Vector Analysis and Optimal Control for the Voltage ...
Transcript of Vector Analysis and Optimal Control for the Voltage ...
Applied Physics Research; Vol. 10, No. 6; 2018 ISSN 1916-9639 E-ISSN 1916-9647
Published by Canadian Center of Science and Education
1
Vector Analysis and Optimal Control for the Voltage Regulation of a Weak Power System with Wind Energy and Power Electronics
Nick Schinas1
1 Technological Educational Institute of Western Greece, Patras, Greece Correspondence: Nick Schinas, Technological Educational Institute of Western Greece, Patras, Greece. E-mail: [email protected], [email protected] Received: September 14, 2018 Accepted: October 1, 2018 Online Published: November 30, 2018 doi:10.5539/apr.v10n6p1 URL: https://doi.org/10.5539/apr.v10n6p1 Abstract This paper deals with the voltage regulation in a weak system which contains large inductive loads and wind turbines using Doubly Fed Induction Generators (FDIGs). The DFIGs demand large amounts of reactive power from the grid and as a result, there is a voltage drop in the system which may be extra deteriorated if large inductive loads and motors are also present in the same line. The problem of the voltage regulation in these cases is treated with the installation of a Static Var Compensator (SVC) besides the capability of the DFIGs to partially regulate the voltage themselves. In this paper, new modeling procedures based on optimal control are developed for the design of the SVC controller and a novel strategy for the grid side converter of the DFIG is presented. The nonlinear system is simulated in the SIMULINK software so that the performance of the new controllers is validated. Keywords: wind turbine, voltage regulation, doubly fed induction generator, optimal control. 1. Introduction The increased need for wind energy development makes the installation of wind turbines in โweakโ ac grids necessary. On the other hand, many voltage instability incidents have taken place around the world the last years (Custem & Vournas, 1998; Berizzi, 2004). The distributed generation with wind power stations installed in weak distribution systems may enlarge this problem especially when large inductive loads are connected to the same line. So, voltage regulation has become a major research area in the field of power systems (Chondrogiannis, 2007; Ledesma, 2002; Kesraoui, 2016). This paper deals with the design of the necessary control loops so that good performance of the grid voltage can be attained in a very weak system which contains a wind park (WP) and large inductive loads. The WP consist of wind turbines with Doubly Fed Induction Generators (DFIGs). The DFIGs demand reactive power from the grid. These amounts of reactive power make the grid voltage very sensitive to load variations. The voltage performance can be improved by means of FACTS devices and better voltage controllers inside the DFIG. The system under study is shown in Figure 1. A medium voltage line is connected to the main grid at bus 1 with short circuit capability of 150 MVA. There is a steam power generation system (SPGS) at bus 2 with rated power of 50 MVA and a wind park at bus 3 connected to this line. This system can be a part of a local grid in an island to which wind parks are to be installed. The WP includes 11 wind turbines each with rated real power of 1.5 MW. At bus 4 there are inductive loads with rated power of 2 MVA and power factor 0.9 lagging. These loads also include three asynchronous motors each rated 300 kW. The nominal line voltage of the system is 25 kV. The variation of the reactive power demanded from the WP causes the load voltage at all buses to deviate from the rated values despite the presence of the SPGS in the system. At t = 50 sec there is an increase in the wind speed from 8 m/s to 14 m/s and at t = 75 sec the large induction motors start to operate. Figure 2 shows the rms value of the load voltage at bus 4. The real power produced from the WP is shown in Figure 3 and the reactive power from the WP is shown in Figure 4. The real power production of the SPGS is kept constant at 15 MW.
apr.ccsenet.
Stea
.org
AC
AC
am power generatio50 MVA
GridShort circuit
150 M
1
n unit
d capability
MVA
X = 4 ฮฉ
Fig
Applied
Medium voltaR = 1 ฮฉ
2
Medium voltage lR = 0.5 ฮฉ X=
Figure 1. S
Figure 2. RM
gure 3. Real po
Physics Resear
2
age line withX=4 ฮฉ 4
ine with2 ฮฉ
InAs
System under s
MS load voltage
ower from the
rch
Medium voltage R = 1 ฮฉ X=
nductive Loads withsynchronous motors
study
e in p.u.
WP (MW)
line with4 ฮฉ 3
Win
s
Vol. 10, No. 6;
AC
nd Park 10 MW
2018
apr.ccsenet.
As we can that the reavoltage drasynchronchange in those of thabout 5% wWhat happinsertion otakes very WP also oline.
.org
see, the grid vactive power frop of about
nous motors in the direction o
he wind turbinewhile without pens if we try toof the motors ay low values nooscillates and d
Figure
voltage starts tofrom the grid to2 % to the swhich the reac
of the reactive es. Finally, thethe presence oo add one more
and Figure 6 show and there isdoes not increa
Figu
Applied
e 4. Reactive p
o decrease fromowards the WPsystem. The gctive power expower stresse
e steady state vof the WP the se wind turbine hows the real ps a serious oscase. This actua
ure 5. Load vol
Physics Resear
3
power from the
m the rated valuP increases. Fr
grid voltage dxchange betwees the various evoltage drop tosteady state vo into the systempower coming cillation after thally means tha
ltage with 12 w
rch
e WP in MVars
ue when the wirom t = 50 sec
decreases signieen the grid andelectronic com
o the load bus aoltage drop wom? Figure 5 shfrom the WP.he start of the at this wind tu
wind turbines
s
ind speed incre up to t = 70 sificantly durind the WP chan
mponents of theafter the insertiould be 3%. ows the rms lo We can see thmotors. The r
urbine cannot b
Vol. 10, No. 6;
eases due to theec the WP caung the start onges direction.e system especion of the moto
oad voltage aftehat the load voreal power frombe connected t
2018
e fact uses a f the This
cially ors is
er the oltage m the o the
apr.ccsenet.
The improCompensaWT. The pthe SVC cinto the sywind turbiin weak sy2. Descrip2.1 MathemThe mathe
where: ฯd, ฯq areid, iq are thvd,vq are thEโq, Eโd arXd (=Xls+XXโd ( = Xd
.org
ovement of theator (SVC) at tpurpose of this ontroller and tstem to stabiliznes to be conn
ystems if good ption and modmatical model
ematical model
e the d,q comphe d,q componehe d,q compone the d,q compXmd) is the d ax- X2
md/Xfd) is
Figure 6. R
e grid voltage the bus 1 and paper is the pr
the grid-side coze the voltage.
nected to the linvoltage regula
deling of the sl of the general in p.u. of the
๐
ponents of the ents of the statnents of the staponents of the xis reactance, the d axis tran
Applied
Real power fro
performance cthe proper desresentation of nontroller of the As we shall sene which meanation is achievystem tor and the posynchronous g=
=๐โฒ = โ๐ธ
๐ = = ๐
stator magnetitor current respator voltage resstator internalXq (=Xls+Xmq
nsient reactance
Physics Resear
4
om the WP wit
can be achievesign of the curnew methods be WT so that thee, the improvens that higher pved.
wer line betwegenerator (SG)= ๐ ๐ + ๐
= ๐ ๐ โ ๐๐ธโฒ โ ๐ โ ๐
= โ๐ธ + ๐= ๐ โ ๐
โ ๐ ๐ โ ๐ic flux respectipectively, spectively, transient volta) is the q axis e,
rch
th 12 wind turb
ed by means orrent regulator based on optimhe right amouned performancpenetration of w
een the buses 1) of the SPGS + ๐ฃ + ๐ฃ
๐ ๐ + ๐ธ โ ๐ ๐
๐ โ ๐
ively,
age respectivereactance,
bines
of the installatof the grid-sid
mal control theonts of reactivee of the voltagwind energy ca
1-2 is (Padiyar, 20
ly,
Vol. 10, No. 6;
tion of a Staticde controller oory for the desi power are inje
ge permits two an be accompli
014):
(
(
(
(
2018
c Var of the gn of ected more ished
(a)
(b)
(1a)
(2a)
(3a)
(4a)
apr.ccsenet.org Applied Physics Research Vol. 10, No. 6; 2018
5
Xโq ( = Xq - X2mq/X1q) is the q axis transient reactance,
ฯ is the rotor electrical angular speed, ๐ is the electrical speed of the magnetic flux, ฮด is the power angle, H is the inertia constant, Tm is the mechanical torque, TD is the damping torque (being neglected from now on). Finally, Tโdo, Tโqo are time constants on field and damper winding respectively (we consider the machine to have one damping winding on q axis) and Rs is the stator resistance. By neglecting the equations regarding the stator magnetic fluxes (equations (a) and (b) above) and replacing the relevant values from the Appendix we have:
= โ0.22 ๐ธโฒ โ 0.31๐ + 0.22๐ธ (1b)
= โ1.5 ๐ธ + 1.7 ๐ (2b)
= ๐ โ 314 (3b)
It also is: ๐ = โ๐ฅ ๐ + ๐ธ , ๐ = โ๐ฅ ๐ โ ๐ธโฒ By replacing to (4a) we finally reach in:
= 180.46 ๐ โ 180.46 ๐ธ ๐ โ 180.46 ๐ธ ๐ + 37.89๐ ๐ (4b)
As we have already seen, the SPGS is connected to the bus 2 and there is a small line up to the main bus 1. The Figure 7 depicts the vectors of the voltages at the buses. The voltage v2 at the bus 2 will be a little ahead of the voltage v1 at the bus 1 (approximately 3 degrees). We consider the main axes D, Q and the axes d, q internally in the synchronous generator to which the various quantities of the SG have been analyzed in the equations (1b)-(4b). We arbitrarily consider that the voltage v2 is lying on the D axis. The stator current i of the SG with its components id, iq onto the axes d, q is also the current I of the line between the buses 2 and 1 with the components ID, IQ onto the axes D, Q respectively. It is (Pai et al., 2014): ๐ = ๐ผ ๐ ๐๐๐ฟ โ ๐ผ ๐๐๐ ๐ฟ, ๐ = ๐ผ ๐๐๐ ๐ฟ + ๐ผ ๐ ๐๐๐ฟ (5a)
D
Q
V1
V2
q
d
ฮด
iq
id
ID
IQ
Figure 7. Vector analysis of the system
Applying the D-Q analysis on the line between the buses 2 and 1 we have firstly on the D axis: ๐ = ๐ + ๐ ๐ผ + ๐ฟ + ๐ ๐ผ โ ๐ โ ๐ = ๐ ๐ผ + ๐ฟ + ๐ ๐ผ
Due to the small value of the angle between v1 and v2 it is approximately: ๐ โ ๐ โ 0. So from the previous equation we conclude to:
apr.ccsenet.org Applied Physics Research Vol. 10, No. 6; 2018
6
๐ฟ = โ๐ ๐ผ โ ๐ ๐ผ
and by replacing the values from the Appendix we have:
= โ0.35๐ผ โ ๐ผ (5)
From the line analysis on the axis Q we have: ๐ = ๐ + ๐ ๐ผ + ๐ฟ ๐๐ผ๐๐ก โ ๐ ๐ผ โ ๐ โ ๐ = ๐ ๐ผ + ๐ฟ ๐๐ผ๐๐ก โ ๐ ๐ผ
By taking into account that v2Q = 0 and by replacing the values we have:
= โ0.35๐ผ + ๐ผ โ 3.8๐ (6)
The equations (1b)-(4b) can be rewritten if we replace the quantities id, iq by ID, IQ using the equation (5a). Then we can reach in:
= โ0.22๐ธ โ 0.31๐ผ ๐ ๐๐๐ฟ + 0.31๐ผ ๐๐๐ ๐ฟ + 0.22๐ธ (1)
= โ1.5๐ธ + 1.7๐ผ ๐๐๐ ๐ฟ + 1.7๐ผ ๐ ๐๐๐ฟ (2)
= ๐ โ 314 (3) ๐๐๐๐ก = 180.46๐ โ 180.46๐ธ ๐ผ ๐๐๐ ๐ฟ โ 180.46๐ธ ๐ผ ๐ ๐๐๐ฟ โ 180.46๐ธ ๐ผ ๐ ๐๐๐ฟ + 180.46๐ธ ๐ผ ๐๐๐ ๐ฟ โ
โ18.95๐ผ ๐ ๐๐2๐ฟ + 18.95๐ผ ๐ ๐๐2๐ฟ โ 37.89๐ผ ๐ผ ๐ ๐๐ ๐ฟ + 37.89๐ผ ๐ผ ๐๐๐ ๐ฟ (4) The nonlinear system that consists of the voltage v1 at the bus 1 and the internal quantities of the SG is actually described by the equations (1)-(6). By linearizing the above equations around the operating point given at the Appendix we finally conclude to the linear system given by the following equations:
โ = โ0.22โ๐ธ โ 0.06โ๐ฟ โ 0.155โ๐ผ + 0.27โ๐ผ + 0.22โ๐ธ (7)
โ = โ1.5โ๐ธ โ 0.385โ๐ฟ + 1.47โ๐ผ + 0.85โ๐ผ (8)
โ = โ๐ โ 314 (9)
โ = โ37.45โ๐ธ โ 40.89โ๐ธ + 20.7โ๐ฟ + 137.2โ๐ผ โ 25โ๐ผ + 180.46โ๐ (10)
โ = โ0.35โ๐ผ โ โ๐ผ (11)
โ = โ0.35โ๐ผ + โ๐ผ โ 3.8โ๐ (12)
The equations (7)-(12) form the linearized model of the SG and the line between the buses 1 and 2.
2.2 Model of the grid โ side converter of the WT A schematic configuration of the grid side converter, is shown in Figure 8 in which the grid phase voltages are denoted as ea, eb, ec and the converter phase voltages as va, vb, vc are respectively. The d,q components (id, iq) of the line currents ia, ib,ic concerning the d,q components (vd, vq) of the converter voltages va, vb, vc can be given by the following equations (Vittal & Ayyanar, 2013):
apr.ccsenet.
The modelchoice of treplacing t
As there ifrequency the paramesystem and
3. Control
3.1 SVC CIn order foinstallationwhich fastThere is aregulates tinto the sysignals in There are voltage at
that is, the variation othe bus 1.
.org
Fi๐ฟ ๐๐ling of the convthe WT d,q, axthe p.u. values๐๐๐๐ก =is the SPGS inis assumed to
eter ฯs has beed not an input
โl SYSTEM DE
Controller Desior the load voltn is proposed at power factor a main voltagethe right suscepystem. A blockthe decision ono such signathe bus 1 to ch
variation of thof the suscepta
igure 8. Grid s๐๐๐๐ก = โ๐ ๐ +verter of the Wxes is arbitrarys from the App= โ0.01๐ + ๐n our system be constant an
en set equal to and so the line= โ0.01โ๐ESIGN
ign age to be kept at the bus 1 of timprovement e controller wptance being cok diagram of Sof the right susals in our studyhange. If only
he magnitude oance of the SVC
Applied
ide converter c+ ๐ ๐ฟ ๐ + ๐WT and its conty. The grid volpendix we reac๐ + 6.67๐ โwith installed
nd that it does n1. Besides, theearized model + โ๐ โ 6.67constant underthe system. Thcan be achiev
which accordinonnected to theSVC voltage csceptance (B) y. The connecthe fundamentโ๐
of the voltage aC (ฮBsvc). X
Physics Resear
7
configuration (โ ๐ฃ , ๐ฟ ๐๐๐๐กtrol design hasltage e has bee
ch in the followโ 6.67๐ฃ , ๐๐๐๐กd capacity mucnot depend on te parameter ed for the control7โ๐ฃ , โ =r reactive powhe SVC is actued. Figure 9 sh
ng to the deviae bus and so th
control is depicof the SVC co
cted susceptanctal componentโ ๐โ๐ต ,
at the bus whereis the equivale
rch
(Vittal & Ayya= โ๐ ๐ โ ๐ nothing to do en chosen to liwing equations
๐ก = โ0.01๐ โch larger that the real power is considered tl design of the= โ0.01โ๐ โ
wer from the WPually a device whows a typicalation of the b
he right amouncted in Figure ontrol, like thece at the bus 1t of Vsvc is con
e the SVC is inent Thevenin re
anar, 2013) ๐ ๐ฟ ๐ โ ๐ฃ with the rest o
ie on the d axis: โ ๐ โ 6.67๐ฃ
the rated outpcoming from tto be a disturba
e grid side convโ๐ โ 6.67โ๐ฃP and the load
with variable sul SVC (TSC-Tbus voltage front of reactive po
10. There cane signals Vs sh1 will force thnsidered, we c
nstalled (i.e. buesistance of th
Vol. 10, No. 6;
of the system, sis and so eq = 0
put of the WPthe WP. As a reance for the coverter is given ๐ฃ (
variations, an usceptance thr
TCR) configuraom the rated vower being insn be more auxihown in Figure magnitude o
can assume tha
us 1) depends ohe grid as seen
2018
so the 0. By
P, the esult,
ontrol by:
(13)
SVC rough ation. value erted iliary e 10.
of the at:
on the from
apr.ccsenet.
The equatiSPGS is co๐In the modwe set for equivalentFor the outthe q-axis voltage. Fr
By lineariz
So, in the added. Thewe can reaspace form
The designvariables aindex has b
.org
Figure 9.
Fi
ions (7)-(12) foonsidered to pr= |๐ |๐ ๐๐๐ฟdel given by (7r now as input t susceptance vtput to be seleccomponent ofrom the equati
zing we get:
linearized mode equations (7)ach in a new 2n
m ๐ฅ = ๐ด๐ฅ + ๐ต ๐ด = โ
n of the SVC and the outputbeen set:
. Typical confi
igure 10. SVC
orm the lineariroduce steady โ ฮ๐ = ๐ ๐๐
7)-(12) the inputhe variation
value in the encted, we need af the stator curion (5a) it is:
del of the equa)-(12) and the nd order system๐ข, ๐ฆ = ๐ถ๐ฅ + ๐ทโ0.5246 00 4.85
controller is t to be minimi
Applied
iguration of SV
voltage contro
zed model conreal power int๐๐ฟ |ฮ๐ | โ ฮut to the systemof the voltaged. a signal that wirrent iq of the s
๐ = ๐ผ ฮ๐ = 0.866
ations (7)-(12) equation (14)
m which is equi๐ท๐ข: 057 , ๐ต = 1.โ0based on the
ized within a s
Physics Resear
8
VC (TSC-TCR
ol with series r
ncerning the SPto the system aฮ๐ = ๐ ๐๐๐ฟm will actuallye ฮV1Q and the
ill follow the vsynchronous g
๐๐๐ ๐ฟ + ๐ผ ๐ ๐๐6ฮ๐ผ + 0.5ฮ๐ผ
the equation (form a system
ivalent to the p
.219.0225 , ๐ถ =optimal contro
specific smallt
rch
R) configuratio
regulator (Padi
PGS and the linand so the angl๐ ๐ ฮ๐ต โy be the change calculated in
voltage variatiogenerator has th
๐ฟ.
โ 0.21ฮ๐ฟ (14) regarding
m of 6th order. Bprevious one. T
โ1.219 3.025ol theory. Wetime period (tf)
on (Padiyar, 20
iyar, 2014)
ne between thele ฮด1 is almost โ ฮ๐ = 0.03
ge in the suscepnput will be tr
on. As we can she same perfo
the output of By means of MThe new system
5 , ๐ท = 0, ๐ฅ =e want the varf). So, the follo
Vol. 10, No. 6;
014)
e buses 1 and 2constant, that 35ฮ๐ต ๐. ๐ข.
ptance ฮ๐ต ransformed int
see from Figurormance as the
(the system mu
MATLAB softwm is given in st
= ๐ฅ๐ฅ (
riations of the owing perform
2018
. The is:
but o the
re 11, load
(14) ust be ware, tate โ
(15)
state mance
apr.ccsenet.
which is o
with F = 0
Riccati equ
where:
The optim
From the pdifferentia
We set the(such as 4feedback aSVC contr
By transfocontroller
We can seFigure 12.
.org
f the form: ๐ฝ =0, R = 1 and
uation needs to๐mal input u is th
previous definal equations for
๐๐e time interval 4 sec) we can sarray K is equroller should b
orming this inp(as it is shown
ee the step resp
Figure 11. Q-
๐ฝ= 12 ๐ฅ ๐ก ๐น ๐ก
๐ = 1 00 1 . A
o be solved: ๐ = โ๐ ๐ก ๐ด๐ = ๐๐
hen:
nitions and the r the various e๐ = 1.049๐= โ4.3324๐ = โ9.714tf equal to 1 se
see that the paal to: K = [0.6
be:
put to the realn in Figure 10)
ponse of the lin
Applied
-axis compone
๐ฝ = 12 ๐ฅ๐ฅ ๐ก + 12
According to t
๐ก โ ๐ด ๐ก ๐ ๐ก๐๐ , with f
u(t) = -Kx(tmatrices A, Blements of the๐ + 1.486๐4๐ + 1.4586๐ + 1.486๐ec. Solving thearameters p11, 67 -4275]. Acc
๐ข = โ0.6l input ฮฮSVC ) is:
๐บ ๐ =nearized mode
Physics Resear
9
ent of the curre
๐ก + ๐ฅ ๐ก +๐ฅ ๐ก ๐ ๐ก ๐ฅ
the optimal co
๐ก + ๐ ๐ก ๐ต ๐กfinal condition
t), with K=R-1
B, C, D given ie matrix P: โ 0.0548๐6๐ ๐ โ 0.02โ 0.0548๐e previous set p12, p13 are: p1
cording to the
67๐ฅ + 4275and by means
. ..el with the opti
rch
ent from the SG
๐ข ๐ก ๐๐ก
๐ฅ ๐ก + ๐ข ๐ก ๐ ontrol theory,
๐ ๐ต ๐ก ๐ ๐กn P (t = tf) = F =
B(t)TP(t). in equation (15
๐ + 0.0005274๐ + 0.00๐ + 0.000of equations f
11 = 0.55, p12 =equation (16)
2๐ฅ s of MATLAB
..
imal control in
G (p.u.)
๐ ๐ก ๐ข ๐ก ๐๐ก
the following
๐ก โ ๐ ๐ก
= 0.
5), we finally
5๐ โ 1, 005๐ ๐ 05๐ โ 1 for time interva= 0, p22 = 190the optimal co
B software, th
nput given by t
Vol. 10, No. 6;
matrix differe
(have the follo
al longer than 0000. So, the ontrol input fo
he equivalent s
(
the equation (1
2018
ential
(16) wing
1 sec state
or the
series
(17)
17) in
apr.ccsenet.
We can se3.2 Grid siThe linear
Due to theconclude t
and we con
be equal to
We set the
๐ปIn order to
By substitu๐ป ๐ฅWe solve t
Now letโs to be indep
.org
Fig
e that the systeide converter cized model is g
๐ฅ =e nature of the to a controller
nsider the opti
o: ๐ฝโ = ๐๐ฅ +e Hamiltonian
๐ฅ ๐ก , ๐ข ๐ก , ๐ฝโ๐ป ๐ฅ ๐ก , ๐ข ๐ก , ๐ฝo have optimiza
ution to the pr๐ก , ๐ข ๐ก , ๐ฝโ =๐ป ๐ฅ ๐กthe HJB equati
๐ป ๐ฅ ๐ก , ๐ข ๐กrecall that x1 =pendent on the
gure 12. Step re
em output getscontrol designgiven by the e๐ฅ = ๐ฅ๐ฅ =โ0.01๐ฅ + ๐ฅabove equatiovery easily. W๐ฝ = ๐imal performan
+ ๐๐ฅ , then i
equation as:
= ๐ ๐ฅ ๐ก , ๐ข ๐ก๐ฝ๐ฅโ = ๐ฅ + ๐ฅation there mu๐๐ป๐๐ข = 0 โevious we hav๐ฅ + ๐ฅ + 11.1๐ก , ๐ข ๐ก , ๐ฝ๐ฅโ =ion: ๐ก , ๐ฝ๐ฅโ + ๐ฝ๐กโ = 0
= ฮid, x2 = ฮiq. e changes in th
Applied
esponse of the
s the desired va quations (13). = โ๐โ , ๐ข =โ 6.67๐ข = ๐
ons we can woWe set the perfo๐ ๐ฅ ๐ก , ๐ข ๐ก ๐
nce index to
it is: ๐ฝโ = โโ
+ ๐ฝโ ๐๐ = ๐ฅ+ ๐ข + ๐ข โust be: โ ๐ขโ = 3.335๐ve: 12๐ ๐ฅ + 11.121 โ 11.12๐0 โ 1 โ 11.12We want the c
he q axis (whi
Physics Resear
10
linearized mo
alue in 0.1 sec
In state space ๐ข๐ข = โ๐ฃโ๐ฃ , ๐ , ๐ฅ = โrk with the Ha
ormance index๐๐ก = ๐ฅ + ๐ฅโโ = ๐๐ฅ๐๐ฅ , ๐ฝโ
๐ฅ + ๐ฅ + ๐ข +0.01๐๐ฅ โ 6.6๐๐ฅ , ๐๐ป๐๐ข = 0 โ2๐ ๐ฅ โ 0.01๐๐ฅโ 0.01๐ ๐ฅ +2๐ โ 0.01๐ ๐ฅchanges in the ch is responsib
rch
odel under opti
c.
form if we set
then we can wโ0.01๐ฅ โ ๐ฅ โamilton - Jacobx as: ๐ฅ + ๐ข + ๐ขโ = โ = 0 .
๐ข + ๐๐ฅ ๐๐ฅ67๐๐ฅ ๐ข โ 0.0โ ๐ขโ = 3.335๐ฅ โ 0.01๐๐ฅ โ+ 1 โ 11.12๐๐ฅ + 1 โ 11.d axis (which ble for the rea
imal control.
t
write: โ 6.67๐ข = ๐bi โ Bellman (
๐๐ก
โ0.01๐ฅ + ๐ฅโ๐ฅ โ 0.01๐ฅ001๐๐ฅ โ 6.65๐๐ฅ . โ 22.24๐ ๐ฅ โ๐ โ 0.01๐ ๐ฅ.12๐ โ 0.01๐is responsible ctive power).
Vol. 10, No. 6;
(HJB) equation
โ 6.67๐ขโ 6.67๐ข 67๐๐ฅ ๐ข
22.24๐ ๐ฅ โ
๐ ๐ฅ = 0 (for the real poIn other words
2018
n and
(18) ower) s, the
apr.ccsenet.
input u mu(18) to be
We keep th
A very sim
4. NonlineThe systemvoltage cois given byt = 50 sec arms value reactive po
There is analmost 1 %start of the
.org
ust make two itrue for every
he negative sig
mple equivalen
ear System Sim as shown in Fontroller is givey the equation and at t = 75 seof the load vo
ower from the
Figu
n important im% voltage drop e large motors
independent stat, then the opt1
gn and hence t
nt series contro
mulation ResFigure 1 has been by the equa(19). As previec the large indoltage at bus 4WP is shown
Figure 13
ure 14. Reactiv
mprovement in in the system
. The voltage d
Applied
ate variables xtimal solution sโ 11.12๐ โthe control law๐ข = โoller for iq is: ๐บ๐ults een simulated nation (17) and tously, there is duction load w4. The real poin Figure 14.
3. Load voltag
ve power towa
the load voltagand the reactivdrop is about 4
Physics Resear
11
x1 and x2. Havishould be as: 0.01๐ = 0 โ
w is: โ๐ฅ , ๐ข = โ๐ฅ๐ ๐ = โ .
now with the athe iq controllea step increase
with the asynchrower produced
ge with the pro
ards the WP wi
ge. The reactivve power towa4 % under the
rch
ing taken this i
๐ โ 0.3. .
addition of an Ser the grid sidee in the wind sronous motors
d from the WP
oposed control
ith the propose
ve power demaards the WP dodemands for r
in mind and in
SVC placed at te converter in espeed from 8 ms is activated. FP is the same a
llers
ed controllers
anded by the Woes not change reactive power
Vol. 10, No. 6;
n order the equ
(
the bus 1. The each WT of them/sec to 14 m/sFigure 13 showas in Figure 3.
WP causes nowdirection unde
r both from the
2018
ation
(19)
SVC e WP sec at
ws the . The
only er the e WP
apr.ccsenet.
and the indvoltage droThe d comconverter cthe other amachine. The proporated powevoltage in
.org
ductive load duop for the start
mponent of the current iq is shand the id is m
osed control strer as 19.5 MWFigure 18.
uring the transting of the motgrid-side convown in Figure
mostly respons
rategy permitsW. The real pow
Figure 15.
Figure 16.
Applied
ient period andtors has been dverter current i16. We can ac
sible for the r
s the addition ower coming ou
D component
Q component
Physics Resear
12
d almost zero idecreased fromid is shown in Fctually see thatreal power out
of two more mut from the WP
of the grid-sid
of the grid-sid
rch
in steady state.m 15 seconds toFigure 15 and t the two comptput while the
machines to be P in this case is
de converter cu
de converter cu
. Besides, the to less than 10 the q compon
ponents are indiq for the rea
added to the Ws shown in Figu
urrent
urrent
Vol. 10, No. 6;
time duration oseconds.
nent of the griddependent one active power o
WP which nowure 17 and the
2018
of the
d-side from
of the
w has load
apr.ccsenet.
5. ConclusThe insertiabsorbed rdifficult. Tperformancan be appproposed increased. ReferenceCustem, TDesineni, NPai, M. A.Krause, P.Jahangir,
RenewAbad, G.,
.org
sions ion of wind tureactive poweThe installationnce in these caplied in any wcontrollers int
es T. V., & Vourn
N. (2003). Op, & Gupta, D. (2010). AnalyH., & Hemanwable Energy.Lopez. J., & R
Figure 17.
Figur
urbines with inder from the gen of SVC devses. The propo
weak or local gto the system
nas, C. D. (199timal Control . P. (2016). Smysis of Electricnshu, R. P. (2. Springer. Rodriguez, M.
Applied
Real power fr
re 18. Load vo
duction generaenerators. The vices along witosed methodol
grid. The load and the pene
8). Voltage StaSystems. CRC
mall Signal Anacal Machinery.2014). Robust
A. (2011). Do
Physics Resear
13
om the WP wi
oltage with 13
ators into weaklow voltage m
th proper contlogies make thvoltage was im
etration of the
ability of the EC Press. alysis of Integr IEEE Press.
t Control for
oubly Fed Indu
rch
ith 13 wind tur
wind turbines
k systems causmay make thetrollers inside the design of thmproved signie rated real po
Electric Power
rated Power Sy
Grid Voltage
uction Machine
rbines
ses extra voltae start of largethe WT can im
he necessary cificantly after tower from the
r System. K. Ac
Systems. Alpha
e Stability: Hi
e. IEEE Press.
Vol. 10, No. 6;
age drops due te induction mmprove the voontrollers easythe insertion o
e wind energy
cademic.
Science.
igh Penetratio
2018
to the motors oltage y and of the y was
on of
apr.ccsenet.org Applied Physics Research Vol. 10, No. 6; 2018
14
Robert, F. S. (1993). Optimal Control and Estimation. Dover publications. Vijay, V., & Raja, A. (2012). Grid Integration and Dynamic Impact of Wind Energy. Springer. Kesraoui, M., & Chaib, A. (2016). Grid voltage local regulation by a doubly fed induction generator-based wind
turbine. Wind Engineering, 41(1), 13-25. Rui, S., & Rui, X. (2015). The design and analysis of wind turbine based on differential speed regulation. Wind
Engineering, 230(2), 221-229. Yu, L., & Zhenlan, D. (2012). Improvement of the low-voltage ride-through capability of doubly fed Induction
Generator Wind Turbines. Wind Engineering, 36(5), 535-551. Appendix Table 1. WT parameters (Abad et al., 2011)
Nominal active power PN 1.5 MW Nominal electrical torque TelN or Tg 9555 Nm Stator voltage VSN 690 V Nominal generator speed ฮทgo 1800 rpm Speed range of generator 900-1850 rpm Pole pairs 4 Blades diameter d 60 m Nominal wind speed VwN 12 m/sec Maximum power coefficient Cp 0.44 Air density 1.125 kg/m3 Nominal turbine speed ฮทto 22.5 rpm Speed range of turbine speed 9-23 rpm TSR optimum 5.43 Grid side components (mฮฉ) Rg = 0.33 Xg = 31.4
Table 2. Medium voltage lines
Rated voltage VN 25 kV Inductive reactance Xo 0.4 ฮฉ/km Resistance Ro 0.1 ฮฉ/km Length between buses 1-2 5 km Length between buses 1-4 and 3-4 10 km
Table 3. Synchronous generator parameters (Pai et al., 2016)
Rated voltage VN 11 kV D axis open circuit time constant Tdoโ 4.5 sec Rated power SN 50 MVA Q axis open circuit time constant Tqoโ 0.67 sec Transient reactance on d axis Xdโ 0.25 p.u. Inertia constant H 0.87 sec Reactance on d axis Xd 1.65 p.u. Stator resistance Rs 0.0045 p.u. Transient reactance on q axis Xqโ 0.46 p.u. Power angle operating point ฮดo =300 Reactance on q axis Xq 1.59 p.u. Bus 1 voltage operating point V1qo=-0.052 p.u. Subt. int.voltage on q axis Eqโ operating point Eโqo=0.68 p.u. Line current on D axis operating point IDo=0.28 p.u. Subt. int.voltage on d axis Edโ operating point Eโdo=0.22 p.u. Line current on Q axis operating point IQo=-0.1 p.u.
Copyrights Copyright for this article is retained by the author(s), with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/).