the grid reaches quickly this new setting.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 512
10
8
6
4
2
0
2x 105 Active Power
time (s)
Activ
e Po
wer
(W)
Active Power to the gridReference Active Power
Figure 9. Plot of the change of the active power
Again, as it is shown in the Figure 10, the dc bus volt-age is following the defined value. The peaks appearon the graphic because of the control dynamics, whichacts during the reference value variations.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 51490
1492
1494
1496
1498
1500
1502
1504
1506
1508
1510DC Bus
time (s)
Volta
ge (V
)
Reference VoltageReal Voltage
Figure 10. Variation of the DC Bus Voltage driven by the activepower change
In Figures 11 and 12, it is plotted both q and dcomponents, either of real and reference currents. Thedifference between them is the input of the IMC tocalculate the voltage values. It can be seen that theq-component of the current presents the same behaviorof the active power profile, its calculation is driven bythe expression (13). In the same way, the d-componentwhich is guide by the expression (12) has a constantzero value due to the reactive power is set zero.
6.3 Change of Reactive PowerThe second scenario presented is the change of the
reactive power reference. Regulation of the reactivepower delivery allows the system to keep constant thevoltage in the connection point. In this scenario, theactive power reference is set in 1MW.In the Figure 13, it can be seen the reactive power ref-erence change, which as in the former case has a stepbehavior. As it is expected the reactive power delivered
-
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5900
800
700
600
500
400
300
200
100
0
100Current qcomponent
time (s)
Curre
nt (A
)
Real qcomponentReference qcomponent
Figure 11. Representation of the q-component of the current withactive power variations
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 525
20
15
10
5
0
5Current dcomponent
time (s)
Curre
nt (A
)
Real dcomponentReference dcomponent
Figure 12. Plotting of the d-component of the current
to the grid matches the new setting with a fast response.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 512
10
8
6
4
2
0
2x 105 Reactive Power
time (s)
Rea
ctiv
e Po
wer
(var)
Reactive Power to the gridReference Reactive Power
Figure 13. Variation of the reference value of the reactive power
In the Figure 14, the dc bus voltage is maintained atconstant value. Again, as in the first scenario, the volt-age presents peaks due to the reactive power changes,that affects the active power sent to the network.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 51485
1490
1495
1500
1505
1510
1515DC Bus
time (s)
Volta
ge (V
)
Reference VoltageReal Voltage
Figure 14. Behavior of the DC Bus Voltage under reactive powervariations
In the Figures 15 and 16, it is plotted the q and dcomponents of the currents. It can be seen that thed-component of the current presents a behavior similarto the reactive power profile. However,in this case,the q-component has not a constant value as it wouldbe expected, since the active power demanded is stillthe same. Then, as the voltage of the network remainsconstants, it is needed a change in the injected current.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5900
800
700
600
500
400
300
200
100
0Current qcomponent
time (s)
Curre
nt (A
)
Real qcomponentReference qcomponent
Figure 15. q-component of the current obtained by the active powerreference
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5900
800
700
600
500
400
300
200
100
0
100Current dcomponent
time (s)
Curre
nt (A
)
Real dcomponentReference dcomponent
Figure 16. d-component of the current obtained by the reactivepower reference
7 ConclusionsThis paper has presented an active and reactive power
control for converters in distributed generation sys-tems. The control of these converters demands the ma-nipulation of three-phase sinusoidal variables, whichmakes difficult to compute the control variables. Tosimplify the design Clark and Park transformationshave been used. In this transformed frame of reference,the control design can be cast as the regulation of DCvariables in two decoupled control loops based on sim-ple proportional-integral-derivative (PID) controllers.Simulation of the control methodology is presented inorder to illustrate the transformation of the variablesand also the ability of the system to regulate the activeand reactive power.
-
ReferencesAkagi, H., Watanabe, E.H and Aredes, M. (2007). In-stantaneous Power Theory and Applications to PowerConditioning. Wiley-Intersciences. New York.
Chung, S.K (2000) A phase tracking system for threephase utility interface inverters. IEEE Transactionson Power Electronics, Vol 15(7), pp. 431438.
Junyent-Ferre`, A. (2007) Modelitzacio i control dunsistema de generacio ele`ctrica de turbina de vent.Masters Thesis, ETSEIB-UPC.
Harneforns, L. and Nee, H.-P. (1998) Model-based cur-rent control of machines using internal model controlmethod IEEE Transactions on Industry Applications,Vol 34(1), pp. 133141.
Ackermann, T., Andersson, G. and Soder, L. (2001)Distributed generation: a definition Electric PowerSystem Research, Vol 57, pp. 195204.
CIGRE working group 37-23 Impact of increasing con-tribution of dispersed generation on the power systemTechnical Report CIGRE.
Karlsson, P., Bjornstedt, J. and Strom, M. (2005)Stability of voltage and frequency control in dis-tributed generation based on parallel-connected con-verters feeding constant power loads ProceedingsEPE, Dresden
Hirodontis, S., Anaya-Lara, O., Burt,G. and McDon-ald, J. (2009) Distributed generation control for fre-quency support Proceedings International Confer-ence on Electricity Distribution, 8-11 June Prague
Wang, F., Duarte, J.L. and Hendrix, M.A.M. (2009)Active and reactive power control schemes for dis-tributed generation systems under voltage dips Pro-ceedings Energy Conversion Congress and Exposi-tion, 20-24 September, San Jose
Hornik, T. and Zhong, Q. (2009) Control of grid-connected DC-AC converters in distributed genera-tion: experimental comparison of different schemesProceedings Compatibility and Power Electronics,
Mohan, N., Undeland, T.M and Robbins, W.P. (2003)Power Electronics: Converters, Applications and De-sign. Wiley. New York.