Research Article Differential Evolution Based IDWNN ...downloads.hindawi.com/journals/tswj/2015/746017.pdfResearch Article Differential Evolution Based IDWNN Controller for Fault Ride-Through

Research ArticleDifferential Evolution Based IDWNN Controller forFault Ride-Through of Grid-Connected Doubly Fed InductionWind Generators

N Manonmani1 V Subbiah2 and L Sivakumar3

1Department of EEE Sri Krishna College of Engineering and Technology Coimbatore 641008 India2Department of EEE PSG College of Technology Coimbatore 641004 India3Sri Krishna College of Engineering and Technology Coimbatore 641008 India

Correspondence should be addressed to N Manonmani manonmaniskcetacin

Received 30 May 2015 Revised 5 July 2015 Accepted 8 July 2015

Academic Editor Gnana Sheela

Copyright copy 2015 N Manonmani et alThis is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

The key objective of wind turbine development is to ensure that output power is continuously increased It is authenticated that windturbines (WTs) supply the necessary reactive power to the grid at the time of fault and after fault to aid the flowing grid voltage Atthis juncture this paper introduces a novel heuristic based controller module employing differential evolution and neural networkarchitecture to improve the low-voltage ride-through rate of grid-connected wind turbines which are connected along with doublyfed induction generators (DFIGs) The traditional crowbar-based systems were basically applied to secure the rotor-side converterduring the occurrence of grid faults This traditional controller is found not to satisfy the desired requirement since DFIG duringthe connection of crowbar acts like a squirrel cage module and absorbs the reactive power from the grid This limitation is takencare of in this paper by introducing heuristic controllers that remove the usage of crowbar and ensure that wind turbines supplynecessary reactive power to the grid during faults The controller is designed in this paper to enhance the DFIG converter duringthe grid fault and this controller takes care of the ride-through fault without employing any other hardware modules The paperintroduces a double wavelet neural network controller which is appropriately tuned employing differential evolution To validatethe proposed controller module a case study of wind farm with 15MW wind turbines connected to a 25 kV distribution systemexporting power to a 120 kV grid through a 30 km 25 kV feeder is carried out by simulation

1 Introduction

For the past few years doubly fed induction generators(DFIGs) occupied the worldrsquos largest share of wind turbinesas a variant to traditional variable speed generators Thedesigned system should be in a position to operate over awide range of wind speeds for achieving optimal efficiencyin order to follow the optimal tip-speed value Henceforththe generatorrsquos rotor is to be designed to operate in a variablerotational speed DFIGs are therefore designed to operateunder both super- and subsynchronous modes with a speedrange of the rotor being in accordance with that of thesynchronous speed The stator circuit module is connecteddirectly to that of the grid on the other hand the windingsof the rotor are connected through slip rings to a three-phase

converter In case of a variable speed system when the speedrange requirements are low the DFIG provides satisfactoryperformance and this is found to be enough for the speedrange employed with the wind power In DFIG modulean AC-to-AC converter is incorporated in the inductiongenerator rotor circuit and the converters are to be rated tohandle a fraction of the total power The total power is therotor power which forms around 32 of nominal generatorpower As a result the occurrences of losses in the convertermodule can be reduced in comparisonwith that of the systemwhere the converter is to handle the complete power andthe entire cost is minimal because of the partial rating ofthe power electronic circuits This section details the earlierstudies and simulations carried out in the applicability anddesign of DFIGs in wind turbines

Hindawi Publishing Corporatione Scientific World JournalVolume 2015 Article ID 746017 15 pageshttpdxdoiorg1011552015746017

2 The Scientific World Journal

Vieira et al [1] carried out a work on genetic algorithmbased optimal controllers to the rotor-side converter ofdoubly fed induction generators (DFIGs) the DFIG con-verter control action performed by proportional and inte-gral controllers The presented approach improves transientstability margin of the power system and also has betterglobal system dynamic behavior during and after the faultperiod In [2]Nagaria et al analyzed various control strategiesfor frequency support in DFIG-based WECS based on thecomparative analysis the following point can observe mini-mum frequency deviation and maximum release of energyobtained by tuning the controller parameter using PSO andimproved performance is achieved by inertial response ofthe system with the PSO optimized proportional controllerBhatt et al [3] presented doubly fed induction generatordynamic participation in automatic generation control fre-quency response of the system is improved by frequencycontrol support function and craziness-based particle swarmoptimization (CRPSO) utilized to obtain the optimal DFIGparameter transient response of area frequency and tie-linepower deviation

A novel robust controller [4] is suggested in a stationaryreference frame for grid-integrated doubly fed inductiongenerators based wind turbines The voltage and flux equa-tions in 120572120573 coordinates used DFIG dynamic model designthe disturbances and uncertainties intrinsic to the systemare accounted for as perturbation terms are included tothe nominal model ANN [5] based adaptive PI controllerson DFIG driven based wind turbine in order to improvetransient performance of DFIG in all wind speed conditionsANN is used to predict the optimal values of parameter andadaptive PI controllers according to different wind speedsdynamically change PI gain values A genetic algorithm [6]based PID controller coefficients adjustment on wind energyconversion system with doubly fed induction generator Liuet al [7] pointed out the DFIG speed controller based onthe nonlinear intelligent model predictive control algorithmSVM is adapted to establish the speed of DFIG and PSO isutilized to solve the rolling optimization the forecasted erroris used to form a closed-loop control Various computationalintelligence based controls of wind turbine generators such asPSOMVO adaptive critic design based adaptive control andshunt FACTS devices are analyzed in [8]

Multiobjective optimization [9] control methodology ofa doubly fed induction generator (DFIG) considering thedistorted grid voltage conditions was dealt with The fourvarious control targets are suggested to restrain the har-monic components in statorrotor current or the ripples inthe stator output activereactive power and electromagnetictorque Vinothkumar and Selvan [10] carried out a workon a new enhancement of fault ride-through capability andsubsequent betterment of rotor speed stability of doubly fedinduction generator based wind farms The wind turbinemechanical energy input is stored during the grid fault andat the moment of fault clearance it is utilized hence thecharging DC-link capacitor by grid-side converter is relievedParticle swarm optimization [11] based optimal VAR controlin grid-connectedDFIG-basedwind farmhas been presentedto optimize reactive power from reactive power capability

(RPC) of DFIGs in order to obtain system active power lossminimization

A hybrid particle swarm optimization [12] (HPSO) algo-rithm for the DFIG optimum design is presented in thiswork and the searching performance is improved by adaptingthe diversity-guided adaptive mutation and fitness-guidedindividual fuzzy inertia weight An optimal grid-integrated-to-the-rotor type doubly fed induction generator design forwind turbine systems is developed in [13] Kriging modelbased on Latin hypercube sampling and genetic algorithm areadapted to maintain the efficiency and maximize the torqueper weight Amelian et al [14] carried out a work on a smallsignal stability improvement based on eigenvalue analysis ofa wind turbine-based doubly fed induction generator in amicro grid environment The stability margin and dampingratio of critical modes are enhanced by using particle swarmoptimization algorithm

An optimal VAR expansion based on capability curve ofDFIG wind farm has been proposed in [15] The proposedapproach reduces the sum of the annual investment costof the new VAR devices and the annual expected energyloss cost Design of adaptive controller [16] using optimalfuzzy logic of doubly fed induction generator (DFIG) windturbine for standalone power system frequency control hasbeen proposed in this workThe particle swarm optimizationis employed to adapt and optimize the fuzzy logic designmembership function and the control rules

Tang et al [17] carried out a work on an optimalcontrol of doubly fed induction generators- (DFIGs-) basedwind generation adapting trajectory and frequency domaininformation based sensitivity analysis and particle swarmoptimization The critical parameters the unified dominatecontrol parameters (UDCP) are identified using sensitivityanalysis and PSO is utilized to obtain the control goal bysearching the optimal values A fuzzy controller is pro-posed in [18] based on fuzzy adaptive theta particle swarmoptimization (FAΘPSO) applied to a doubly fed inductiongenerator (DFIG) to run at medium voltage FAΘPSO algo-rithm is used to obtain optimum values of membershipfunctions A load frequency control (LFC) of micro gridsconnected with main grids in a regulated and deregulatedenvironment the PSO (particle swarm optimization) tunedPI controller gains have improved performance compared tothe conventional PI controller parameters proposed in [19]An adaptive neurofuzzy inference system [20] controller ispointed out for doubly fed induction generator based windturbine The presented approach controls the current andvoltage ripple within 001 pu as well as reducing the powerloss

Kong et al [21] proposed data-driven based modeling ofdoubly fed induction generator wind turbine system adaptingneural networksTheDFIGmodel is developed based onneu-ral networks and neurofuzzy networks by using large amountof input-output online measurement data from the selectedmonths An RSC controller [22] is proposed for DFIG-basedwind power generation system In this case the stator-flux-oriented (SFO) vector control based design is adapted andRSC controller for DFIG with double closed-loop system ismodeled using rotor current control loop and power control

The Scientific World Journal 3

Turbine

WindDrive train

Rotor

Pitch angle

Stator

Inductiongenerator

AC DC AC

ACDCAC converter

Control

Vr Vgc

Crotor Cgrid

L

Three-phasegrid

Figure 1 Wind turbine and the doubly fed induction generator system

loop alongwith proportional integral (PI) controllerWu et al[23] suggested combination of particle swarm optimization(PSO) and orthogonal method for wind turbines parametertuning of doubly fed induction generator systems A virtualresistance control approach for high voltage ride-through[24] of doubly fed induction wind generators using particleswarm optimization is presented in this work Zhang etal [25] pointed out design of additional impedance-basedSSR damping controller for power system with DFIG windturbine the additional SSR damping controller effectivelysuppresses the SSR

It is to be noted that in all the above discussed earlierstudies on design and control of DFIG the control strategiesare complex in nature yielding unstable nature of the systemand voltage rating of rotor is high enough in comparisonwiththat of the megawatt rating of the wind turbine generatorswhich possess stator to rotor winding turns ratio between25 and 3 In this work a novel control strategy employingDE based IDWNN controller is proposed which makes theDFIG-based system of high stable nature and the designedsystem simulation is carried out choosing appropriate turnsratio This paper also attempts to handle the fault ride-through capability of wind turbines and to make them morestringent The remaining part of the paper is organized asfollows Section 2 presents themodeling ofDFIG and the pro-posed controller design employing DE-IDWNN controlleris discussed in Section 3 Section 4 presents the applicabilityof the proposed DE-IDWNN for the fault ride-through ofDFIG module The simulation results and the validationof the proposed controller are performed in Section 5 Theconclusions with the findings of the proposed work aresummarized in Section 6

2 Modeling of DFIG Wind Turbine System

The schematic diagram of the wind turbine and DFIG is asshown in Figure 1

The AC-DC-AC converter system of DFIG is found tobe divided into two components the rotor-side converter(119862ROTOR) and the grid-side converter (119862GRID) Both 119862ROTORand 119862GRID are voltage-sourced converters which employforced commutated power electronic devices (IGBTs) forsynthesizing an AC voltage from a DC voltage source Onthe DC side a capacitor is connected which acts as a DCvoltage source For connecting 119862GRID to the grid a couplinginductor 119871 has been used Slip rings and brushes are usedto connect the three-phase rotor winding to 119862ROTOR andthe three-phase stator winding is directly connected to thegrid The power acquired by the wind turbine is convertedinto electrical power by the induction generator and this istransmitted to the grid by the stator and the rotor windingsThe control system module is employed to generate thepitch angle command and the voltage command signals 119881

119877

and 119881119866for 119862ROTOR and 119862GRID respectively These signals

tend to control the power of the wind turbine the DCbus voltage and the voltage or reactive power at the gridterminals

The equations employed for computing mechanicalpower and the stator electric power output are as follows

119875119898= 119879119898120596119903

119875119904= 119879119890120596119904

(1)

where 119875119898is mechanical power captured by the wind turbine

and transmitted to the rotor 119875119904is stator electrical power

output 119879119898

is mechanical torque applied to rotor 119879119890is

electromagnetic torque applied to the rotor by the generatorand 120596

119903and 120596

119904are the rotational speed of rotor and of the

magnetic flux in the air-gap of the generator respectively Incase of lossless generator the mechanical equation is given by

119869119889120596119903119889119905

= 119879119898minus 119879119890 (2)

Considering steady state at constant speed in case of losslessgenerator 119879

119898= 119879119890and 119875

119898= 119875119904+ 119875119903 where 119875

119903is the rotor


electrical power output and 119869 is the rotor and wind turbineinertia coefficient The rotor electrical power output is givenby

119875119903= 119875119898minus 119875119904= 119879119898120596119903minus 119879119890120596119904= minus119879119898120596119904minus120596119903120596119904

120596119904

= minus119904119879119898120596119904= minus119904119875

119904

(3)

where ldquo119904rdquo is defined as the slip of the generator 119904 = (120596119904minus

120596119903)120596119904 The turbine and tracking characteristics are as given

in Figure 2In this case power is controlled such that it follows set

power-speed characteristics called tracking characteristicsThe measurement of actual turbine speed is done and itsrespective mechanical power is considered as the referencepower for the power control loop In Figure 2 ABCD curverepresents the tracking characteristics reference power iszero from zero to A position the power between A and Bis a straight line and the speed of B is to be greater thanA In between B and C the characteristic is the locus ofthe maximum turbine power Further the characteristic isstraight line from C to D

21 DFIG Modeling [26] DFIG modeling is carried with aninduction machine controlled in a synchronously rotating119889119902-axis frame with the 119889-axis found to be oriented along thestator-flux vector position This approach forms stator-fluxorientation (SFO) control and is used for modeling DFIGFollowing this manner a decoupled control existing betweenthe electrical torque and the rotor excitation current is notedAs a result the active and reactive powers are controlledindependently of each other Based on this the stator androtor voltage equations are given by

V119889119904= 119877119904119894119889119904minus 120596119904120582119902119904+

119889120582119889119904

119889119905

V119902119904= 119877119904119894119902119904+ 120596119904120582119889119904+

119889120582119902119904

119889119905

V119889119903= 119877119903119894119889119903minus (119904120596119904) 120582119902119903+

119889120582119889119903

119889119905

V119902119903= 119877119903119894119902119903+ (119904120596119904) 120582119889119903+

119889120582119902119903

119889119905

(4)

In the above equations stator voltages and stator currents aregiven by V

119904and 119894119904 rotor voltages and currents are represented

as V119903and 119894119903 stator and rotor resistances are given by119877

119904and119877

119903

and 120582119904and 120582

119903are respectively stator and rotor flux linkage

components As the machine is rotating in 119889119902-axis frame theindex ldquo119889rdquo represents direct axis component of the referenceframe and ldquo119902rdquo represent quadrature axis components of thereference frame In each of the equations (4) the given fluxlinkages are defined by

120582119889119904= minus119871119904119894119889119904+ 119871119898119894119889119903

120582119902119904= minus119871119904119894119902119904+ 119871119898119894119902119903

Turbine power characteristics (pitch angle beta = 0deg)

Turbine power characteristics

Tracking characteristics

A

B

C

D

5ms

12ms

162ms16

14

12

1

08

06

04

02

0

06 07 08 09 1 11 12 13Turb

ine o

utpu

t pow

er (p

u of

nom

inal

mec

hani

cal p

ower

)

Turbine speed (pu of generator synchronous speed)

Figure 2 Turbine and tracking characteristics

120582119889119903= 119871119903119894119889119903+ 119871119898119894119889119904

120582119902119903= 119871119903119894119902119903+ 119871119898119894119902119904

(5)

where 119871 in each of the cases represent their inductances and119871119898is the magnetizing inductance Equations (4) through (5)

represent the modeled equations for the considered DFIGsystem

22 Description of the Conventional DFIG Control SystemNumerous papers [20ndash26] have presented the conventionalDFIG control system which possess two control modulesrotor-side converter (RSC) control system and grid-sideconverter (GSC) control system This section discusses thedescription of the conventional DFIG control systemwhereinin this module no steps are taken to handle the faultride-through (FRT) of the DFIG This conventional controlmodule with RSC and GSC will generally be merged withthat of the hardware like STATCOM or crowbar systemwith the limitations discussed in Section 1 The conventionalDFIG control system is presented here in order to make thelucid difference between the applicability of the proposedoptimized control module and that of the conventionalcontrol system module

221 Rotor-Side Converter Control System Figure 3 depictsthe rotor-side converter control system The reactive currentflowing in the rotor-side converter 119862ROTOR controls thevoltage or the reactive power at the grid terminals The mainobjective of RSC is to regulate the stator active and reactivepower independently To have decoupled control of the statoractive (119875

119904) and reactive power (119876

119904) the rotor current gets

transformed to 119889-119902 components employing the referenceframe oriented with that of the stator fluxThe 119902-axis currentcomponent (119868

119902119903) controls the stator active power (119875

119904) and

the reference value of the active power (119875ref ) as in Figure 3


V

I

V

V

I

I

120596r

120596r

Is and IrIgc

AC voltagemeasurement

DroopXs

Varmeasurement

Trackingcharacteristic

Powermeasurement

Powerloosers

Vac

Vref

Qref

VX119904

Q

Pref

P

P1

+

+

+

+

+

minusminus

minus

minusminus

minus

minus

Ir

AC voltageregulator

Varregulator

Currentmeasurement

Powerregulator

Currentregulator

Vr

(Vdr Vqr)

Iqr_ref

Iqr

Idr

Idr_ref

Figure 3 Rotor-side converter control system

is obtained using the tracking characteristics via MaximumPower Point tracking The measured stator active powerwill get subtracted from the reference value of the activepower and the error is driven to the power regulator (powercontroller module) The output of the power regulator is119868119902119903 ref which is the reference value of the 119902-axis rotor currentThis value will be compared with that of the actual value(119868119902119903) and that error is flowed through the current regulator

(current controller) whose output is V119902119903(reference voltage for

the 119902-axis component)Until the reactive current is within the maximum current

values given by the converter rating the voltage gets regulatedat the reference voltage On operating the wind turbine inVAR regulation mode the grid terminals reactive power ismaintained constant by a VAR regulator In this work aDFIG which fed a weak AC grid is considered henceforthAC voltage regulator component is used and the reactivecontroller (VAR regulator) module is not used The outputof the AC voltage regulator is the reference 119889-axis current(119868119889119903 ref ) which must be injected in the rotor by the rotor-

side converter (119862ROTOR)The same current regulator (currentcontroller) as that of the power regulator is used to regulatethe actual (119868

119889119903) component of positive-sequence current to

its reference value The output from this regulator is the 119889-axis voltage (119881

119889119903) generated by rotor-side converter (119862ROTOR)

The current regulator is provided with the feed forward termsthat predict 119881

119889119903 119881119889119903

and 119881119902119903

denote the 119889-axis and 119902-axisvoltages respectively The magnitude of the reference rotorcurrent 119868

119903 ref is given by

119868119903 ref = 119868

119889119903 ref + 119868119902119903 ref (6)

Vdc

Vdc_ref

Igc

+

+

+

minus

minus

minus

Idgc

Iqgc

Idgc_ref

Iq_ref

Vgc

DC voltageregulator

Currentmeasurement

Currentregulator

Figure 4 Grid-side converter control system

The maximum value of this rotor current is limited to 1 puOn the other hand when 119868

119889119903 ref and 119868119902119903 ref are such that theirmagnitude is higher than 1 pu then the 119868

119902119903 ref component isreduced to bring the magnitude to 1 pu

222 Grid-SideConverter Control System Figure 4 shows theschematic of grid-side converter control system The GSCcontrol system is designed to regulate the voltage of theDC bus capacitor that is to maintain the DC-link voltageconstant Also GSC module is used to generate or absorbreactive power

GSC control systems consist of a measurement system tomeasure the 119889 and 119902 components of AC positive-sequencecurrents which is to be controlled as well as the DC voltage


StartInitialize the populationWhen (stopping condition is false) perform the followingBegin

Perform Mutation Operation [32]Update the new solution valuePerform Crossover operationModify the trial population vectorEvaluate the fitness functionPerform Selection operation to select the best operator

EndStop

Pseudocode 1 Pseudocode of standard DE

119881dc The outer regulation loop consists of a DC voltageregulator DC voltage regulator output is going to be thereference current (119868

119889gc ref ) to be fed for the current regulatorThe current in phase (119868

119889gc) with that of the grid voltagethat controls active power flow is also measured and thedifference between 119868

119889gc and 119868119889gc ref is one of the inputs to thecurrent regulator A current regulator forms the inner currentregulation loop and this regulator controls the magnitudeand phase of the voltage generated by 119862GRID from that of119868119889gc ref developed by the DC voltage regulator and given 119868

119902 ref It should be noted that the magnitude of the reference gridconverter current is given by

119868gc ref = 119868119889gc ref + 119868

119902119903 ref (7)

Themaximum value of this 119868gc ref is limited to a value definedby the converter maximum power at nominal voltage When119868119889gc ref and 119868

119902 ref are such that the magnitude is higher thanthis converter maximum power then the 119868

119902 ref componentis minimized to revert back the magnitude to the maximumvalueThe DC voltage is controlled with the signal 119894

119889gc ref andthe reactive power is controlled by means of 119868

119902 ref from thereactive power regulator [27ndash29]

3 Description of the Proposed DE-IDWNNOptimized Control System

The conventional DFIG control system discussed in Section 2does not take into account the fault ride-through of theinduction generator Thus in this paper attempt is made tocarry out ride-through fault eliminating the extra hardwarecomponent The optimized control module is designed in amanner to accomplish optimal synchronization between therotor-side converter control system and grid-side convertercontrol system and can handle the disturbances occurringin the system because of the fault This is taken care of inthe system even with the wind turbine feeding a weak ACgrid At this juncture it should be noted that the proposedcontroller is to perform effectively within a short span of timeand has to be not influenced by themeasured noise thatmightinterrupt in the system This controller is also to be designed

to take care of lackingmachine parameters information of thesystem

All the above discussed conditions will result in addednonlinearity of the system Hence to address the saiddifficulties as well as handle the unavoidable nonlinear-ity introduced in the system the proposed DE-IDWNNcontroller design lay on the evolutionary computationalintelligence techniques and will provide a better solution incomparison with that of the conventional approaches Tobe more accurate the controller design is performed witha double wavelet neural network controller whose weightsare optimized and tuned employing differential evolutionThe controllers employed are IDWNN controllers and thetuning of IDWNN controller meets the fault ride-throughIDWNN-FRT is carried out employing DE Further theapplicability of DE also performs weight optimization of theneural network controller to achieve better solution and fasterconvergence

31 Differential Evolution Algorithm An Overview A pop-ulation based stochastic evolutionary algorithm introducedby Storn and Price [30] is differential evolution algorithmwhich is in a way similar to genetic algorithms [31] employ-ing the operators crossover mutation and selection andsearches the solution space based on the weighted differencebetween the two population vectors To differentiate geneticalgorithmic approach relies on crossover and differentialevolution approach relies on mutation operation Mutationoperation is used in DE algorithm as a search mechanismand the selection operation directs the search towards theprospective regions of the search space Initially in DEpopulations of solution vectors are generated randomly at thebeginning and this initially generated population is improvedover generations using mutation crossover and selectionoperators In the progress of DE algorithm each of the newsolutions which resulted competes with that of the mutantvector and the better ones in the race win the competitionThe standard DE is as given in Pseudocode 1

32 Proposed Inertia Based Double Wavelet Neural NetworkThe neural network architecture of the proposed inertia


x1

x2

xn

f(x1)

f(x2)

h1

h0h2

hn

u(k) y(k)Σ

f(xn)

Figure 5 Architecture of inertia based double wavelet neuralnetwork

based double wavelet neural network is shown in Figure 5 Itconsists of input layer hidden layer 1 hidden layer 2 and theoutput layer Hidden layer 1 contains 119899-wavelet synapses withℎ1wavelet functions and hidden layer 2 contains one wavelet

synapse with ℎ2wavelet functions

Vector signal is as follows

119909 (119896) = (1199091 (119896) 1199092 (

119896) 119909119899 (119896)) (8)

where 119896 = 0 1 2 119899 which is the number of sample inputsin the training set

The output of the proposed inertia based double waveletneural network is expressed as

119910 (119896) = 119891(

119899

sum

119894=1

119891 (119909119894 (119896))) = 119891 (119906 (119896))

=

ℎ2

sum

119902=0

1205931199020(

119899

sum

119894=1

ℎ1

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896))1199081198950

119910 (119896) =

ℎ2

sum

119902=0

12059320(119906 (119896)1199081199020 (

119896))

(9)

where 119908119895119894(119896) represents synaptic weights

The wavelets differ between each other by dilationtranslation and bias factors are realized in each waveletsynapse The architecture of proposed inertia based waveletneural network with nonlinear wavelet synapses is shown inFigure 6

The expression for tuning of the output layer is given by

119864 (119896) =

1

2

(119889 (119896) minus 119910 (119896))2=

1

2

1198902(119896) (10)

where 119889(119896) is external training signalThe learning algorithmof output layer of the proposed neural network is

1199081198950 (

119896 + 1) = 1205781199081198950 (

119896) + 1205950 (119896) 119890 (119896) 1205931198950 (

119906 (119896)) (11)

Equation (11) can be represented in vector form as

1199080 (119896 + 1) = 120578119908

0 (119896) + 120595

0 (119896) 119890 (119896) 1205930 (

119906 (119896)) (12)

12059311

12059321

120593k1

12059312

12059322

120593k2

1205931n

1205932n

120593kn

12059310

12059320

120593k0

w11

w21

wh1

w12

w22

wh2

w1n

w2n

whn

w10

w00

w0n

w02

w01

w20

wh0

Σ

Σ

Σ

Σ Σ

x1

x2

xn

f(x1)

f(x2)

f(xn)

u y

Figure 6 Architecture of proposed IDWNNwith nonlinear waveletsynapse

where 119890(119896) is learning error 1205950(119896) is learning rate parameter

and 120578 is inertia factorThe analogy of the learning algorithm is given by

1199080 (119896 + 1) = 119908

0 (119896) +

119890 (119896) 1205930 (119906 (119896))

1205741199080

119894(119896)

1205741199080

119894(119896 + 1) = 120572120574

1199080

119894(119896) +

10038171003817100381710038171205930 (119906 (119896 + 1))

1003817100381710038171003817

2

(13)

The range of 120572 is between 0 and 1 The expression for tuningof the hidden layer is

119864 (119896) =

1

2

(119889 (119896) minus 1198910 (119906 (119896)))

2

=

1

2

(119889 (119896) minus 1198910(

119899

sum

119894=1

ℎ2

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896)))

2

(14)

Learning algorithm of hidden layer of the proposed neuralnetwork is

119908119895119894 (119896 + 1) = 120578119908

119895119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119895119894

(119909119894 (119896))

(15)


119908119894 (119896 + 1) = 120578119908

119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

(16)

where 119890(119896) is learning error 120595(119896) is learning rate parameterand 120578 is inertia factor


StartPhase I

Initialize the IDWNN weights learning rate and inertia factorRandomly generate weight valuesPresent the input samples to the network modelCompute net input of the network consideredApply activations to compute output of calculated net inputPerform the above process for hidden layer to hidden layer and hidden layer to output layerFinally compute the output from the output layer

Perform weight updation till stopping condition metPhase II

Present the computed outputs from IDWNNmodel into DEInvoke DEPerform MutationPerform CrossoverPerform SelectionDo carry out generations until fitness criteria metNote the points at which best fitness is achievedPresent the points of best fitness back to IDWNN

Phase IIIInvoke IDWNN with the inputs from the output of Phase IITune the weights to this IDWNN employing DE

Invoke DEPerform Phase I process for tuned weights of DEContinue Phase II

Repeat until stopping condition met (Stopping conditions can be number of iterationsgenerations orreaching the optimal fitness value)

Stop

Pseudocode 2 Pseudocode for DE-IDWNN controller

The analogy of the learning algorithm in (15) is given by

119908119894 (119896 + 1) = 119908

119894 (119896) +

119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

120574119908119894

119894(119896)

120574119908119894

119894(119896 + 1) = 120572120574

119908119894

119894(119896) +

1003817100381710038171003817120593119894(119909119894 (119896 + 1))

1003817100381710038171003817

2

(17)

The range of 120572 is between 0 and 1 The properties of this pro-posed algorithm have both smoothing and approximatingThe inertia parameter helps in increasing the speed of conver-gence of the system The inertia parameter also prevents thenetwork from getting converged to local minimaThe inertiafactor is between values 0 and 4

33 Proposed DE Based IDWNN Optimization ControllerBased on the given inputs and outputs of the system theproposed inertia based double wavelet neural networkmodelis designed and the DE approach is employed to tune theweights of the IDWNN and to tune the outputs of the neuralnetwork model The proposed pseudocode for DE basedIDWNN optimization controller is as given in Pseudocode 2

The proposedDE-IDWNNcontroller is applied to handlefault ride-through of grid-connected doubly fed inductiongenerators and the methodology adopted for handling faultride-through using this proposed controller is presented inthe following section

Proposed optimized controller

Fault detection

dqabc PWM

Regular RSC controlsystem

Current regulator Vqr

Vdr

+

minus

Ncrf

VS_ref

VS

XIDWNNFRT

V998400qr

ilowastr

Vlowastdc

Figure 7 Proposed DE-IDWNN optimized control system

4 Fault Ride-Through of DFIG Module UsingProposed DE-IDWNN Controller

DFIG module with fault ride-through employing proposedDE-IDWNN controller is as shown in Figure 7 In theproposed design GSC control system remains the same asthat of the conventional DFIG module and the modificationoccurs in the RSC control system module


The proposed optimized controller starts its operationonly when the AC voltage V

119904exceeds ten percent higher

than that of the set reference voltage The constraints thatshould be taken care of for safe guarding the DFIG arethe DC-link overvoltage and the rotor overcurrent Boththese constraints should not exceed their set limits in theconsidered restoring period Also care should be taken inorder to transfer the additional energy generated in the rotorvia the converters to the gridThis process of transferring theadditional energy induced will enable the DC-link voltageand rotor current to maintain their respective normal valuesA setback on performing this operation is that when therotor current is decreased by fast transfer of the stored energyfrom rotor to grid there might be a chance that the DC-linkvoltage increases abruptly and it may get deviated from thenormal limits On the other hand without fast mechanismslowly decreasing the rotor current such that the DC-linkvoltage does not exceed the limit may result in the rotorcurrent reaching abnormal transient points Henceforth forimplementing an effective and efficient fault ride-throughthe transition signals of the rotor current should also considerthe nominal values of theDC-link voltageThe proposed con-troller is the inertia double wavelet neural network controllerand the input to IDWNNFRT is 119881lowastdc and 119894

lowast

119903and the output of

the neural network controller is 119873crf 119881lowast

dc and 119894lowast

119903act as the

inputs to the IDWNN optimized controller and are given by

119881lowast

dc =119881dc minus 119881ss119881max minus 119881ss

119894lowast

119903=

119894119903minus 119894ss

119894max minus 119894ss

(18)

where119881ss and 119894ss indicate the steady state values119881max and 119894maxspecify the maximum values 119894

119903is the rotor rms current and

119881dc is the DC-link voltageBoth the input values are normalized before they are

fed into the neural network controller Design of neuralnetwork controller is performed as shown in Table 1 and thetraining process of the controller is carried out to obtain119873crf Originally the controller is designed by consideringrandomweights into the IDWNN and thenDE is adopted fortuning theweight values of the IDWNNcontroller Nonlinearwavelet synapse is employed during the training process forfaster convergence

Along with the proposed IDWNN controller differentialevolution is employed to tune the weights of IDWNNcontroller as well as minimize the objective function or thefitness function as given in

min119891 = [

119881dcmax minus 119881ss119881max minus 119881ss

]

2

+ [

119894119903max minus 119894ss119894max minus 119894ss

]

2

(19)

119881max and 119894119903max represent the maximum value of the signals

during the entire restoring duration On applying DE for theconsidered problem each set of variables is represented bybinary strings called chromosomes and chromosomes madeup individual entities called genes All the chromosomesgenerated result in the formation of population In thisproblem under consideration population size is 20 that is 20

Table 1 Proposed IDWNN controller parameters

Parameters of the proposed IDWNN controller Set valuesNumber of input neurons 2Number of output neurons 1Inertia factor 175Number of hidden layer neurons 8Learning rate parameter 1

05

0

minus05

1

05

0 002

0406

081

DC-link voltage Rotor current

IDW

NN

out

putN

crf

Figure 8 Three-dimensional output for the trained controller

chromosomes and the length of chromosome is taken to be8 that is 8 genes comprise a chromosomeThe initial 6 genesrepresent the binary strings of the range of input parametersand the remaining genes represent the output range Theprocess of DE is carried out as given by the pseudocodein Pseudocode 1 DE is invoked and activated to search forminimizing the optimization function given in (19) In DEprocess mutation is carried out at the initial process andthen crossover and selection operations are performed Thesearch process is carried out to find the maximized value for(19) and the procedure is continued until a specified stoppingcondition is reached

The importance of the fitness function in (19) is as followsgenerally for a variety of problems the objective functionwill be an integral function where the output is the completebehavior of the system in a particular time interval In thisproblem domain the aim is to limit the instantaneous valueof rotor current and DC-link voltage so as to eliminate thetripping ofDFIGAs a result the objective function is the sumof the squared components that are to be minimized Also itis to be noted in this case that the aim here is not only tominimize the specified objective function in (19) but also tomaintain the balance between the rotor current and DC-linkvoltage in the range of acceptable values Figure 8 shows thethree-dimensional graph for the IDWNNFRT output119873crf withrespect to119881lowastdc and 119894

lowast

119903after the DE optimizationThe proposed

DE-IDWNN controller can be employed for various sizes ofmachines The training of the neural network controller andthe DE optimization process remains the same in all cases


AC grid

Load 1

Line 1

Load 3

Synchronous generator

Load 2

WT

PCC

Line 3 Line 4

Line 2

Fault

20kV400V

20kV690V

Figure 9 Electrical system considered for validating the proposed controller

5 Numerical Experimentation andSimulation Results

TheproposedDE-IDWNNcontroller is validated by applyingthe said control strategy for 15MWwind turbines connectedto a 25 kV distribution system exporting power to a 120 kVgrid through a 30 km 25 kV feeder The electrical system [26]considered for validating the proposed control design is givenin Figure 9 Also this work basically deals with handlingthree-phase symmetrical grid faultsThe occurrence of three-phase faults is noted at 05 seconds and the simulation iscarried out The simulation response is noted for both thetraditional DFIG control system and that of the proposedcontrol system design From the simulation response of thetraditional control system it can be noted that this systemrequires an auxiliary system for fault ride-through optionand observes the limitations as discussed in Section 1 Duringthe process of simulation it is observed that the DFIGalong with the proposed controller handles the completeresponse at fault periods and after fault periods and getsride-through fault eliminating the applicability of auxiliaryhardware Table 2 shows the specifications for the parametersof the DFIG system and the grid system [26] considered forsimulation The entire proposed control strategy was run inMATLABR2009 environment and executed in Intel Core2Duo Processor with 227GHz speed and 200GB RAMSimulink environment in MATLAB is used to model theconverter modules All the simulations are carried out forwind speed of 12ms

Figures 10ndash17 show the response of the traditional controlsystem generated for the wind speed of 12ms From Figures10 and 11 it can be observed that the DC voltage exceeds theset limit resulting in damaging the capacitor present Alsothe rotor current increases in an advert manner nearly 100compared to that of the acceptable value in the rotor-sideconverter control systemDuring this conventional controlleraction it is noted that the entire response of the system variesin an erroneous manner resulting in more fluctuations in the

Table 2 System specifications of DFIG system and grid module

Specifications of DFIG systemRated power 15MWRated stator voltage 690VNominal wind speed 12msStator resistance 000706 puRotor resistance 0005 puStator leakage inductance 01716 puRotor leakage inductance 0156 puMagnetizing inductance 29 puTurns ratio 27Rotational inertia 504 sNominal DC-link rated voltage 1200VDC bus capacitor 60millifarads

Specifications of AC gridRated voltage 20 kVFrequency 50HzShort circuit ratio 223Load 1 power 400 kW amp 120 kVarLoad 2 power 500 kW amp 150 kVarLoad 3 power 50KW amp 15 kVar

Length of transmission linesLine 1 and Line 3 15 kmLine 2 and Line 4 30 kmSynchronous machine speed 1500 rpmSynchronous machine power 85 kVA

grid sideThus this conventional control design ismodified toenable appropriate handling of ride-through fault conditionsThe proposed DE-IDWNN controller is simulated for 100generations and the entire response observed during the sim-ulation response is as shown in Figure 18 through Figure 25Nonlinear wavelet synapse is utilized for the updating processof IDWNN controller and the DE is initiated to tune the


145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)

Figure 10 Response of three-phase voltage for traditional controlsystem

1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2

Figure 11 Response of DC-link voltage for traditional controlsystem

14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)

Figure 12 Response of rotor current for traditional control system

14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)

Figure 13 Response of stator current for traditional control system

145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)

Figure 14 Response of wind turbine active power output fortraditional control system

14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02

Figure 15 Response of wind turbine reactive power output fortraditional control system

14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)

Figure 16 Response of rotor speed for traditional control system

10 11 12 13 14 15 16 17 18 19 20minus7minus6minus5minus4minus3minus2minus1

0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)

Figure 17 Response of 119902-component of rotor voltage for traditionalcontrol system


14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)

Figure 18 Response of three-phase stator voltage for proposed DE-IDWNN controller

146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)

Figure 19 Response of rotor current for proposed DE-IDWNNcontroller

weights of IDWNN controller and that of the objective func-tion as given in (19) The evaluation of the fitness functionis studied at 85 voltage dip The response of the observedparameters during simulation process proves the removalof fluctuations occurring during the conventional controldesign and as well it is noted that it reaches the steady stateat the earliest The fault ride-through of the DFIG module isachieved at the point wherein optimal solution is arrived atemploying the proposed controller design At this junctureovervoltages noted at DC-link point and overcurrents at therotor side are observed to be well within the limit of themaximum set threshold values As a result the damaging ofthe capacitor is protected and fault and postfault occurrenceto the grid are controlled in an effective manner Figures19 and 20 show the rotor current and stator current withfluctuations reduced obtaining the steady state value at afaster rate

Figures 21 and 22 show the response of the wind turbineoutput active and reactive power Fundamentally in thisproposed controller design RSC will not be cut during thefault condition and this RSC provides the required reactivepower to the grid in case of handling the sufficient voltagedrops occurring But the proposed DE-IDWNN controlleracts to the fault effectively arriving at an optimal solutionand thus it will prevent more amount of reactive power frombeing transferred from DFIG after the fault In the proposeddesign DFIG supplies the grid with the required amount of

144 146 148 15 152 154 156 158minus6minus4minus2

02468

101214

Time (s)

Stat

or cu

rren

t (kA

)

Figure 20 Response of stator current for proposed DE-IDWNNcontroller

146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)

Figure 21 Response of wind turbine output active power forproposed DE-IDWNN controller

reactive power and is found to sustain that of the grid voltageAt the time of fault the rotor speed of the proposed controllerincreases which is seen in Figure 23 The increase in speed ofthe rotor is noted due to the capacity of the wind turbine tostore huge energy At the occurrence of fault there will be ahuge drop in the voltage and the wind power is transferred askinetic energy to the rotor without dissipating that of the gridAt the end of the fault duration the grid receives this energyand the increase in the speed of the rotor will automaticallyget settled to the earlier value (ie the value before the faulthas occurred)

Figures 24 and 25 show the 119902-component and themodified 119902-component of the rotor voltage of the proposedcontroller design The occurrence of transients is noted andthe proposed controller acts in amanner to bring back theACvoltage to its set value resulting in the above said transientIn case during simulation process higher voltage dip occursDFIG is allowed to disconnect This is the bad conditionon getting connected to the grid Thus the entire simulationstudy is carried out for wind speed at 12ms and voltage dip of85 The proposed controller found that ride-through faultoccurred Table 3 shows the fitness function values evolvedduring various generations At the end of the 100 generationsit is noted that the fitness value reached 40021 The DC-link voltage and the rotor current are found to be maintainedwithin the said permissible limits at this optimal point


13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

Figure 22 Response of wind turbine output reactive power forproposed DE-IDWNN controller

minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)

Figure 23 Response of rotor speed for proposed DE-IDWNNcontroller

Table 3 Fitness function evolved during generations

Generations Fitness function (119891) as in (19)10 8674120 8725930 6421340 7093250 4321860 5983170 4765180 4312490 42618100 40021

From Table 4 it can be observed that the proposedcontroller achieves a minimal fitness function value of 40021in comparisonwith that of the earlier othermethods availablein the literature proving the effectiveness of the controllerFigure 26 shows the variation of fitness function with respectto generations and from Figure 26 it can be inferred thatthe search process enables the fitness function to reach aminimum value

14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)

Figure 24 Response of 119902-component of rotor voltage for proposedDE-IDWNN controller

10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)

Figure 25 Response of the modified 119902-component of rotor voltagefor proposed DE-IDWNN controller

Table 4 Comparison of the fitness function values

Methods employed Optimal best value of ldquo119891rdquo in(19)

GA based approach [26] 95612ANN controller [5] 72314ANFIS controller [20] 71230Proposed DE-IDWNN controller 40021

6 Conclusion

This paper proposed a novel heuristic based controllermodule employing differential evolution and neural networkarchitecture to improve the low-voltage ride-through rate ofgrid-connected wind turbines which are connected alongwith doubly fed induction generators (DFIGs)The limitationof the traditional control system is taken care of in this workby introducing heuristic controllers that remove the usageof crowbar and ensure that wind turbines supply necessaryreactive power to the grid during faults The controllerdesigned in this paper enhances the DFIG converter duringthe grid fault and this controller takes care of the ride-throughfault without employing any other hardware modules Theproposed controller design is validated for a case study ofwind farm with 15MW wind turbines connected to a 25 kVdistribution system exporting power to a 120 kV grid Theresults simulated prove the effectiveness of the controller


10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859

Number of generations

Com

pute

d fit

ness

val

ues

Figure 26 Variation of fitness functions with respect to number ofgenerations

design in comparison with that of the methods available inthe literature

Conflict of Interests

The authors declare that there is no conflict of interests forpublishing this paper in this journal

References

[1] J P A Vieira M V A Nunes U H Bezerra and A CDo Nascimento ldquoDesigning optimal controllers for doubly fedinduction generators using a genetic algorithmrdquo IET Genera-tion Transmission amp Distribution vol 3 no 5 pp 472ndash4842009

[2] D Nagaria G N Pillai and H O Gupta ldquoComparison ofcontrol schemes for frequency support in DFIG based WECSrdquoInternational Energy Journal vol 11 no 1 pp 17ndash28 2010

[3] P Bhatt R Roy and S P Ghoshal ldquoDynamic participationof doubly fed induction generator in automatic generationcontrolrdquo Renewable Energy vol 36 no 4 pp 1203ndash1213 2011

[4] J P da Costa H Pinheiro T Degner and G Arnold ldquoRobustcontroller for DFIGs of grid-connected wind turbinesrdquo IEEETransactions on Industrial Electronics vol 58 no 9 pp 4023ndash4038 2011

[5] B Dong S Asgarpoor and W Qiao ldquoANN-based adaptive PIcontrol for wind turbine with doubly fed induction generatorrdquoin Proceedings of the 43rd North American Power Symposium(NAPS rsquo11) pp 1ndash6 IEEE August 2011

[6] N Ghadimi ldquoGenetically adjusting of PID controllers coef-ficients for wind energy conversion system with doubly fedinduction generatorrdquoWorld Applied Sciences Journal vol 15 no5 pp 702ndash707 2011

[7] J Liu D Xu and Y Lu ldquoNonlinear model predictive controlof the speed of double-fed induction generatorrdquo Power SystemTechnology vol 35 no 4 pp 159ndash163 2011

[8] W Qiao J Liang G K Venayagamoorthy and R G HarleyldquoComputational intelligence for control of wind turbine gen-eratorsrdquo in Proceedings of the IEEE Power and Energy SocietyGeneral Meeting pp 1ndash6 San Diego Calif USA July 2011

[9] Y Song H Nian J Hu and J-W Li ldquoMulti-objective opti-mization control of DFIG system under distorted grid voltageconditionsrdquo in Proceedings of the International Conference on

Electrical Machines and Systems (ICEMS rsquo11) pp 1ndash6 IEEEAugust 2011

[10] K Vinothkumar and M P Selvan ldquoNovel scheme for enhance-ment of fault ride-through capability of doubly fed inductiongenerator based wind farmsrdquo Energy Conversion and Manage-ment vol 52 no 7 pp 2651ndash2658 2011

[11] B P Numbi J L Munda and A A Jimoh ldquoOptimal VARcontrol in grid-integrated DFIG-based wind farm using swarmintelligencerdquo in Proceedings of the IEEE Power and EnergySociety Conference and Exposition in Africa (PowerAfrica rsquo12)pp 1ndash6 IEEE Johannesburg South Africa July 2012

[12] H-M Wang C-L Xia Z-W Qiao and Z-F Song ldquoA hybridparticle swarm optimization algorithm for optimum electro-magnetic design of DFIG-based wind power systemrdquo Journalof Tianjin University Science and Technology vol 45 no 2 pp140ndash146 2012

[13] Y-M You T A Lipo and B-I Kwon ldquoOptimal design of agrid-connected-to-rotor type doubly fed induction generatorfor wind turbine systemsrdquo IEEE Transactions on Magnetics vol48 no 11 pp 3124ndash3127 2012

[14] M Amelian R Hooshmand A Khodabakhshian and HSaberi ldquoSmall signal stability improvement of a wind turbine-based doubly fed induction generator in a micro grid environ-mentrdquo in Proceedings of the 3th International eConference onComputer andKnowledge Engineering (ICCKE rsquo13) pp 384ndash389IEEE November 2013

[15] E E El-Araby ldquoOptimal VAR expansion considering capabilitycurve of DFIGwind farmrdquo in Proceedings of the IEEE Power andEnergy Society General Meeting (PES rsquo13) 5 p 1 IEEE can July2013

[16] N Sa-Ngawong and I Ngamroo ldquoOptimal fuzzy logic-basedadaptive controller equipped with DFIG wind turbine forfrequency control in stand alone power systemrdquo in Proceedingsof the IEEE Innovative Smart Grid Technologies (ISGT Asia rsquo13)IEEE November 2013

[17] Y Tang P Ju H He C Qin and F Wu ldquoOptimized controlof DFIG-based wind generation using sensitivity analysis andparticle swarm optimizationrdquo IEEE Transactions on Smart Gridvol 4 no 1 pp 509ndash520 2013

[18] S Beheshtaein ldquoFAΘPSO based fuzzy controller to enhanceLVRT capability of DFIG with dynamic referencesrdquo in Pro-ceedings of the 23rd International Symposium on IndustrialElectronics (ISIE rsquo14) pp 471ndash478 IEEE Istanbul Turkey June2014

[19] S S Dhillon S Marwaha and J S Lather ldquoRobust loadfrequency control of micro grids connected with main grids ina regulated and deregulated environmentrdquo in Proceedings of theInternational Conference on Recent Advances and Innovations inEngineering (ICRAIE rsquo14) pp 1ndash9 IEEE May 2014

[20] N Govindarajan and D Raghavan ldquoDoubly fed induction gen-erator based wind turbine with adaptive neuro fuzzy inferencesystem controllerrdquoAsian Journal of Scientific Research vol 7 no1 pp 45ndash55 2014

[21] X Kong X Liu and K Y Lee ldquoData-driven modelling of adoubly fed induction generator wind turbine system based onneural networksrdquo IET Renewable Power Generation vol 8 no8 pp 849ndash857 2014

[22] W Wang and Y M Chu ldquoDFIG based wind power generationsystem RSC controller designrdquo Applied Mechanics and Materi-als vol 513ndash517 pp 3911ndash3914 2014

[23] F Wu H Wang Y Jin and P Ju ldquoParameter tuning ofdoubly fed induction generator systems for wind turbines


based on orthogonal design and particle swarm optimizationrdquoAutomation of Electric Power Systems vol 38 no 15 pp 19ndash242014

[24] Z Xie and X Li ldquoThe virtual resistance control strategyfor HVRT of doubly fed induction wind generators basedon particle swarm optimizationrdquo Mathematical Problems inEngineering vol 2014 Article ID 350367 11 pages 2014

[25] S Zhang Q Jiang Y Chen W Zhang and J Song ldquoDesign ofadditional SSR damping controller for power systemwithDFIGwind turbinerdquo Electric Power Automation Equipment vol 34no 6 pp 36ndash43 2014

[26] T D Vrionis X I Koutiva and N A Vovos ldquoA geneticalgorithm-based low voltage ride-through control strategy forgrid connected doubly fed induction wind generatorsrdquo IEEETransactions on Power Systems vol 29 no 3 pp 1325ndash13342014

[27] G Cai C Liu and D Yang ldquoRotor current control of DFIGfor improving fault ridemdashthrough using a novel sliding modecontrol approachrdquo International Journal of Emerging ElectricPower Systems vol 14 no 6 pp 629ndash640 2013

[28] J-H Liu C-C Chu and Y-Z Lin ldquoApplications of nonlinearcontrol for fault ride-through enhancement of doubly fedinduction generatorsrdquo IEEE Journal of Emerging and SelectedTopics in Power Electronics vol 2 no 4 pp 749ndash763 2014

[29] M Mohseni and S M Islam ldquoReview of international gridcodes for wind power integration diversity technology anda case for global standardrdquo Renewable and Sustainable EnergyReviews vol 16 no 6 pp 3876ndash3890 2012

[30] R Storn and K PriceDifferential EvolutionmdashA Simple and Effi-cient Adaptive Scheme for Global Optimization over ContinuousSpaces ICSI Berkeley Calif USA 1995

[31] R Storn and K Price ldquoDifferential evolutionmdasha simple andefficient heuristic for global optimization over continuousspacesrdquo Journal of Global Optimization vol 11 no 4 pp 341ndash359 1997

[32] K Price RM Storn and J A LampinenDifferential EvolutionA Practical Approach to Global Optimization Springer 2006

Submit your manuscripts athttpwwwhindawicom

Computer Games Technology

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Distributed Sensor Networks


Advances in

FuzzySystems

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014


ReconfigurableComputing

Hindawi Publishing Corporation httpwwwhindawicom Volume 2014


Applied Computational Intelligence and Soft Computing

thinspAdvancesthinspinthinsp

Artificial Intelligence

HindawithinspPublishingthinspCorporationhttpwwwhindawicom Volumethinsp2014

Advances inSoftware EngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Electrical and Computer Engineering

Journal of

Journal of

Computer Networks and Communications


Hindawi Publishing Corporation

httpwwwhindawicom Volume 2014

Advances in

Multimedia


Biomedical Imaging


ArtificialNeural Systems

Advances in


RoboticsJournal of



Computational Intelligence and Neuroscience

Industrial EngineeringJournal of


Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Human-ComputerInteraction

Advances in

Computer EngineeringAdvances in



Vieira et al [1] carried out a work on genetic algorithmbased optimal controllers to the rotor-side converter ofdoubly fed induction generators (DFIGs) the DFIG con-verter control action performed by proportional and inte-gral controllers The presented approach improves transientstability margin of the power system and also has betterglobal system dynamic behavior during and after the faultperiod In [2]Nagaria et al analyzed various control strategiesfor frequency support in DFIG-based WECS based on thecomparative analysis the following point can observe mini-mum frequency deviation and maximum release of energyobtained by tuning the controller parameter using PSO andimproved performance is achieved by inertial response ofthe system with the PSO optimized proportional controllerBhatt et al [3] presented doubly fed induction generatordynamic participation in automatic generation control fre-quency response of the system is improved by frequencycontrol support function and craziness-based particle swarmoptimization (CRPSO) utilized to obtain the optimal DFIGparameter transient response of area frequency and tie-linepower deviation

A novel robust controller [4] is suggested in a stationaryreference frame for grid-integrated doubly fed inductiongenerators based wind turbines The voltage and flux equa-tions in 120572120573 coordinates used DFIG dynamic model designthe disturbances and uncertainties intrinsic to the systemare accounted for as perturbation terms are included tothe nominal model ANN [5] based adaptive PI controllerson DFIG driven based wind turbine in order to improvetransient performance of DFIG in all wind speed conditionsANN is used to predict the optimal values of parameter andadaptive PI controllers according to different wind speedsdynamically change PI gain values A genetic algorithm [6]based PID controller coefficients adjustment on wind energyconversion system with doubly fed induction generator Liuet al [7] pointed out the DFIG speed controller based onthe nonlinear intelligent model predictive control algorithmSVM is adapted to establish the speed of DFIG and PSO isutilized to solve the rolling optimization the forecasted erroris used to form a closed-loop control Various computationalintelligence based controls of wind turbine generators such asPSOMVO adaptive critic design based adaptive control andshunt FACTS devices are analyzed in [8]

Multiobjective optimization [9] control methodology ofa doubly fed induction generator (DFIG) considering thedistorted grid voltage conditions was dealt with The fourvarious control targets are suggested to restrain the har-monic components in statorrotor current or the ripples inthe stator output activereactive power and electromagnetictorque Vinothkumar and Selvan [10] carried out a workon a new enhancement of fault ride-through capability andsubsequent betterment of rotor speed stability of doubly fedinduction generator based wind farms The wind turbinemechanical energy input is stored during the grid fault andat the moment of fault clearance it is utilized hence thecharging DC-link capacitor by grid-side converter is relievedParticle swarm optimization [11] based optimal VAR controlin grid-connectedDFIG-basedwind farmhas been presentedto optimize reactive power from reactive power capability

(RPC) of DFIGs in order to obtain system active power lossminimization

A hybrid particle swarm optimization [12] (HPSO) algo-rithm for the DFIG optimum design is presented in thiswork and the searching performance is improved by adaptingthe diversity-guided adaptive mutation and fitness-guidedindividual fuzzy inertia weight An optimal grid-integrated-to-the-rotor type doubly fed induction generator design forwind turbine systems is developed in [13] Kriging modelbased on Latin hypercube sampling and genetic algorithm areadapted to maintain the efficiency and maximize the torqueper weight Amelian et al [14] carried out a work on a smallsignal stability improvement based on eigenvalue analysis ofa wind turbine-based doubly fed induction generator in amicro grid environment The stability margin and dampingratio of critical modes are enhanced by using particle swarmoptimization algorithm

An optimal VAR expansion based on capability curve ofDFIG wind farm has been proposed in [15] The proposedapproach reduces the sum of the annual investment costof the new VAR devices and the annual expected energyloss cost Design of adaptive controller [16] using optimalfuzzy logic of doubly fed induction generator (DFIG) windturbine for standalone power system frequency control hasbeen proposed in this workThe particle swarm optimizationis employed to adapt and optimize the fuzzy logic designmembership function and the control rules

Tang et al [17] carried out a work on an optimalcontrol of doubly fed induction generators- (DFIGs-) basedwind generation adapting trajectory and frequency domaininformation based sensitivity analysis and particle swarmoptimization The critical parameters the unified dominatecontrol parameters (UDCP) are identified using sensitivityanalysis and PSO is utilized to obtain the control goal bysearching the optimal values A fuzzy controller is pro-posed in [18] based on fuzzy adaptive theta particle swarmoptimization (FAΘPSO) applied to a doubly fed inductiongenerator (DFIG) to run at medium voltage FAΘPSO algo-rithm is used to obtain optimum values of membershipfunctions A load frequency control (LFC) of micro gridsconnected with main grids in a regulated and deregulatedenvironment the PSO (particle swarm optimization) tunedPI controller gains have improved performance compared tothe conventional PI controller parameters proposed in [19]An adaptive neurofuzzy inference system [20] controller ispointed out for doubly fed induction generator based windturbine The presented approach controls the current andvoltage ripple within 001 pu as well as reducing the powerloss

Kong et al [21] proposed data-driven based modeling ofdoubly fed induction generator wind turbine system adaptingneural networksTheDFIGmodel is developed based onneu-ral networks and neurofuzzy networks by using large amountof input-output online measurement data from the selectedmonths An RSC controller [22] is proposed for DFIG-basedwind power generation system In this case the stator-flux-oriented (SFO) vector control based design is adapted andRSC controller for DFIG with double closed-loop system ismodeled using rotor current control loop and power control


Turbine

WindDrive train

Rotor

Pitch angle

Stator

Inductiongenerator

AC DC AC

ACDCAC converter

Control

Vr Vgc

Crotor Cgrid

L

Three-phasegrid







119877




119875119898= 119879119898120596119903

119875119904= 119879119890120596119904

(1)



output 119879119898



119903and 120596



119869119889120596119903119889119905

= 119879119898minus 119879119890 (2)


119898= 119879119890and 119875

119898= 119875119904+ 119875119903 where 119875

119903is the rotor




120596119904

= minus119904119879119898120596119904= minus119904119875

119904

(3)






V119889119904= 119877119904119894119889119904minus 120596119904120582119902119904+

119889120582119889119904

119889119905

V119902119904= 119877119904119894119902119904+ 120596119904120582119889119904+

119889120582119902119904

119889119905

V119889119903= 119877119903119894119889119903minus (119904120596119904) 120582119902119903+

119889120582119889119903

119889119905

V119902119903= 119877119903119894119902119903+ (119904120596119904) 120582119889119903+

119889120582119902119903

119889119905

(4)




119904and119877

119903

and 120582119904and 120582



120582119889119904= minus119871119904119894119889119904+ 119871119898119894119889119903

120582119902119904= minus119871119904119894119902119904+ 119871119898119894119902119903




A

B

C

D

5ms

12ms

162ms16

14

12

1

08

06

04

02

0

06 07 08 09 1 11 12 13Turb

ine o

utpu

t pow

er (p

u of

nom

inal

mec

hani

cal p

ower

)



120582119889119903= 119871119903119894119889119903+ 119871119898119894119889119904

120582119902119903= 119871119903119894119902119903+ 119871119898119894119902119904

(5)









119904) and



V

I

V

V

I

I

120596r

120596r

Is and IrIgc


DroopXs

Varmeasurement


Powermeasurement

Powerloosers

Vac

Vref

Qref

VX119904

Q

Pref

P

P1

+

+

+

+

+

minusminus

minus

minusminus

minus

minus

Ir

AC voltageregulator

Varregulator

Currentmeasurement

Powerregulator

Currentregulator

Vr

(Vdr Vqr)

Iqr_ref

Iqr

Idr

Idr_ref











119889119903 119881119889119903

and 119881119902119903



119868119903 ref = 119868

119889119903 ref + 119868119902119903 ref (6)

Vdc

Vdc_ref

Igc

+

+

+

minus

minus

minus

Idgc

Iqgc

Idgc_ref

Iq_ref

Vgc

DC voltageregulator

Currentmeasurement

Currentregulator










EndStop







119868gc ref = 119868119889gc ref + 119868

119902119903 ref (7)













x1

x2

xn

f(x1)

f(x2)

h1

h0h2

hn

u(k) y(k)Σ

f(xn)





119909 (119896) = (1199091 (119896) 1199092 (

119896) 119909119899 (119896)) (8)



119910 (119896) = 119891(

119899

sum

119894=1

119891 (119909119894 (119896))) = 119891 (119906 (119896))

=

ℎ2

sum

119902=0

1205931199020(

119899

sum

119894=1

ℎ1

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896))1199081198950

119910 (119896) =

ℎ2

sum

119902=0

12059320(119906 (119896)1199081199020 (

119896))

(9)




119864 (119896) =

1

2

(119889 (119896) minus 119910 (119896))2=

1

2

1198902(119896) (10)


1199081198950 (

119896 + 1) = 1205781199081198950 (

119896) + 1205950 (119896) 119890 (119896) 1205931198950 (

119906 (119896)) (11)


1199080 (119896 + 1) = 120578119908

0 (119896) + 120595

0 (119896) 119890 (119896) 1205930 (

119906 (119896)) (12)

12059311

12059321

120593k1

12059312

12059322

120593k2

1205931n

1205932n

120593kn

12059310

12059320

120593k0

w11

w21

wh1

w12

w22

wh2

w1n

w2n

whn

w10

w00

w0n

w02

w01

w20

wh0

Σ

Σ

Σ

Σ Σ

x1

x2

xn

f(x1)

f(x2)

f(xn)

u y




1199080 (119896 + 1) = 119908

0 (119896) +

119890 (119896) 1205930 (119906 (119896))

1205741199080

119894(119896)

1205741199080

119894(119896 + 1) = 120572120574

1199080

119894(119896) +

10038171003817100381710038171205930 (119906 (119896 + 1))

1003817100381710038171003817

2

(13)


119864 (119896) =

1

2

(119889 (119896) minus 1198910 (119906 (119896)))

2

=

1

2

(119889 (119896) minus 1198910(

119899

sum

119894=1

ℎ2

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896)))

2

(14)


119908119895119894 (119896 + 1) = 120578119908

119895119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119895119894

(119909119894 (119896))

(15)


119908119894 (119896 + 1) = 120578119908

119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

(16)



StartPhase I







Stop



119908119894 (119896 + 1) = 119908

119894 (119896) +

119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

120574119908119894

119894(119896)

120574119908119894

119894(119896 + 1) = 120572120574

119908119894

119894(119896) +

1003817100381710038171003817120593119894(119909119894 (119896 + 1))

1003817100381710038171003817

2

(17)





Fault detection

dqabc PWM



Vdr

+

minus

Ncrf

VS_ref

VS

XIDWNNFRT

V998400qr

ilowastr

Vlowastdc








lowast



dc and 119894lowast

119903act as the


119881lowast


119894lowast

119903=

119894119903minus 119894ss


(18)






min119891 = [


]

2

+ [


]

2

(19)





05

0

minus05

1

05

0 002

0406

081


IDW

NN

out

putN

crf




lowast




AC grid

Load 1

Line 1

Load 3


Load 2

WT

PCC

Line 3 Line 4

Line 2

Fault

20kV400V

20kV690V











145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)


1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2


14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)


14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)


145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)


14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02


14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)



0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)



14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




Turbine

WindDrive train

Rotor

Pitch angle

Stator

Inductiongenerator

AC DC AC

ACDCAC converter

Control

Vr Vgc

Crotor Cgrid

L

Three-phasegrid







119877




119875119898= 119879119898120596119903

119875119904= 119879119890120596119904

(1)



output 119879119898



119903and 120596



119869119889120596119903119889119905

= 119879119898minus 119879119890 (2)


119898= 119879119890and 119875

119898= 119875119904+ 119875119903 where 119875

119903is the rotor




120596119904

= minus119904119879119898120596119904= minus119904119875

119904

(3)






V119889119904= 119877119904119894119889119904minus 120596119904120582119902119904+

119889120582119889119904

119889119905

V119902119904= 119877119904119894119902119904+ 120596119904120582119889119904+

119889120582119902119904

119889119905

V119889119903= 119877119903119894119889119903minus (119904120596119904) 120582119902119903+

119889120582119889119903

119889119905

V119902119903= 119877119903119894119902119903+ (119904120596119904) 120582119889119903+

119889120582119902119903

119889119905

(4)




119904and119877

119903

and 120582119904and 120582



120582119889119904= minus119871119904119894119889119904+ 119871119898119894119889119903

120582119902119904= minus119871119904119894119902119904+ 119871119898119894119902119903




A

B

C

D

5ms

12ms

162ms16

14

12

1

08

06

04

02

0

06 07 08 09 1 11 12 13Turb

ine o

utpu

t pow

er (p

u of

nom

inal

mec

hani

cal p

ower

)



120582119889119903= 119871119903119894119889119903+ 119871119898119894119889119904

120582119902119903= 119871119903119894119902119903+ 119871119898119894119902119904

(5)









119904) and



V

I

V

V

I

I

120596r

120596r

Is and IrIgc


DroopXs

Varmeasurement


Powermeasurement

Powerloosers

Vac

Vref

Qref

VX119904

Q

Pref

P

P1

+

+

+

+

+

minusminus

minus

minusminus

minus

minus

Ir

AC voltageregulator

Varregulator

Currentmeasurement

Powerregulator

Currentregulator

Vr

(Vdr Vqr)

Iqr_ref

Iqr

Idr

Idr_ref











119889119903 119881119889119903

and 119881119902119903



119868119903 ref = 119868

119889119903 ref + 119868119902119903 ref (6)

Vdc

Vdc_ref

Igc

+

+

+

minus

minus

minus

Idgc

Iqgc

Idgc_ref

Iq_ref

Vgc

DC voltageregulator

Currentmeasurement

Currentregulator










EndStop







119868gc ref = 119868119889gc ref + 119868

119902119903 ref (7)













x1

x2

xn

f(x1)

f(x2)

h1

h0h2

hn

u(k) y(k)Σ

f(xn)





119909 (119896) = (1199091 (119896) 1199092 (

119896) 119909119899 (119896)) (8)



119910 (119896) = 119891(

119899

sum

119894=1

119891 (119909119894 (119896))) = 119891 (119906 (119896))

=

ℎ2

sum

119902=0

1205931199020(

119899

sum

119894=1

ℎ1

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896))1199081198950

119910 (119896) =

ℎ2

sum

119902=0

12059320(119906 (119896)1199081199020 (

119896))

(9)




119864 (119896) =

1

2

(119889 (119896) minus 119910 (119896))2=

1

2

1198902(119896) (10)


1199081198950 (

119896 + 1) = 1205781199081198950 (

119896) + 1205950 (119896) 119890 (119896) 1205931198950 (

119906 (119896)) (11)


1199080 (119896 + 1) = 120578119908

0 (119896) + 120595

0 (119896) 119890 (119896) 1205930 (

119906 (119896)) (12)

12059311

12059321

120593k1

12059312

12059322

120593k2

1205931n

1205932n

120593kn

12059310

12059320

120593k0

w11

w21

wh1

w12

w22

wh2

w1n

w2n

whn

w10

w00

w0n

w02

w01

w20

wh0

Σ

Σ

Σ

Σ Σ

x1

x2

xn

f(x1)

f(x2)

f(xn)

u y




1199080 (119896 + 1) = 119908

0 (119896) +

119890 (119896) 1205930 (119906 (119896))

1205741199080

119894(119896)

1205741199080

119894(119896 + 1) = 120572120574

1199080

119894(119896) +

10038171003817100381710038171205930 (119906 (119896 + 1))

1003817100381710038171003817

2

(13)


119864 (119896) =

1

2

(119889 (119896) minus 1198910 (119906 (119896)))

2

=

1

2

(119889 (119896) minus 1198910(

119899

sum

119894=1

ℎ2

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896)))

2

(14)


119908119895119894 (119896 + 1) = 120578119908

119895119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119895119894

(119909119894 (119896))

(15)


119908119894 (119896 + 1) = 120578119908

119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

(16)



StartPhase I







Stop



119908119894 (119896 + 1) = 119908

119894 (119896) +

119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

120574119908119894

119894(119896)

120574119908119894

119894(119896 + 1) = 120572120574

119908119894

119894(119896) +

1003817100381710038171003817120593119894(119909119894 (119896 + 1))

1003817100381710038171003817

2

(17)





Fault detection

dqabc PWM



Vdr

+

minus

Ncrf

VS_ref

VS

XIDWNNFRT

V998400qr

ilowastr

Vlowastdc








lowast



dc and 119894lowast

119903act as the


119881lowast


119894lowast

119903=

119894119903minus 119894ss


(18)






min119891 = [


]

2

+ [


]

2

(19)





05

0

minus05

1

05

0 002

0406

081


IDW

NN

out

putN

crf




lowast




AC grid

Load 1

Line 1

Load 3


Load 2

WT

PCC

Line 3 Line 4

Line 2

Fault

20kV400V

20kV690V











145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)


1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2


14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)


14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)


145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)


14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02


14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)



0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)



14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in






120596119904

= minus119904119879119898120596119904= minus119904119875

119904

(3)






V119889119904= 119877119904119894119889119904minus 120596119904120582119902119904+

119889120582119889119904

119889119905

V119902119904= 119877119904119894119902119904+ 120596119904120582119889119904+

119889120582119902119904

119889119905

V119889119903= 119877119903119894119889119903minus (119904120596119904) 120582119902119903+

119889120582119889119903

119889119905

V119902119903= 119877119903119894119902119903+ (119904120596119904) 120582119889119903+

119889120582119902119903

119889119905

(4)




119904and119877

119903

and 120582119904and 120582



120582119889119904= minus119871119904119894119889119904+ 119871119898119894119889119903

120582119902119904= minus119871119904119894119902119904+ 119871119898119894119902119903




A

B

C

D

5ms

12ms

162ms16

14

12

1

08

06

04

02

0

06 07 08 09 1 11 12 13Turb

ine o

utpu

t pow

er (p

u of

nom

inal

mec

hani

cal p

ower

)



120582119889119903= 119871119903119894119889119903+ 119871119898119894119889119904

120582119902119903= 119871119903119894119902119903+ 119871119898119894119902119904

(5)









119904) and



V

I

V

V

I

I

120596r

120596r

Is and IrIgc


DroopXs

Varmeasurement


Powermeasurement

Powerloosers

Vac

Vref

Qref

VX119904

Q

Pref

P

P1

+

+

+

+

+

minusminus

minus

minusminus

minus

minus

Ir

AC voltageregulator

Varregulator

Currentmeasurement

Powerregulator

Currentregulator

Vr

(Vdr Vqr)

Iqr_ref

Iqr

Idr

Idr_ref











119889119903 119881119889119903

and 119881119902119903



119868119903 ref = 119868

119889119903 ref + 119868119902119903 ref (6)

Vdc

Vdc_ref

Igc

+

+

+

minus

minus

minus

Idgc

Iqgc

Idgc_ref

Iq_ref

Vgc

DC voltageregulator

Currentmeasurement

Currentregulator










EndStop







119868gc ref = 119868119889gc ref + 119868

119902119903 ref (7)













x1

x2

xn

f(x1)

f(x2)

h1

h0h2

hn

u(k) y(k)Σ

f(xn)





119909 (119896) = (1199091 (119896) 1199092 (

119896) 119909119899 (119896)) (8)



119910 (119896) = 119891(

119899

sum

119894=1

119891 (119909119894 (119896))) = 119891 (119906 (119896))

=

ℎ2

sum

119902=0

1205931199020(

119899

sum

119894=1

ℎ1

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896))1199081198950

119910 (119896) =

ℎ2

sum

119902=0

12059320(119906 (119896)1199081199020 (

119896))

(9)




119864 (119896) =

1

2

(119889 (119896) minus 119910 (119896))2=

1

2

1198902(119896) (10)


1199081198950 (

119896 + 1) = 1205781199081198950 (

119896) + 1205950 (119896) 119890 (119896) 1205931198950 (

119906 (119896)) (11)


1199080 (119896 + 1) = 120578119908

0 (119896) + 120595

0 (119896) 119890 (119896) 1205930 (

119906 (119896)) (12)

12059311

12059321

120593k1

12059312

12059322

120593k2

1205931n

1205932n

120593kn

12059310

12059320

120593k0

w11

w21

wh1

w12

w22

wh2

w1n

w2n

whn

w10

w00

w0n

w02

w01

w20

wh0

Σ

Σ

Σ

Σ Σ

x1

x2

xn

f(x1)

f(x2)

f(xn)

u y




1199080 (119896 + 1) = 119908

0 (119896) +

119890 (119896) 1205930 (119906 (119896))

1205741199080

119894(119896)

1205741199080

119894(119896 + 1) = 120572120574

1199080

119894(119896) +

10038171003817100381710038171205930 (119906 (119896 + 1))

1003817100381710038171003817

2

(13)


119864 (119896) =

1

2

(119889 (119896) minus 1198910 (119906 (119896)))

2

=

1

2

(119889 (119896) minus 1198910(

119899

sum

119894=1

ℎ2

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896)))

2

(14)


119908119895119894 (119896 + 1) = 120578119908

119895119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119895119894

(119909119894 (119896))

(15)


119908119894 (119896 + 1) = 120578119908

119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

(16)



StartPhase I







Stop



119908119894 (119896 + 1) = 119908

119894 (119896) +

119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

120574119908119894

119894(119896)

120574119908119894

119894(119896 + 1) = 120572120574

119908119894

119894(119896) +

1003817100381710038171003817120593119894(119909119894 (119896 + 1))

1003817100381710038171003817

2

(17)





Fault detection

dqabc PWM



Vdr

+

minus

Ncrf

VS_ref

VS

XIDWNNFRT

V998400qr

ilowastr

Vlowastdc








lowast



dc and 119894lowast

119903act as the


119881lowast


119894lowast

119903=

119894119903minus 119894ss


(18)






min119891 = [


]

2

+ [


]

2

(19)





05

0

minus05

1

05

0 002

0406

081


IDW

NN

out

putN

crf




lowast




AC grid

Load 1

Line 1

Load 3


Load 2

WT

PCC

Line 3 Line 4

Line 2

Fault

20kV400V

20kV690V











145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)


1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2


14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)


14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)


145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)


14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02


14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)



0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)



14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




V

I

V

V

I

I

120596r

120596r

Is and IrIgc


DroopXs

Varmeasurement


Powermeasurement

Powerloosers

Vac

Vref

Qref

VX119904

Q

Pref

P

P1

+

+

+

+

+

minusminus

minus

minusminus

minus

minus

Ir

AC voltageregulator

Varregulator

Currentmeasurement

Powerregulator

Currentregulator

Vr

(Vdr Vqr)

Iqr_ref

Iqr

Idr

Idr_ref











119889119903 119881119889119903

and 119881119902119903



119868119903 ref = 119868

119889119903 ref + 119868119902119903 ref (6)

Vdc

Vdc_ref

Igc

+

+

+

minus

minus

minus

Idgc

Iqgc

Idgc_ref

Iq_ref

Vgc

DC voltageregulator

Currentmeasurement

Currentregulator










EndStop







119868gc ref = 119868119889gc ref + 119868

119902119903 ref (7)













x1

x2

xn

f(x1)

f(x2)

h1

h0h2

hn

u(k) y(k)Σ

f(xn)





119909 (119896) = (1199091 (119896) 1199092 (

119896) 119909119899 (119896)) (8)



119910 (119896) = 119891(

119899

sum

119894=1

119891 (119909119894 (119896))) = 119891 (119906 (119896))

=

ℎ2

sum

119902=0

1205931199020(

119899

sum

119894=1

ℎ1

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896))1199081198950

119910 (119896) =

ℎ2

sum

119902=0

12059320(119906 (119896)1199081199020 (

119896))

(9)




119864 (119896) =

1

2

(119889 (119896) minus 119910 (119896))2=

1

2

1198902(119896) (10)


1199081198950 (

119896 + 1) = 1205781199081198950 (

119896) + 1205950 (119896) 119890 (119896) 1205931198950 (

119906 (119896)) (11)


1199080 (119896 + 1) = 120578119908

0 (119896) + 120595

0 (119896) 119890 (119896) 1205930 (

119906 (119896)) (12)

12059311

12059321

120593k1

12059312

12059322

120593k2

1205931n

1205932n

120593kn

12059310

12059320

120593k0

w11

w21

wh1

w12

w22

wh2

w1n

w2n

whn

w10

w00

w0n

w02

w01

w20

wh0

Σ

Σ

Σ

Σ Σ

x1

x2

xn

f(x1)

f(x2)

f(xn)

u y




1199080 (119896 + 1) = 119908

0 (119896) +

119890 (119896) 1205930 (119906 (119896))

1205741199080

119894(119896)

1205741199080

119894(119896 + 1) = 120572120574

1199080

119894(119896) +

10038171003817100381710038171205930 (119906 (119896 + 1))

1003817100381710038171003817

2

(13)


119864 (119896) =

1

2

(119889 (119896) minus 1198910 (119906 (119896)))

2

=

1

2

(119889 (119896) minus 1198910(

119899

sum

119894=1

ℎ2

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896)))

2

(14)


119908119895119894 (119896 + 1) = 120578119908

119895119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119895119894

(119909119894 (119896))

(15)


119908119894 (119896 + 1) = 120578119908

119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

(16)



StartPhase I







Stop



119908119894 (119896 + 1) = 119908

119894 (119896) +

119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

120574119908119894

119894(119896)

120574119908119894

119894(119896 + 1) = 120572120574

119908119894

119894(119896) +

1003817100381710038171003817120593119894(119909119894 (119896 + 1))

1003817100381710038171003817

2

(17)





Fault detection

dqabc PWM



Vdr

+

minus

Ncrf

VS_ref

VS

XIDWNNFRT

V998400qr

ilowastr

Vlowastdc








lowast



dc and 119894lowast

119903act as the


119881lowast


119894lowast

119903=

119894119903minus 119894ss


(18)






min119891 = [


]

2

+ [


]

2

(19)





05

0

minus05

1

05

0 002

0406

081


IDW

NN

out

putN

crf




lowast




AC grid

Load 1

Line 1

Load 3


Load 2

WT

PCC

Line 3 Line 4

Line 2

Fault

20kV400V

20kV690V











145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)


1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2


14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)


14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)


145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)


14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02


14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)



0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)



14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in






EndStop







119868gc ref = 119868119889gc ref + 119868

119902119903 ref (7)













x1

x2

xn

f(x1)

f(x2)

h1

h0h2

hn

u(k) y(k)Σ

f(xn)





119909 (119896) = (1199091 (119896) 1199092 (

119896) 119909119899 (119896)) (8)



119910 (119896) = 119891(

119899

sum

119894=1

119891 (119909119894 (119896))) = 119891 (119906 (119896))

=

ℎ2

sum

119902=0

1205931199020(

119899

sum

119894=1

ℎ1

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896))1199081198950

119910 (119896) =

ℎ2

sum

119902=0

12059320(119906 (119896)1199081199020 (

119896))

(9)




119864 (119896) =

1

2

(119889 (119896) minus 119910 (119896))2=

1

2

1198902(119896) (10)


1199081198950 (

119896 + 1) = 1205781199081198950 (

119896) + 1205950 (119896) 119890 (119896) 1205931198950 (

119906 (119896)) (11)


1199080 (119896 + 1) = 120578119908

0 (119896) + 120595

0 (119896) 119890 (119896) 1205930 (

119906 (119896)) (12)

12059311

12059321

120593k1

12059312

12059322

120593k2

1205931n

1205932n

120593kn

12059310

12059320

120593k0

w11

w21

wh1

w12

w22

wh2

w1n

w2n

whn

w10

w00

w0n

w02

w01

w20

wh0

Σ

Σ

Σ

Σ Σ

x1

x2

xn

f(x1)

f(x2)

f(xn)

u y




1199080 (119896 + 1) = 119908

0 (119896) +

119890 (119896) 1205930 (119906 (119896))

1205741199080

119894(119896)

1205741199080

119894(119896 + 1) = 120572120574

1199080

119894(119896) +

10038171003817100381710038171205930 (119906 (119896 + 1))

1003817100381710038171003817

2

(13)


119864 (119896) =

1

2

(119889 (119896) minus 1198910 (119906 (119896)))

2

=

1

2

(119889 (119896) minus 1198910(

119899

sum

119894=1

ℎ2

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896)))

2

(14)


119908119895119894 (119896 + 1) = 120578119908

119895119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119895119894

(119909119894 (119896))

(15)


119908119894 (119896 + 1) = 120578119908

119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

(16)



StartPhase I







Stop



119908119894 (119896 + 1) = 119908

119894 (119896) +

119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

120574119908119894

119894(119896)

120574119908119894

119894(119896 + 1) = 120572120574

119908119894

119894(119896) +

1003817100381710038171003817120593119894(119909119894 (119896 + 1))

1003817100381710038171003817

2

(17)





Fault detection

dqabc PWM



Vdr

+

minus

Ncrf

VS_ref

VS

XIDWNNFRT

V998400qr

ilowastr

Vlowastdc








lowast



dc and 119894lowast

119903act as the


119881lowast


119894lowast

119903=

119894119903minus 119894ss


(18)






min119891 = [


]

2

+ [


]

2

(19)





05

0

minus05

1

05

0 002

0406

081


IDW

NN

out

putN

crf




lowast




AC grid

Load 1

Line 1

Load 3


Load 2

WT

PCC

Line 3 Line 4

Line 2

Fault

20kV400V

20kV690V











145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)


1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2


14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)


14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)


145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)


14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02


14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)



0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)



14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




x1

x2

xn

f(x1)

f(x2)

h1

h0h2

hn

u(k) y(k)Σ

f(xn)





119909 (119896) = (1199091 (119896) 1199092 (

119896) 119909119899 (119896)) (8)



119910 (119896) = 119891(

119899

sum

119894=1

119891 (119909119894 (119896))) = 119891 (119906 (119896))

=

ℎ2

sum

119902=0

1205931199020(

119899

sum

119894=1

ℎ1

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896))1199081198950

119910 (119896) =

ℎ2

sum

119902=0

12059320(119906 (119896)1199081199020 (

119896))

(9)




119864 (119896) =

1

2

(119889 (119896) minus 119910 (119896))2=

1

2

1198902(119896) (10)


1199081198950 (

119896 + 1) = 1205781199081198950 (

119896) + 1205950 (119896) 119890 (119896) 1205931198950 (

119906 (119896)) (11)


1199080 (119896 + 1) = 120578119908

0 (119896) + 120595

0 (119896) 119890 (119896) 1205930 (

119906 (119896)) (12)

12059311

12059321

120593k1

12059312

12059322

120593k2

1205931n

1205932n

120593kn

12059310

12059320

120593k0

w11

w21

wh1

w12

w22

wh2

w1n

w2n

whn

w10

w00

w0n

w02

w01

w20

wh0

Σ

Σ

Σ

Σ Σ

x1

x2

xn

f(x1)

f(x2)

f(xn)

u y




1199080 (119896 + 1) = 119908

0 (119896) +

119890 (119896) 1205930 (119906 (119896))

1205741199080

119894(119896)

1205741199080

119894(119896 + 1) = 120572120574

1199080

119894(119896) +

10038171003817100381710038171205930 (119906 (119896 + 1))

1003817100381710038171003817

2

(13)


119864 (119896) =

1

2

(119889 (119896) minus 1198910 (119906 (119896)))

2

=

1

2

(119889 (119896) minus 1198910(

119899

sum

119894=1

ℎ2

sum

119895=0

120593119895119894(119909119894 (119896)) 119908119895119894 (

119896)))

2

(14)


119908119895119894 (119896 + 1) = 120578119908

119895119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119895119894

(119909119894 (119896))

(15)


119908119894 (119896 + 1) = 120578119908

119894 (119896)

+ 120595 (119896) 119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

(16)



StartPhase I







Stop



119908119894 (119896 + 1) = 119908

119894 (119896) +

119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

120574119908119894

119894(119896)

120574119908119894

119894(119896 + 1) = 120572120574

119908119894

119894(119896) +

1003817100381710038171003817120593119894(119909119894 (119896 + 1))

1003817100381710038171003817

2

(17)





Fault detection

dqabc PWM



Vdr

+

minus

Ncrf

VS_ref

VS

XIDWNNFRT

V998400qr

ilowastr

Vlowastdc








lowast



dc and 119894lowast

119903act as the


119881lowast


119894lowast

119903=

119894119903minus 119894ss


(18)






min119891 = [


]

2

+ [


]

2

(19)





05

0

minus05

1

05

0 002

0406

081


IDW

NN

out

putN

crf




lowast




AC grid

Load 1

Line 1

Load 3


Load 2

WT

PCC

Line 3 Line 4

Line 2

Fault

20kV400V

20kV690V











145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)


1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2


14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)


14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)


145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)


14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02


14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)



0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)



14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




StartPhase I







Stop



119908119894 (119896 + 1) = 119908

119894 (119896) +

119890 (119896) 1198911015840

0(119906 (119896)) 120593119894

(119909119894 (119896))

120574119908119894

119894(119896)

120574119908119894

119894(119896 + 1) = 120572120574

119908119894

119894(119896) +

1003817100381710038171003817120593119894(119909119894 (119896 + 1))

1003817100381710038171003817

2

(17)





Fault detection

dqabc PWM



Vdr

+

minus

Ncrf

VS_ref

VS

XIDWNNFRT

V998400qr

ilowastr

Vlowastdc








lowast



dc and 119894lowast

119903act as the


119881lowast


119894lowast

119903=

119894119903minus 119894ss


(18)






min119891 = [


]

2

+ [


]

2

(19)





05

0

minus05

1

05

0 002

0406

081


IDW

NN

out

putN

crf




lowast




AC grid

Load 1

Line 1

Load 3


Load 2

WT

PCC

Line 3 Line 4

Line 2

Fault

20kV400V

20kV690V











145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)


1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2


14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)


14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)


145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)


14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02


14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)



0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)



14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in







lowast



dc and 119894lowast

119903act as the


119881lowast


119894lowast

119903=

119894119903minus 119894ss


(18)






min119891 = [


]

2

+ [


]

2

(19)





05

0

minus05

1

05

0 002

0406

081


IDW

NN

out

putN

crf




lowast




AC grid

Load 1

Line 1

Load 3


Load 2

WT

PCC

Line 3 Line 4

Line 2

Fault

20kV400V

20kV690V











145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)


1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2


14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)


14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)


145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)


14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02


14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)



0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)



14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




AC grid

Load 1

Line 1

Load 3


Load 2

WT

PCC

Line 3 Line 4

Line 2

Fault

20kV400V

20kV690V











145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)


1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2


14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)


14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)


145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)


14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02


14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)



0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)



14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




145 15 155 16 1650506070809

1111213

Time (s)

Thre

e-ph

ase v

olta

ge (p

u)


1351414515155

0

2

4

6

8

10

Time (s)

DC-

link

volta

ge (V

)

minus2


14 142 144 146 148 15 152 154 156 158 16minus2minus1

0123456

Time (s)

Roto

r cur

rent

(kA

)


14 145 15 155 16 165 170

02040608

112141618

2

Time (s)

Stat

or cu

rren

t (kA

)


145 15 155 16 165 17 175 18minus2

minus1

0

1

2

3

4

5

Time (s)

WT

outp

ut ac

tive p

ower

(MW

)


14 15 16 17 18 19 20

002040608

11214

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)

minus02


14 145 15 155 16 165 17minus02

002040608

11214

Time (s)

Roto

r spe

ed (p

u)



0123

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)



14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




14 145 15 155 16 165 170506070809

111121314

Time (s)

Thre

e-ph

ase s

tato

r vol

tage

(pu)


146 148 15 152 154 156 158 16minus05

005

115

225

335

445

5

Time (s)

Roto

r cur

rent

(kA

)





02468

101214

Time (s)

Stat

or cu

rren

t (kA

)


146 148 15 152 154 156 158 16

070809

11112131415

Time (s)

WT

outp

ut ac

tive p

ower

(mW

)





13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




13 135 14 145 15 155 16 165 17 175 18minus05

005

115

225

335

445

5

Time (s)

WT

outp

ut re

activ

e pow

er (m

Var)


minus2 0 2 4 6 8 10minus05

0

05

1

15

2

25

3

Time (s)

Roto

r spe

ed (p

u)





14 145 15 155 16 165 17 175 18minus2minus1

012345678

Time (s)

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)


10 11 12 13 14 15 16 17 18 19 20minus1

minus05

0

05

1

15

2

Time (s)

Mod

ified

q-c

ompo

nent

of t

hero

tor v

olta

ge (p

u)





6 Conclusion



10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




10 20 30 40 50 60 70 80 90 1004

455

556

657

758

859


Com

pute

d fit

ness

val

ues





References











































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in





















Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in










Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in



Documents

Research Article Differential Evolution Based IDWNN ...downloads.hindawi.com/journals/tswj/2015/746017.pdfResearch Article Differential Evolution Based IDWNN Controller for Fault Ride-Through