Adaptive voltage control for large scale solar PV power plant ...

1

Adaptive Voltage Control for Large Scale Solar PVPower Plant Considering Real Life Factors

Hazem Karbouj, Zakir Hussain Rather, Member, IEEE and Bikash C. Pal, Fellow, IEEE

Abstract—This paper presents an accurate and realistic es-timation of reactive power capability of solar photovoltaic (PV)inverters considering ambient temperature, solar irradiance, andterminal voltage. Based on the accurate estimation of reactivepower capability, a self-adaptive voltage controller is proposedto enable solar PV power plant participation in voltage controlancillary service. The proposed accurate and realistic estimationhas revealed the possibility of solar PV power plant failing tocomply with grid code requirements under extreme weatherconditions. On the other hand, the proposed control strategyhas shown significantly better effectiveness to utilise solar PVinverter capability, and provide better voltage control supportservice to the grid.

Index Terms—Ancillary service, reactive power capability,large scale solar PV power plant.

I. INTRODUCTION

Grid integration of solar photovoltaic (PV) power has re-cently experienced a significant shift from rooftop installationto large utility scale solar PV power plant, which can beattributed to several factors, such as falling prices of solarPV technology, regulatory and policy measures for largescale renewable energy integration. GW-scale solar PV powerplants (PVPP) are currently under construction or already inoperation in various places around the world. In India, forexample, 2 GW large scale PVPP is partly in operation, withtwo other PVPP of 2.25 and 5 GW under construction [1].

Due to displacement of conventional generation by largescale PVPP and other renewable energy sources (RESs),transmission system operators (TSOs) encounter various chal-lenges in secure and stable operation of RES integrated powersystems, particularly, at high RES penetration level. RESdriven displacement of synchronous machines, the main sourceof dynamic reactive power in conventional power systems,result in diminished reactive power reserve and poor dynamicvoltage control in the power system. Several countries arealready facing diminishing reactive power issue due to highRES penetration, and these countries are tackling such issuesthrough several measures, such as, installing new synchronouscondensers [2], and procuring reactive power from ancillaryservice market [3]. Therefore, in large-scale RES integratedpower system, renewable energy power plants, including largescale PVPP, are required to provide dynamic reactive powerand voltage control support for secure and stable grid opera-tion. In large-scale solar PV power integrated systems, largescale PVPP are expected to take leading role in grid voltage

H. Karbouj and Z. H. Rather are with the Department of Energy Scienceand Engineering, Indian Institute of Technology Bombay, Mumbai 400076,India (e-mail: {hazem.karbouj & zakir.rather}@iitb.ac.in). B. C. Pal is withthe Department of Electrical and Electronic Engineering, Imperial CollegeLondon, London SW7 2AZ, U.K (e-mail: [email protected]).

control beyond current mandatory grid code requirements,potentially through procurement of such service in ancillaryservice market or by mandating such services from large scalePVPP.

Solar PV inverters can actively participate in reactive powersupport in daylight and night-time [4], [5]. Currently rooftopsolar PV inverters are not required to participate in voltagecontrol/dynamic reactive power control. However, large scalePVPP, typically connected at transmission voltage level, arerequired to participate in dynamic voltage control [6]. Theauthors in [7] presented a field implementation of 300 MWsolar PV power plant participating in dynamic reactive powersupport, and plant’s compliance with grid code regulations.The authors in [8] and [9] presented control strategies toutilize solar PV inverters as reactive power static compensators(STATCOM) during day and night-time. Moreover, the authorsin [10] and [11] analysed the cost of operating a solar PVpower plant at night as STATCOM. However, while therelated studies reported in the literature so far have focusedon reactive power support from solar PV inverters/plants, allthe studies have ignored accurate calculation or estimationof reactive power (Q) capability of such inverter interfacedsolar PV systems, as several practical factors, particularly,weather conditions, can significantly affect actual reactivepower capability of such plants.

This paper addresses the above stated practical limitationsthrough two main contributions, viz.: i) A methodology foraccurate calculation of solar PV inverter reactive power ca-pability considering ambient temperature, solar irradiance andinverter terminal voltage has been proposed to find realisticreactive power support capability of solar PV plant, and ii)Based on the accurate estimation of reactive power capability,a self-adaptive voltage control is proposed to enable largescale PVPP participation in voltage control ancillary service.Accurate reactive power capability of solar PV inverters iscalculated under extreme weather conditions for two locationsin Saudi Arabia and India. The proposed Q capability calcula-tion is compared with the existing conventional approach thatdoes not consider effect of ambient temperature and inverterterminal voltage on reactive power capability of solar PV in-verters. The proposed methodology of Q capability calculationand self-adaptive voltage control scheme has been validatedon a PVPP integrated IEEE 39 bus system developed inDIgSILENT PowerFactory platform, with detailed modellingof PVPP.

The rest of the paper is organised as follows, in SectionII mandatory reactive power requirements from PVPP isdiscussed for different grid code regulations. Accurate reactivepower capability of solar PV inverter is formulated in Section

Adaptive voltage control for large scale solar PV power plant considering real life factors

This is a peer-reviewed, accepted author manuscript of the following article: Karbouj, H., Rather, Z., & Pal, B. C. (2021). Adaptive voltage control for large scale solar PV power plant considering real life factors. IEEE Transactions on Sustainable Energy, 12(2), 990-998. https://doi.org/10.1109/TSTE.2020.3029102

2

V

Q/Pmax

1.05

1.10

Q=0.52, V=1.118

0.95

Q=0.52, V=0.9Q=-0.5, V=0.9

0.9

-0.25-0.5

Q=-0.5, V=1.118

V

Q/Pmax

1.05

1.10

0.95

-0.25-0.50.25 0.250.5

Q=-0.328, V=1.1

Q=0.328, V=0.90.9

0.5

Ireland(EirGrid) USA (CAISO)

V

Q/Pmax

1.05

1.10

0.95

-0.25-0.5 0.25

Q=-0.3, V=1.068

Q=0.3, V=0.930.9

0.5

Spain(REE)

V

Q/Pmax

1.05

1.10

0.95

-0.25-0.5 0.25

Q=-0.328, V=1.15

Q=0.41, V=0.8770.9

0.5

Germany(TenneT)

Q=0.41, V=1.113Q=-0.328, V=1.123

Q=0.328, V=0.877

Q=0.328, V=1.113

Fig. 1: Reactive power requirements from large scale PVPP for different countries under different voltage magnitudes. Irishrequirements are for 110 kV and 220 kV grids. Spanish and German requirements are for 220 kV grid. [7], [12]–[14]

III. The proposed self-adaptive voltage controller is presentedin Section IV. Results are presented in Section V. The paperis discussed and concluded in Section VI.

II. MANDATED REACTIVE POWER REQUIREMENTS FROMLARGE SCALE PVPP

A. Mandated Reactive Power Requirements from large scalePVPP

System regulators/TSOs require solar PV inverters to pro-vide minimum reactive power support based on the steady stateoperating condition. Further, solar PV inverters are required tofollow specific control rule within a given capability range asdescribed in Section II-B.

Minimum Solar PV inverter reactive power capability re-quired in grid code regulation depends on the rating of thepower plant, and point of common coupling (PCC) voltagelevel. Moreover, minimum reactive power requirements varyfrom system to system depending on the system strength andcharacteristics.

Fig. 1 shows mandated reactive power requirements fromPVPP in four different TSO jurisdictions, where Pmax refersto the maximum active power generation of the solar PVarray/plant. It can be observed that the strictest requirementof reactive power operating range is in Ireland where solarPV inverters at transmission voltage level are required tohave a mandatory reactive power capability of more thanhalf of the installed solar PV array power [13]. CaliforniaIndependent System Operator (CAISO), Spanish and Germangrid operators have relatively less strict mandatory reactivepower requirement. In this study, Germany’s minimum reactivepower requirement is considered to be mandatory reactivepower reserve Qmand.

B. Reactive Power Control Modes of Large Scale PVPP

Decoupled control of power inverter has enabled the inverterto operate in different reactive power control modes withoutcompromising maximum power point tracking (MPPT) opera-tion. Solar PV inverters are generally controlled in one of threemodes, reactive power, power factor or voltage control mode.In reactive power control mode, inverter exchange reactivepower with the grid based on the reference value of Q set bythe plant operator. In power factor control, on the other hand,reactive power is exchanged by the inverter to maintain thedesired power factor at the inverter terminal. Reactive power

or power factor modes are primarily used in solar PV invertersconnected to distribution system. However, voltage controlmode, primarily based on Q-voltage (Q-V) droop, is generallyadopted in inverters connected to transmission network. It isimportant to note that in all the three control (Q, power factor,voltage) modes, inverter reactive power is controlled withinthe mandated reactive power range described in Section II-A.As large scale PVPPs are typically connected to transmissionsystems, voltage control mode is followed in this paper.

In voltage control mode, Q-V droop slope is set as requiredby TSO to regulate the grid voltage within acceptable range,[0.9-1.1] pu. Reactive power support from PVPP beyondmandatory requirement (described in Section II-A), whichmight be required to bring voltage to its set point is consideredas ancillary service [13], [15].

Solar PV inverters are designed based on peak power ratingof solar PV arrays, however, for major portion of the daysolar irradiance is less than the maximum designed irradiance(1000 W/m2). Therefore, due to high margin for reactivepower support, PVPPs have great potential to participate involtage control ancillary service, thus potentially allowing theplant operator to earn additional revenue from Q ancillaryservice market. On the other hand, from TSO prospective, par-ticipation of large scale PVPP would help securing the systemvoltage without the need of installing additional reactive powercompensation devices in the grid.

III. PROPOSED METHOD FOR ACCURATE REACTIVEPOWER CAPABILITY OF A SINGLE SOLAR PV INVERTER

In order to estimate reactive power reserve from solar PVplant, reactive power capability curves of individual solar PVinverters should be known. While there has been reasonablefocus on finding reactive power capability of wind power plant[16]–[18], calculation of reactive power capability of solar PVpower plant has received a limited attention so far.

Typical solar PV inverter system is comprised of severalsolar PV modules connected in series and parallel to builddesired voltage and current, which in turn is interfaced to thegrid through inverter and a step-up transformer, as shown inFig. 2. For realistic estimation of reactive power capability ofsolar PV plant, it is important to consider critical factors incalculation of individual solar PV inverter Q capability. Suchfactors that affect the reactive power capability of solar PVinverter can be classified into three categories, as describedbelow.


3

Inverter control and protection

Local measurement

Plant control commands

Pulses

Ns×Np

Grid

Fig. 2: Solar PV inverter system

0 500 1000 1500VPV(V)

0

0.2

0.4

0.6

0.8

1

P/P m

ax

1000

W/m

2

800 W/m

2

600 W/m2

400 W/m2

200 W/m2

25oC

35oC

50oC

Fig. 3: Solar PV array performance under different irradianceand temperature

A. Impact of solar irradiance and ambient temperature onsolar PV array output

Active power produced by solar PV arrays is highly de-pendent on solar irradiance and ambient temperature. Mathe-matical equations that govern active power produced by solarPV array are given in [19]. Fig. 3 shows performance of asolar PV array (specifications provided in the Appendix) underdifferent solar irradiance and ambient temperature. It can beobserved from Fig. 3, that active power output of solar arrays,is highly sensitive to solar irradiance and ambient temperature,with reduced capability at higher ambient temperature. Con-sequently, as active power delivered to grid by the inverterchanges with such factors, reactive power reserve/margin ofthe inverter changes accordingly as given in (1).

Qres(G,Ta) =√S2r − P 2(G,Ta) (1)

Where Sr is the rated apparent power of inverter, P (G,Ta)is the active power produced by solar array connected tothe inverter as a function of ambient temperature (Ta) andirradiance (G).

B. Impact of ambient temperature on solar PV inverter per-formance

Ambient temperature directly influences inverter current car-rying capability, as increased ambient temperature decreasesheat dissipation from power electronic switches to ambient

TABLE I: Specifications of solar PV inverters

# Manufacturer S1

(MVA)S2

(MVA)S3

(MVA)T0

(◦C)T1

(◦C)T2

(◦C)T3

(◦C)

1 SMA [23] 4.6 3.91 0 -25 25 50 602 ABB [24] 4.4 4 - -20 35 50 -3 SMA [23] 4 3.4 0 -25 25 50 604 SMA [25] 3 2.7 0 -25 35 50 605 ABB [26] 2.4 2 - -20 20 45 -6 SMA [25] 2.2 2 0 -25 35 50 607 SMA [27] 2.2 2.06 - -25 40 50 -

environment, which in turns lead to increased junction tem-perature of power electronic switches [20]. Therefore, forsafe operation of the device/switches under high ambienttemperature, additional remedial actions are required.

To avoid inverter component overheating due to increasedambient temperature, solar PV inverter manufacturers advisederated operation of the inverter, thus maximum power that canbe delivered by the inverter under high ambient temperature[21], [22]. While the maximum temperature has nonlineardependence on ambient temperature [20], solar PV invertermanufacturers tend to use a linear derating characteristic [21],[22], as given by (2).

Srt(Ta) =

Sr = S1, if T0 ≤ Ta ≤ T1

Sr × S1−S2

T1−T2, if T1 < Ta ≤ T2

Sr × S2−S3

T2−T3, if T2 < Ta ≤ T3

0, if T3 < Ta

(2)

Where Srt is the derated power of the inverter, T0...T3 andS1...S3 are ambient temperature in ◦C and dearted power inMVA, respectively, with their typical values given in Table I. Itcan be observed from Table I that inverter derating starts whenambient temperature exceeds particular threshold, mostly inthe range of (25-35 ◦C). Further, beyond ambient temperatureof around 50 ◦C, sharp decrease in power carrying capacity isrequired, before the inverter is tripped when ambient temper-ature exceeds 60 ◦C. While the inverter is disconnected whenambient temperature exceeds 60 ◦C.

Since maximum current carrying capacity of solar PVinverter decreases with the increase in ambient temperature,inverter reactive power capability is also significantly affected.Therefore, to calculate realistic reactive power capability ofinverter, it is important to take inverter derating into considera-tion, hence, inverter reactive power reserve (1) can be modifiedby replacing Sr with Srt(Ta) as given in (3).

Qres(G,Ta) =√S2rt(Ta) − P 2(G,Ta) (3)

Reactive power capability of the 4.6 MVA SMA inverterwith its parameters provided in the Appendix, for specificsolar irradiance and a varying ambient temperature is shownin Fig. 4. It can be observed from Fig. 4 that for a given activepower output from solar PV inverter, reactive power capabilityis significantly reduced due to increased ambient temperaturedriven derating of the inverter. It is, however, worthy to note


4

0 1 2 3 4 5Active power(MW)

-5

0

5R

eact

ive

pow

er(M

VA

r)

25

30

35

40

45

50

Am

bien

t tem

pera

ture

(°C

)

3.5 3.6 3.7-3

-2

-1

0

1

2

3

Thermal limits of inverter

Thermal limits of inverter

Thermal limitsof PV modules

Fig. 4: Solar PV system performance under different ambienttemperatures for G=1000 W/m2 and V=1 pu.

Fig. 5: Solar PV system Q capability under different ambienttemperatures and irradiance, V=1 pu.

that reduction of solar PV array power generation due tothe increased ambient temperature can to some extent offsetdecrease in inverter reactive power reserve, as can also beconcluded from (3).

Impact of ambient temperature on the SMA test inverterreactive power capability under different solar irradiance isalso investigated with results shown in Fig. 5 assuming solarPV inverter operates at maximum power point. It can beobserved from Fig. 5 that ambient temperature has a significantimpact on inverter reactive power capability, particularly, athigh irradiance level.

C. Impact of inverter terminals voltage on PV inverter reactivepower capability

Inverter terminal voltage can affect inverter reactive powercapability as the terminal voltage influences current magnitudethrough inverter switches for a given apparent power. Toincorporate inverter terminal voltage in Q capability equation,(3) can be modified as given in (4).

Qres(G,Ta, V ) =√S2rtv(Ta, V ) − P 2(G,Ta) (4)

0 1 2 3 4Active power(MW)

-6

-4

-2

0

2

4

6

Rea

ctiv

e po

wer

(MV

Ar)

25

30

35

40

45

50

Am

bien

t tem

pera

ture

(°C

)V=0.9 pu

V=1.1 pu

Fig. 6: Solar PV system performance under different ambienttemperatures for G=1000 W/m2 for V=0.9 pu and V=1.1 pu.

Fig. 7: Solar PV system Q capability under different ambienttemperatures and irradiance and voltages

Where Srtv = V × Srt, V is the solar PV inverter terminalsvoltage in pu, in the range of (0.9-1.1 pu).

Reactive power capability of the SMA test inverter underirradiance of 1000 W/m2, varying ambient temperature, andtwo terminal voltages, 0.9 pu and 1.1 pu has also been studied,as shown in Fig. 6. It can be observed that decrease in terminalvoltage at high ambient temperature results in decreasedreactive power capability of the inverter. On the other hand,higher terminal voltage can result in higher reactive powerreserve under the same conditions, which can be attributed tolower current required from the inverter for a given power.Fig. 7 shows reactive power capability curves under differentirradiance, ambient temperature, and inverter terminal voltage.It can be observed from Fig. 7 that inverter reactive powercapability is minimum at high ambient temperature, high solarirradiance and low terminal voltage.

IV. PROPOSED ADAPTIVE VOLTAGE CONTROL SCHEMEFOR PVPP ANCILLARY SERVICE SUPPORT

As described in Section II, large scale PVPP are connectedat transmission voltage level, and can support the grid voltage


5

V

Q

Qmand

Qmax=f(P,V,Ta)

kQ V=f(Qm ax )

Qmand

-Qmax=f(P,V,Ta)

-Vmax

Vmax

-Vdb

Vdb

`

kQ V=f(Qm ax)`

Fig. 8: Conventional fixed QV droop (bold line) and proposedadaptive QV droop (dashed lines)

by operating in voltage control mode. Fig. 8 shows conven-tional fixed Q-V droop characteristic (bold black line) used tocomply with mandatory reactive power control requirements,where Qmand and Vmax values are taken from grid coderegulation, Fig. 1.

Solar PV inverter Q capability, as discussed in Section III,is significantly influenced by variation in ambient temperature,solar irradiance, and PV inverter terminal voltage. The fixedQ-V droop can be expressed as given by (5).

Q = kQV ∗ ∆V (5)

Where kQV is fixed, predefined QV droop and ∆V is thevoltage deviation from the set point at the inverter terminal.

On the other hand, in practice, solar irradiance is not at peakfor most of the day, which allows inverter to supply/consumeadditional reactive power beyond mandatory requirements.Therefore, to utilise this untapped reactive power reserve, andsupport the grid beyond mandatory requirements, self-adaptiveQ-V droop control is proposed for large scale PVPP. In theproposed control scheme, QV droop adapts to varying reactivepower capability of the inverter as formulated in (4) as shownin Fig. 8. In other words, increase of reactive power capabilityresult in higher QV droop, and hence larger Q contributionfrom the solar PV inverter towards regulating its terminalsvoltage. The self-adaptive QV droop can be formulated asgiven in (6).

Q = k′QV ∗ ∆V (6)

Where k′QV = −Qmax(P, Ta, V )

Vmax − Vdband Qmax is calculated as

given in (4). Vdb is the deadband voltage, which is used toavoid undesirable control triggering, as shown in Fig. 8.

To implement the proposed self-adaptive Q-V droop, thecontrol scheme shown in Fig. 9 has been used, where max-imum reactive power capability is calculated based on mea-sured terminals voltage, active power and ambient temperature.It is worth to note that high power inverters are equipped withambient temperature measurement provision, which is usedfor inverter overheating protection [22], [23]. The maximumreactive power capability of the inverter calculated for a givenoperating conditions (irradiance, ambient temperature, and

Capability calculationVinv

PTa

Qmax,Qmin

Droop calculation

Q

V max

min

Qref

MPPTVPVIPV

Pcom

P

PI- Active power

control mode

Vdcref PI- Idrefmax

min

Q-

PIIqref

PWM InverterPulses

Fig. 9: Proposed adaptive QV droop control

terminal voltage) is then used for adaptive droop calculationand to limit maximum reactive power exchange of the inverter.Reactive power reference for the PV inverter is generatedfrom the adaptive droop block based on the measured terminalvoltage as given in (6).

Active power supplied by the inverter, on the other hand, iscontrolled based on MPPT reference, or set by the solar unitoperator command (Pcom). Commanded reactive and activepower from the inverter are compared with the measured val-ues and the error is passed through proportional-integral (PI)controllers to generate current commands. The dq componentsof current commands are then used to generate PWM pulsesto control the inverter switches.

V. RESULTS

The results, in line with the methodologies described inSection III and IV, are done in two stages. In the first stage, themethodology of calculating accurate reactive power capabilityof inverter is implemented, and participation of large scalePVPP in voltage control ancillary service under differentambient temperature, solar irradiance, and terminal voltage isevaluated. In the second stage, performance of the proposedself-adaptive Q-V droop control scheme for large scale PVPPunder various sets of disturbances is performed in time domainsimulation.

A. Impact of ambient temperature on solar PV inverter reac-tive power capability

To study the impact of high ambient temperature on solarPV inverter reactive power capability, 24 hour solar irradianceand ambient temperature data of two locations have been usedin a SMA solar PV inverter with its specification provided inthe Appendix.

The first location considered is in Al Qaseem, Saudi Arabia,whose solar irradiance and ambient temperature for 24 hoursare shown in Fig. 10(a) [28]. It can be observed from Fig.10(a) that ambient temperature reaches to approximately 45 ◦Cin afternoon hours, while solar irradiance touches 1000 W/m2

around noon time. Fig. 10 (b) shows active power generatedfrom the studied solar PV system with and without consideringambient temperature. It is assumed that solar PV system isoperating in MPPT mode during the entire day, while the solarPV inverter stays grid connected at night for voltage support,


6

(a) (b) (c)

Fig. 10: PV system performance under weather conditions: Al Qaseem, Saudi Arabia, 04/08/2002 [28]. (a) Ambient temperature and solarirradiance, (b) active power and reactive power reserve for v=1 pu, (c) active power and reactive power reserve for v=0.9 pu

(a) (b) (c)

Fig. 11: PV system performance under weather conditions: Bikaner, India, 24/06/2002 [29]. (a) Ambient temperature and solar irradiance,(b) active power and reactive power reserve for v=1 pu, (c) active power and reactive power reserve for v=0.9 pu

and terminals voltage is initially assumed to be 1 pu duringthe entire period. Minimum reactive power requirement, as perGerman grid code regulation, is represented by the bottomhorizontal line, as shown in Fig. 10. The blue shaded areacalculated using conventional approach represents untappedreactive power reserve that can be utilised for voltage controlancillary service beyond the grid code requirements. However,the Q reserve (shaded area) is not accurate as it is calculatedwithout considering impact of ambient temperature on PV ar-ray and PV inverter. A significant difference in reactive powerreserve can be observed between the proposed methodology(hatched area) and conventional approach of Q-reserve calcu-lation as the ambient temperature rises over 25 ◦C. Therefore,it can be concluded that ignoring ambient temperature in the Qreserve calculation, may result in overestimation of the reserve.Fig. 10(c) presents active power and reactive power reserveof the inverter under the same conditions, however, at lowerterminals voltage, 0.9 pu. An important observation under lowterminal voltage condition is that by considering the realisticconstraints, even mandatory reactive power requirements maynot be met for short period of time during the day whenambient temperature and solar irradiance are high, as shownin Fig. 10(c). Therefore, ignoring ambient temperature inQ reserve calculation results in false overestimation of theinverter capability, and thus the bottleneck operating pointswhich may even result in breaching mandatory grid coderegulation cannot be observed under conventional approachof Q-reserve calculation.

Similar study has been carried out for another location,Bikaner, in Rajasthan state of India, and corresponding resultsare shown in Fig. 11. Solar irradiance and ambient temperaturedata for this location is taken from [29]. From Fig. 11(a), itcan be observed that ambient temperature for Bikaner locationrises to as high as 48 ◦C, however, with solar irradiance lessthan 1000 W/m2. This reduction in irradiance has a mitigationeffect on Q reserve, as can be observed from Fig. 11(b)compared with that in Fig. 10(b). The significant reductionof Q reserve during night in Fig. 11(b) compared to that inFig. 10(b) at night is due to the high ambient temperature inthe former case (Indian site) compared to the later one (Saudisite). Moreover, due to the reduced solar irradiance, mandatoryreactive power requirements are almost met without deficit forlow terminal voltage (0.9 pu) as shown in Fig. 11(c).

It is uesful to note that in both the studied locations,maximum solar irradiance and ambient temperature peak donot coincide, as the ambient temperature peak occurs withsome delay. This delay helps in improving minimum Q reservethat would otherwise be further reduced in case the two factorscoincide with each other.

B. Performance of the proposed self-adaptive QV droop con-trol scheme

In order to study the performance of the proposed controlscheme, a detailed large scale PVPP developed in DIgSILENTPowerFactory is connected to bus 38 in IEEE 39 bus system, asshown in Fig. 12, with PVPP layout as given in [7], [30]. The


7

G

G1

2

3

30

25 26

27

28

1718

37

29

38

16 21

24 36

23

22

3533343231

1112

1020

1913

144

56

7

39

15

9

8GGG

G G

G

G

LSPVPP

Fig. 12: IEEE 39 bus system [31].

capacity of PVPP is same as that of the displaced synchronousgenerator connected at bus 38. The system and the PVPP arebuilt and simulated in DIgSILENT PowerFactory.

An additional reactive power load is connected througha circuit breaker at bus 28 to simulate a voltage dip atPVPP terminal at t=0.5 sec. Solar irradiance level and ambi-ent temperature are assumed to be 1000 W/m2 and 25 ◦C,respectively. Response of large scale PVPP to the voltagedip is shown in Fig. 13(a), where it can be observed thatreactive power generated by the PVPP connected at bus 38is higher in the proposed adaptive QV droop compared tothe fixed droop, which can be attributed to utilisation ofuntapped reactive power reserve of the solar PV inverters bythe proposed adaptive control scheme, described in SectionIV. This additional reactive power supply has resulted in animprovement in the terminals voltage as can be observedfrom PCC voltage shown in Fig. 13(a). Response of a singleinverter to the voltage event at the PVPP terminal is shownin Fig. 13(b), where higher QV droop gain can be observedwhen adaptive control scheme is used compared to fixed droopscheme. The adaptive droop is reduced after the disturbancedue to the reduced inverter voltage terminals, which in turnreduces Qmax.

For a comprehensive analysis of the proposed controlscheme, a time series load variation is introduced in IEEE 39bus system to apply voltage variations at PVPP terminals. Fur-ther, to incorporate seasonal variation, three different weatherconditions are considered, viz., high solar irradiance with lowand high ambient temperature, and low solar irradiance andambient temperature as shown in Fig. 14. It can be observedfrom Fig. 14 that large scale PVPP reduces terminal voltagevariations in case of adaptive QV droop compared to fixedQV droop, which can be attributed to higher dynamic reactivepower compensation under the adaptive QV droop. Moreover,PVPP under adaptive QV control in the first case, Fig. 14(a),outperforms the PVPP dynamic behaviour compared to case2, Fig. 14(b). This is due to the higher reactive power reserveof the PVPP under lower ambient temperature, which isreflected in higher QV droop in the first case compared tothe second case. In the third case, Fig. 14(c), PVPP showsa better performance than the first two cases due to lower

(a)

(b)

Fig. 13: (a) PVPP response to voltage dip. (b) A single solarPV inverter response to voltage dip.

solar irradiance level, and hence having higher reactive powerreserve and better QV droop.

VI. DISCUSSIONS AND CONCLUSIONS

In this paper, reactive power capability of solar PV invertersis studied considering variation in solar irradiance, ambienttemperature and terminal voltage. Based on the proposedaccurate and realistic reactive power capability calculation, aself-adaptive QV droop control is proposed to utilise reactivepower reserve of large scale PVPP for voltage control ancillaryservice. The following concluding points are drawn from thispaper.• Considering ambient temperature impact on solar PV

arrays only to estimate reactive power capability of solarPV inverter may lead to over estimation of reactivepower capability of the system. Incorporation of theimpact of ambient temperature on solar PV inverter givesmore accurate and realistic estimation of reactive powercapability of solar PV conversion system.

• The geographical location of PVPP plays a significantrole in determining its reactive power capability. Loca-tions with a very high temperature may require oversizingsolar PV inverter or installing additional reactive powersources, such as capacitors, to comply with the grid codemandatory requirements or to increase PVPP participa-tion in voltage control ancillary service. In cold countries,


8

0 1 2 3 4 5 60.9

0.95

1

1.05

1.1V

PCC

(V)

0 1 2 3 4 5 6Time (sec)

-600

-400

-200

0

200

400

Q(M

VA

R)

0.4 0.5 0.6100

130

160

3.5 3.55 3.6100

150

200

1.9 1.95 2

-400

-300

(a) G=1000 W/m2, T= 25 ◦C

0 1 2 3 4 5 60.9

0.95

1

1.05

1.1

VPC

C (V

)

0 1 2 3 4 5 6Time (sec)

-600

-400

-200

0

200

400

Q(M

VA

R)

0.4 0.5 0.6100

130

160

1.9 1.95 2

-400

-300

3.5 3.55 3.6100

150

200

(b) G=1000 W/m2, T= 40 ◦C

0 1 2 3 4 5 60.9

0.95

1

1.05

1.1

VPC

C (V

)

0 1 2 3 4 5 6Time (sec)

-600

-400

-200

0

200

400

Q(M

VA

R)

(c) G=700 W/m2, T= 25 ◦C

Fig. 14: PVPP voltage control response to voltage disturbances under various irradiance and ambient temperature conditions.

TABLE II: Specifications of Solar PV module and inverter

Inverter [23] Solar PV module @ STC [32]Nominal AC voltage 690 V Nominal Power 355 W

Frequency 50 Hz VMPP 43.4 VRated AC current at 25 ◦C 3850 A IMPP 8.18 ARated AC current at 50 ◦C 3273 A VOC 51.9 VRated AC power at 25 ◦C 4.6 MVA ISC 8.68 ARated AC power at 50 ◦C 3.91 MVA Power temp. coef. -0.37%/◦C

Max. DC voltage 1500 V Voltage temp. coef. -176.5 mV/◦CMax. efficiency 98.7% Current temp. coef. 3.6 mA/◦C

on the other hand, the change in reactive power capabilityof solar PV inverters is trivial.

• Large scale PVPP reactive power capability in practicediffers significantly from solar PV inverter capabilityunder standard test conditions. Incorporating practicalaspects such as, weather conditions provides an accurateestimation of PVPP, which is critical for PVPP operatoras well as transmission system operator.

APPENDIX

Table II shows solar PV system parameters. The number ofparallel and series PV modules are Np = 370 and Ns = 28.

REFERENCES

[1] P. Fairley, “The pros and cons of the world’s biggest solar park,” IEEESpectrum, 2020.

[2] Z. H. Rather, Z. Chen, P. Thøgersen, and P. Lund, “Dynamic reactivepower compensation of large-scale wind integrated power system,” IEEETrans. Power Sys., vol. 30, no. 5, pp. 2516–2526, 2015.

[3] EirGrid and SONI, “DS3 System Services: Portfolio Capability Analy-sis,” 2014, online: www.eirgrid.com.

[4] H. Yu, J. Pan, and A. Xiang, “A multi-function grid-connected pv systemwith reactive power compensation for the grid,” Solar energy, vol. 79,no. 1, pp. 101–106, 2005.

[5] F. L. Albuquerque et al., “Photovoltaic solar system connected to theelectric power grid operating as active power generator and reactivepower compensator,” Solar Energy, vol. 84, no. 7, pp. 1310–1317, 2010.

[6] A. Ellis et al., “Reactive power interconnection requirements for pv andwind plants-recommendations to NERC,” Sandia National Laboratories,NM, USA, Tech. Rep., 2012.

[7] C. Loutan et al., “Demonstration of essential reliability services by a300-MW solar photovoltaic power plant,” NREL, Tech. Rep., 2017.

[8] R. K. Varma et al., “Nighttime application of pv solar farm as statcomto regulate grid voltage,” IEEE Trans. Energy Conv., vol. 24, no. 4, pp.983–985, 2009.

[9] R. K. Varma et al., “New control of PV solar farm as STATCOM (PV-STATCOM) for increasing grid power transmission limits during nightand day,” IEEE Trans. Power Del., vol. 30, no. 2, pp. 755–763, 2014.

[10] L. F. N. Lourenco, de Camargo Salles et al., “Technical Cost ofOperating a Photovoltaic Installation as a STATCOM at Nighttime,”IEEE Trans. Sustain. Energy, vol. 10, no. 1, pp. 75–81, 2018.

[11] L. Lourenco, R. Monaro, M. Salles et al., “Evaluation of the reactivepower support capability and associated technical costs of photovoltaicfarms operation,” Energies, vol. 11, no. 6, p. 1567, 2018.

[12] K. Loudiyi et al., “Grid code status for wind farms interconnectionin northern Africa and Spain: Descriptions and recommendations fornorthern Africa,” Renew. Sustain. Energy Rev., vol. 81, 2018.

[13] EirGrid, “Grid code, version 8,” Online: http://www.eirgrid.com, 2019.[14] Tennet TSO GmbH, “Grid code, high and extra high voltage,” Germany,

November, 2015.[15] “Enhanced Reactive Power: A guide to the services procured by National

Grid to manage voltages on the transmission system,” National Grid,available at http://www2.nationalgrid.com.

[16] N. R. Ullah, K. Bhattacharya, and T. Thiringer, “Wind farms as reactivepower ancillary service providerstechnical and economic issues,” IEEETrans. Energy Conv., vol. 24, no. 3, pp. 661–672, 2009.

[17] H. Karbouj and Z. H. Rather, “Voltage control ancillary service fromwind power plant,” IEEE Trans. Sustain. Energy, vol. 10, no. 2, pp.759–767, 2018.

[18] M. Sarkar, M. Altin, P. E. Sørensen, and A. D. Hansen, “Reactive powercapability model of wind power plant using aggregated wind powercollection system,” Energies, vol. 12, no. 9, p. 1607, 2019.

[19] D. Sera, R. Teodorescu, and P. Rodriguez, “PV panel model based ondatasheet values,” in IEEE Int. Symp. Indust. Electron., 2007.

[20] H. Zhang and L. M. Tolbert, “Efficiency impact of silicon carbide powerelectronics for modern wind turbine full scale frequency converter,”IEEE Trans. Indust. Electron., vol. 58, no. 1, pp. 21–28, 2010.

[21] R. G. Wagoner and D. Smith, “Power system junction temperaturecontrol,” Mar. 15 2016, US Patent 9,287,766.

[22] “Technical Information: Temperature derating,” SMA Solar Tech-nology AG, Tech. Rep., available at https://files.sma.de/dl/7418/Temp-Derating-TI-en-15.pdf.

[23] “Sunny Central UP datasheet,” SMA Solar Technology AG, available athttps://www.sma.de/.

[24] “Solar Inverters: ABB Compact Skid for US market — PVS980-CS-US2 to 4.4 MVA,” ABB, available at https://www.abb.com/.

[25] “Sunny Central 2200/2475/2500-EV/2750-EV/3000-EV datasheet,”SMA Solar Technology AG, available at https://www.sma.de/.

[26] “Solar Inverters: ABB inverter station — PVS800-IS 1.75 to 2 MW,”ABB, available at https://www.abb.com/.

[27] “Product specifications Conext SmartGenTM

CS2200 Model,” SchneiderElectric, available at https://solar.schneider-electric.com/.

[28] National Renewable Energy Laboratory (NREL). NASA remotesensing validation data: Saudi Arabia. [Online]. Available: https://www.nrel.gov/grid/solar-resource/saudi-arabia.html

[29] NREL. National Solar Radiation Database (NSRDB). [Online].Available: https://maps.nrel.gov/nsrdb-viewer

[30] Cabrera-Tobar et al., “Review of advanced grid requirements for theintegration of large scale photovoltaic power plants in the transmissionsystem,” Renew. and Sustain. Energy Rev., vol. 62, pp. 971–987, 2016.

[31] P. Demetriou et al., “Dynamic IEEE test systems for transient analysis,”IEEE Syst. J., vol. 11, no. 4, pp. 2108–2117, 2017.

[32] “SunPower R© Performance Series 1500 Volt p17,” SunPower, availableat https://us.sunpower.com/.


Adaptive voltage control for large scale solar PV power plant ...

Documents