Investigation of the interaction between step voltage regulators …381149/Ruifeng... · 2019. 10. 11. · This paper is published in IET Renewable Power Generation, Vol. 10, 2016.This

This paper is published in IET Renewable Power Generation, Vol. 10, 2016. This is the authors’

version of the paper.

1

Investigation of the interaction between step voltage regulators and large-scale

photovoltaic systems regarding voltage regulation and unbalance

Mohammed Imran Hossain, Ruifeng Yan,* and Tapan Kumar Saha

School of Information Technology and Electrical Engineering,

University of Queensland,

Brisbane, QLD 4072, Australia

[email protected]; [email protected]; [email protected]

Abstract: The number of installed photovoltaic (PV) systems has been

increasing in an unprecedented rate every year throughout the world. In some situations,

the PV systems can provide more power than the demand in certain parts of a network

and cause reverse power flow, which the traditional distribution networks are not

designed for. Consequently, these instances may result in adverse interactions between

photovoltaic systems and voltage regulation devices such as step voltage regulators

(SVRs). This can potentially drive voltage magnitudes beyond acceptable limits and

possibly damage consumers’ appliances. In the literature, such interaction cases with

consideration of realistic three-phase four-wire unbalanced networks have not yet been

reported. Therefore, this paper analyses the interaction between SVR and PV systems in

a real world network by utilizing the quasi-static time series technique for a long-term

statistical assessment. This evaluation focuses on examining voltage regulation issues in

an unbalanced three-phase four-wire network along with an open-delta step voltage

regulator configuration, and the corresponding solution is proposed to resolve the

concerning issues. The results of this study will provide valuable information on

* Dr. Ruifeng Yan, School of Information Technology and Electrical Engineering, The

University of Queensland, St. Lucia, QLD 4072, Australia. Tel: +61 7 33653394 , Email:

[email protected]

mailto:[email protected]:[email protected]:[email protected]:[email protected]



2

interaction characteristics to academia and utilities for assessing large-scale PV system

integration.

Keywords: photovoltaic integration, step voltage regulator, unbalanced system,

voltage regulation.

1. Introduction

The capacity of photovoltaic (PV) systems has grown with an average of 50%

per annum around the world [1]. The growth is even higher in the context of Australia,

where it has increased from 474MW in November 2010 to 3,897MW in November

2014 [2-3]. It is clear that PV growth is constantly rising and substantial increment is

expected in the future. This large volume of PV generation is dependent on weather

conditions and can have rapid and substantial fluctuations due to cloud coverage in a

local area [4]. As most of these generation systems are connected to power networks [5],

such PV generation fluctuations can result in a significant amount of variations on

network voltage magnitudes, which may violate the voltage standards and cause harm to

consumer appliances.

In the literature, there are a number of studies focusing on the effects of PV

fluctuations on voltage amplitudes [6]. Tonkoski investigated the impacts of high

photovoltaic penetration on voltage profiles for low voltage level residential

neighbourhoods [7]. This study considered the role of feeder impedance, feeder length

and transformer short circuit resistance, but the unbalanced nature of the low-level

network was not taken into account. With network unbalance in consideration, analyses

have been conducted on the sensitivity and stability of voltage in distribution networks

with high PV penetration [8, 9]. It was found that fluctuations of PV power generation

could cause voltage magnitude to change rapidly and unevenly across the phases. Later,

Alam has investigated the impacts of PV fluctuation and proposed a computational tool



3

to identify excessively changing characteristics of voltage levels, reverse power flow,

feeder power loss, voltage unbalance, and tap operation [10, 11]. However, these studies

have not addressed long-term dynamic impacts of PV integration on voltage variations,

which can reveal essential information on the statistical performance of a distribution

network.

There are also several studies in the literature on the interaction between

photovoltaic systems and voltage regulating devices. A dynamic adjustment method of

on-load tap changer (OLTC) parameters with distributed generation reactive power

support has been proposed in a time domain assessment [12]. This study has considered

the interaction of distributed generation with the operation of tap changer regulated to

ensure effective voltage control. Similarly, to regulate voltage magnitudes, coordination

between OLTC and static VAr compensator with distributed generation has been

investigated by considering these devices in uniform or independent operation across

phases [13]. However, the networks in these studies can still have critical voltage

concerns if the interactions between PV and tap changing regulators are considered with

weather dependent PV generation. Yearlong examinations on tap changing operations

have been conducted considering cloud induced PV fluctuations at different PV

penetration levels [14]. Although this investigation has detailed impacts on voltage

amplitudes, the study was not performed on an unbalanced network. An online method

was developed to control voltage by regulating OLTCs and another tap changer

operated device - step voltage regulators (SVRs), in a network with a high level of

distributed generators [15]. However, this research needs to address the impacts of the

interactions of PV fluctuations and SVRs with the consideration of different SVR

connection types (such as open-delta, open-wye etc.).



4

The impacts of PV generation uncertainty on voltage magnitude and unbalance

have been investigated with load variability in [16-20], and these works have utilized

the probabilistic load flow method for statistical voltage assessment. However, the

probabilistic load flow approach cannot provide sequential results for analysing the

characteristics of time dependent devices such as SVRs, which rely on sequentially

changing tap positions for voltage regulation. Since PV output uncertainty may

substantially increase the number of tap changes and further affect SVR lifetime and

maintenance, realistic SVR operation needs to be evaluated, which cannot be achieved

via the probabilistic load flow technique. Therefore, the quasi-static load flow method

that is based on successive time domain processing is utilized in this study to examine

SVR performance and network voltage quality.

In summary, the effects of PV generation fluctuations need to be addressed for

three-phase four-wire unbalanced networks with examination in time series analysis for

a considerable amount of time to attain statistical significance. Further, the

characteristics of different SVR connection types, which can reveal points of concern in

the interaction with PV systems, should also be taken into account. Finally, the quasi-

static load flow method should be utilized to examine the performance of SVR in

tandem with PV systems. This paper analyses the detailed interactions of SVR and PV

systems to evaluate the voltage quality in terms of voltage magnitude and fluctuation in

a real world unbalanced distribution network. Considering these analyses, a novel

solution that utilizes power line communication is developed to address the concerns

that rise from the investigation.



5

2. Investigated System

2.1 Gatton campus network in The University of Queensland

The system studied in this paper is the Gatton campus network in the University

of Queensland (UQ), Australia with a large-scale solar PV array under construction. Fig.

1 is the electric network map of the Gatton campus of UQ. As can be seen from this

map an 11kV line is bifurcated almost 3 km after the Gatton zone substation and they

stretch to the northern and southern point of the campus. Between these two lines, the

southern connection is operational for the campus while the northern connection

provides other surrounding suburbs [21]. The line length from the zone substation to the

southern SVR is about 4 km.

Fig. 1. Map of University of Queensland, Gatton [21]

The SVR is a Cooper Power system product and is arranged in an open-delta

configuration. Two tap changing regulators of the SVR are connected across Phases A-

B and Phases B-C, which will be referred as AB and BC respectively in this paper. This



6

SVR has the ability to regulate voltage in both forward and reverse operation by sensing

current (mostly real current, sometimes reactive current) directions. The SVR operates

with the load drop compensator (LDC) which estimates the voltage value at the load

centre (the UQ Gatton campus in this study) and it accordingly changes tap positions to

regulate this load centre voltage towards a predefined target.

The rest of line after the southern SVR is around 3 km long until the UQ Gatton

campus. This 11kV line is connected to the 415 V network inside the campus via a

number of delta to wye-ground transformers. From the local electricity distributor

records, it has been gathered that the total demand of the campus is normally between 1

MW to 3 MW [21]. The PV system is connected to the 11kV line via a separate delta to

wye-ground transformer. The solar array is a 3.3MWp power generation plant over a

10-hectare area. The network components in this system have been enlisted with

pertinent parameters in Table 1.

Table 1. Description of network components

Items Parameters

Gatton Zone Substation 33/11kV, 25MW [21]

Underground Cable (UG) 240mm2 conductor [22, 23]

Racoon Conductor (0.44+j0.28)Ω/km [24]

Moon Conductor (0.28+j0.25) Ω/km [25]

Step Voltage Regulator Open-Delta connection. Between A-B

phases and B-C phases [26]

Solar Array 3.3MWp [21]

Delta to Wye-Ground Transformer 11kV/433V [21]

Campus Load Maximum 3MW and minimum 1MW [21]

2.2 Representation of the network

A representation of the Gatton campus network used for the analysis is shown in

Fig. 2, which closely resembles the actual network (Fig. 1). To conduct the three-phase



7

four-wire unbalanced network analysis, load flow program based on Current Injection

Method [27] has been developed in MATLAB [28]. The unbalanced characteristics of

the network have been realized by acquiring unbalanced pole top configurations of lines

and loads from the local utility company [29]. The model of the three-phase four-wire

unbalanced line for network analysis is given in (1) [30, 31], where 𝑧𝑙𝑖𝑛𝑒 is the line

impedance matrix, i and j represent system phases (Phase-A, Phase-B, Phase-C or

Phase-N), 𝑧𝑠𝑖 is the self-impedance of Phase-i and 𝑧𝑚𝑖𝑗 is the mutual impedance

between Phase-i and Phase-j.

Fig. 2. Representation of Gatton network used in the simulation

𝑧𝑙𝑖𝑛𝑒 = [

𝑧𝑠𝑎 𝑧𝑚𝑎𝑏 𝑧𝑚𝑎𝑐 𝑧𝑚𝑎𝑛𝑧𝑚𝑏𝑎 𝑧𝑠𝑏 𝑧𝑚𝑏𝑐 𝑧𝑚𝑏𝑛𝑧𝑚𝑐𝑎 𝑧𝑚𝑐𝑏 𝑧𝑠𝑐 𝑧𝑚𝑐𝑛𝑧𝑚𝑛𝑎 𝑧𝑚𝑛𝑏 𝑧𝑚𝑛𝑐 𝑧𝑠𝑛

] (1)

The self and mutual impedance is calculated by (2) and (3), where 𝑟𝑖, 𝑓, 𝜌, 𝐺𝑀𝑅𝑖,

𝐷𝑖𝑗 and 𝐿 denote resistance of Phase-i, system frequency, Earth resistivity, Geometric

Mean Radius of Phase-i, and distance separation between Phase-i and Phase-j and line

length [30].



8

𝑧𝑠𝑖 = [𝑟𝑖 + 0.00158836𝑓 + 𝑗0.00202237𝑓 (𝑙𝑛1

𝐺𝑀𝑅𝑖+ 7.6786 +

1

2𝑙𝑛

𝜌

𝑓)]. 𝐿 (2)

𝑧𝑚𝑖𝑗 = [0.00158836𝑓 + 𝑗0.00202237𝑓 (𝑙𝑛1

𝐷𝑖𝑗+ 7.6786 +

1

2𝑙𝑛

𝜌

𝑓)]. 𝐿 (3)

The model for delta-to-wye-ground transformer has been adopted and modified

from [32, 33] and expressed in (4), where 𝑦𝑡𝑟𝑎𝑛𝑠𝑓𝑜𝑟𝑚𝑒𝑟 represents the admittance matrix

of a three-phase four-wire delta to wye-ground transformer and 𝑦𝑙 is the transformer

leakage admittance.

𝑦𝑡𝑟𝑎𝑛𝑠𝑓𝑜𝑟𝑚𝑒𝑟 =

[

2

3𝑦𝑙 −

1

3𝑦𝑙 −

1

3𝑦𝑙 −

1

√3𝑦𝑙

1

√3𝑦𝑙 0 0

−1

3𝑦𝑙

2

3𝑦𝑙 −

1

3𝑦𝑙 0 −

1

√3𝑦𝑙

1

√3𝑦𝑙 0

−1

3𝑦𝑙 −

1

3𝑦𝑙

2

3𝑦𝑙

1

√3𝑦𝑙 0 −

1

√3𝑦𝑙 0

−1

√3𝑦𝑙 0

1

√3𝑦𝑙 𝑦𝑙 0 0 −𝑦𝑙

1

√3𝑦𝑙 −

1

√3𝑦𝑙 0 0 𝑦𝑙 0 −𝑦𝑙

01

√3𝑦𝑙 −

1

√3𝑦𝑙 0 0 𝑦𝑙 −𝑦𝑙

0 0 0 −𝑦𝑙 −𝑦𝑙 −𝑦𝑙 3𝑦𝑙 ]

(4)

The other model which is integral for this investigation is the open-delta SVR

model developed from Cooper Power Systems manual of the CL-6 SVR [26] and [34],

which is given in (5), where 𝑟1 and 𝑟2 are the effective turns ratios of AB and BC

regulators in the SVR and 𝑦𝑙 is the leakage admittance of the regulators.

𝑦𝑜𝑝𝑒𝑛−𝛥−𝑆𝑉𝑅 =

[

𝑦𝑙 −𝑦𝑙 0 −𝑟1𝑦𝑙 𝑟1𝑦𝑙 0−𝑦𝑙 2𝑦𝑙 −𝑦𝑙 𝑟1𝑦𝑙 −(𝑟1 + 𝑟2)𝑦𝑙 𝑟2𝑦𝑙0 −𝑦𝑙 𝑦𝑙 0 𝑟2𝑦𝑙 −𝑟2𝑦𝑙

−𝑟1𝑦𝑙 𝑟1𝑦𝑙 0 𝑦𝑙𝑟12 −𝑦𝑙𝑟1

2 0

𝑟1𝑦𝑙 −(𝑟1 + 𝑟2)𝑦𝑙 𝑟2𝑦𝑙 −𝑦𝑙𝑟12 𝑦𝑙𝑟1𝑟2 −𝑦𝑙𝑟2

2

0 𝑟2𝑦𝑙 −𝑟2𝑦𝑙 0 −𝑦𝑙𝑟22 𝑦𝑙𝑟2

2 ]

(5)

The demand in the network has been acquired primarily from the total load report

prepared by Gatton campus authority. Next, the unbalanced and distributive

characteristics of the loads have been addressed by observing the time domain demand

trend [35] of individual buildings in the campus and report from distribution utility [36].



9

In the modelling, the first seven buses have been selected from all possible phase

loading configurations to investigate the different unbalanced effects. A sample load

profile for a moment with 2MW total load has been enlisted in Table 2.

Table 2. Sample load profile in the distribution network for a moment

Power demand per phase (kW)

Bus Number Phase A Phase B Phase C

22 37.5 41.7 45.8

23 37.5 45.8 41.7

24 41.7 37.5 45.8

25 45.8 41.7 37.5

26 45.8 37.5 41.7

27 41.7 45.8 37.5

28 41.7 41.7 41.7

29 45.9 42.2 38.6

30 42.5 40.0 37.5

31 40.7 39.4 37.9

32 42.6 38.9 35.0

33 46.4 44.1 41.6

34 40.6 39.7 38.6

35 42.5 39.9 37.3

36 47.3 45.3 43.2

37 49.5 45.7 42.1

The 3.3 MWp UQ Gatton campus solar plant is under construction and yet to be

commissioned. Therefore, the actual PV output data is not available. Fortunately, UQ

has a 1.2 MWp solar plant at its St. Lucia campus, which has recorded 1-minute interval

data of PV instantaneous power generation from 2011 with the variable characteristics

of PV output due to different weather conditions [37]. Though the St. Lucia solar plant

is smaller in terms of capacity, it is scattered over a comparable geographical area to

that of the Gatton PV plant. This presents an opportunity of having a similar percentage



10

of PV generation fluctuation due to cloud coverage transition. Therefore, the PV data

from St. Lucia has been utilized and scaled up to the Gatton campus 3.3 MWp PV

arrays to be used for long-term PV impact assessment.

2.3 Related standards

In this investigation the results are analysed with respective standards of voltage

magnitude, phase unbalance and voltage fluctuation, which are followed by the local

electricity distributor. Firstly, the acceptable allowance of voltage magnitude at a low

voltage level is ±6% of the base voltage of 240V [38]. Temporary voltage violation is

accepted up to 1% of measurements or 100 minutes in a week [39]. Next, according to

IEEE Std 1159-2009 supply voltages are typically regulated at less than 1% unbalance,

although 2% is not unusual [40]. To calculate the unbalance percentage the following

equations [40] have been utilized, where 𝑉𝐴𝐵,𝑉𝐵𝐶 , 𝑎𝑛𝑑 𝑉𝐶𝐴 are phase to phase voltage

between Phase A-B, Phase B-C and Phase C-A.

𝛽 = |𝑉𝐴𝐵|

4+|𝑉𝐵𝐶|4+|𝑉𝐶𝐴|

4

(|𝑉𝐴𝐵|2+|𝑉𝐵𝐶|2+|𝑉𝐶𝐴|2)2 (6)

𝑉𝑢𝑛𝑏𝑎𝑙𝑎𝑛𝑐𝑒(%) = √1− √3−6𝛽

2

1+ √3−6𝛽2

2× 100% (7)

Finally, the AS 2279.4-1991 standard [41] is followed for evaluating voltage

fluctuations by the Queensland distributor which will be presented in detail in Section 3.

3. Analysis of interaction between SVR and PV on long-term voltage

performance

To analyse the interaction between SVR and PV in terms of voltage magnitude

and fluctuation in a long-term, a few considerations have to be taken into account.

Firstly, based on the SVR setting in Gatton network [42], a 2-minute response delay has

been adopted for the SVR in this study. The investigation utilizes 1-minute interval for

quasi-static time series analysis, which is adequate to incorporate the SVR operation



11

delay. Next, due to cloud induced PV output fluctuations, voltage variations are

expected. Therefore, the standards of perceptibility and irritability (terms taken from

[41]) are used for evaluating the impacts of PV influenced voltage fluctuations. Table 3

lists the standard criteria of the relative voltage change thresholds and the respective

permitted violation allowance.

Table 3. Values from voltage fluctuation standard [41]

Acceptable Maximum Occurrences 5 10 20 30 60

Daily relative voltage change thresholds for

threshold of perceptibility (%)

1.60 1.30 1.10 1.00 0.80

Daily relative voltage change thresholds for

threshold of irritability (%)

4.20 3.70 3.00 2.70 2.40

3.1 Simulation result analysis for a sample day

3.1.1 Simulation without PV

The simulation for a sample day has been conducted without PV to establish the

base case of the network and pave the way to examine the interaction between SVR and

PV. Generally, distribution systems have smooth demand changes. As a result, rapid

variations of voltage magnitudes are expected only due to SVR tap changes, and it is

normally not a serious concern for the network.

The investigation on tap operation of the SVR and voltage magnitude at Bus 27

of this network without PV has been depicted in Fig. 3, which presents the status of the

system. It can be seen that, in a 14-hour period of a typical day around 5 tap changes

can be expected per regulator, and the phase voltage variation range is less than 8%.



12

Fig. 3. Daylong tap operation, voltage magnitude results from simulations without PV

3.1.2 Simulation with PV – LDC voltage target setting of 0.99pu

With the introduction of PV, fluctuations of power generation due to cloud

transient can cause more voltage variations than that of the scenario without PV

integration (Fig. 4). The target value for the SVR has been selected as 0.99pu. Tap

changing activity has risen to 47 times for AB and 43 times for BC of the SVR [Fig.

4(b)]. The voltage variations [Fig. 4(c)] can be observed due to PV output fluctuation

and the lower limit of 0.94pu was violated for about 5 minutes. The total voltage

variation range has increased to over 11% around the day.

In summary, the voltage magnitudes have been severely affected than those

without PV (Section 3.1.1). The total voltage range covered in a sample day has risen

from less than 8% to more than 11%. The allowable limits for voltage are 6% above and

below the base, which gives a 12% range for regulation. However, the study indicates

the margin left for regulation reduces from more than 4% to less than 1%. As a result,

the network becomes more vulnerable to any inaccurate settings and additional

disturbances such as LDC voltage target settings and load changes. Furthermore, the tap

6AM 8AM 10AM 12PM 2PM 4PM 6PM

0.95

1

1.05

(c) Voltage at Bus 27

Time of day

Vo

lta

ge

(p

u)

2

4

6

(b) Tap position (open-delta)

Ta

p p

ositio

n

0.5

1(a) Load Profile (per phase)

Po

we

r (M

W)

A

B

C

Limit

AB

BC

A

B

C



13

change activity has increased about nine times more with large-scale PV, which will

affect the durability of the SVR.

Fig. 4. Daylong tap operation, voltage magnitude results from simulations with PV

(LDC voltage target setting of 0.99pu)

3.1.3 Simulation with PV – LDC voltage target setting of 1.01pu

The LDC target value of the SVR is an important factor for network voltage

regulation. As the distribution network is perpetually changing, it is quite hard to predict

the suitable target value for SVR deployment. Therefore, the SVR target value can be at

a certain level that was fixed long time ago, which may not be entirely ideal for the

current network conditions [43]. Non-ideal target value may drive voltage magnitudes

to violate the acceptable limits more frequently. In order to study such cases, a different

LDC target has been selected in this section. As undervoltage was observed with an

LDC target of 0.99pu (Section 3.1.2), a higher target of 1.01pu is selected to elevate the

voltage magnitudes in the network.

By comparing the results between two different LDC target settings – 0.99pu

(Fig. 4) and 1.01pu (Fig. 5), the number of tap changing actions are very similar.

However, voltage magnitudes [Fig. 5 (c)] are now violating the acceptable limits for a


0.95

1

1.05


Time of day

Vo

lta

ge

(p

u)

2

4

6


Ta

p p

ositio

n

0

0.5

1

(a) PV Generation Profile (per phase)

Po

we

r (M

W)

A

B

C

Limit

AB

BC



14

longer time, from 5 minutes undervoltage (below 0.94pu) violation with the 0.99pu

voltage target to 23 minutes of overvoltage (over 1.06pu) with the 1.01pu voltage target.

Whereas the weekly acceptable limit is 100 minutes, the non-compliance limit for a day

is breached only due to the SVR target value change. This violation occurs because

there is very little room for imprecise settings due to the reduction of leftover voltage

variation margin (less than 1%) caused by the PV power fluctuations.

From Figs. 5-7 an increased level of tap operation, voltage magnitude violation,

and voltage fluctuation can be observed. However, violations in a day do not necessarily

lead to any serious concerns. In order to achieve statistical significance of a detailed

impact of these phenomena, examination over a much longer period should be

conducted.

Fig. 5. Daylong tap operation, voltage magnitude results from simulations with PV

(LDC voltage target setting of 1.01pu)

3.2 Analysis of yearlong performance

3.2.1 Voltage magnitude

Firstly, Table 4 is comprised of the voltage violation data in minutes for an

entire year. Without PV systems, a maximum of 0, 4, and 28 minutes violations in a


0.95

1

1.05


Time of day

Vo

lta

ge

(p

u)

2

4

6


Ta

p p

ositio

n

0

0.5

1


Po

we

r (M

W)

A

B

C

Limit

AB

BC



15

week can be observed in three phases. These numbers rise up to 30, 22, and 220 minutes

respectively with PV, which violates the 100 minutes per week limit according to the

utility standard [39].

Table 4. Voltage violation in terms of minutes over a year

Annual Total Maximum in a week

Allowable Limit - 100min/week

Phases A B C A B C

Total time beyond acceptable voltage

limit without PV system (minutes) 0 130 1456 0 4 28


limit with PV system (minutes) 332 168 8308 30 22 220

3.2.2 Tap operation

Next, the number of tap changing operation has been shown in terms of annual

total and daily maximum in Table 5. The number of tap changes across AB and BC has

increased about 5 and 6 times annually and around 13 times for both in an extreme day.

This amount of operation increase indicates decreased durability and increased

maintenance cost of the SVR.

Table 5. Step voltage regulator operation over a year

Annual Total Maximum in a day

Step voltage regulator connection AB BC AB BC

Number of tap changes without PV system 1924 1612 6 6

Number of tap changes with PV system 9998 9013 80 77

3.2.3 Voltage fluctuation

Finally, the last concern is the voltage fluctuation, which has been investigated

through a yearlong study. Table 6 is enlisted with the fluctuation violation criteria for

both concerns of perceptibility and irritability respectively. When the SVR is regulating



16

the voltage of the network (Fig. 2) without PV, the average violations in a day are very

well within the limit (Table 6). However, with PV systems, the standard is more likely

to be breached as it can be seen that the average occurrence of daily voltage violation

regarding the 0.80% threshold has risen from 4.5/day to 59.5/day for Phase-A. This will

significantly violate the 60/day limit for this threshold in a year.

Occurrences of exceeding the threshold for irritability are rarer than those for

perceptibility. Although, the irritability examination shows low levels of average

number of violation, occurrences increase with the introduction of high PV generation.

The average values for irritability do not signify the same level of concerns as for

perceptibility. However, the average number of violations does not entirely reflect the

severity of voltage fluctuations caused by large-scale PV integration.

Table 6. Average occurrences of violation of threshold of perceptibility and irritability

for a year

Perceptibility examination (phases) A B C A B C

Relative voltage change thresholds

(Maximum daily acceptable occurrences)

1.30 %

(10)

0.80 %

(60)

Average number of daily violations without

PV generation 4.5 3.8 0.4 4.5 3.8 4.6

Average number of daily violations with

PV generation 38.7 29.8 22.6 59.5 49.4 46.8

Irritability examination (phases) A B C A B C

Relative voltage change thresholds

(Maximum daily acceptable occurrences)

3.70 %

(10)

2.40 %

(60)

Average number of daily violations without

PV generation 0.7 0.6 0.4 1.3 1.0 0.7

Average number of daily violations with

PV generation 3.2 1.3 2.0 8.5 5.0 12.4

To demonstrate the complete fluctuation impacts on voltage Fig. 6 has been

delineated with the entire annual data. The figure has been plotted into boxplots where



17

blue boxes display the 25 to 75 percentile of results, and the central red lines represent

the median value. The extended dashed black lines illustrate the extreme data points

except the outliers, and the red plus symbols are the outliers, which are outstanding days

that have severe fluctuation incidences.

Fig. 6(a) shows that the median of daily voltage fluctuation incidences are over

the permitted limit, which creates some concern. At the same time, Fig. 6(b) depicts the

median values are just below the acceptable threshold. However, large portions of the

days in both figures are above the standard of perceptibility limits, which indicates that

the fluctuations on these days are perceivable to the consumers and their appliances.

Fig. 6. Daily fluctuation violation over a year

Fig. 6(c) and Fig. 6(d) depict the threshold of irritability with the relative voltage

change limits of 3.70% and 2.40%. These results indicate that most of the data lie below

the permitted limits. However, there are quite a few outstanding data points above the

standard limits with PV integration, which suggests that there are a number of days with

severe fluctuation violation. Such high PV fluctuations cause intolerable experience to

the consumers in these days.

A B C A B C

0

50

100

Without PV With PV

(a) Daily violations over 1.30% relativevoltage change (perceptibility standard)

Nu

mb

er

of vio

latio

ns

A B C A B C

0

100

200

Without PV With PV

(b) Daily violations over 0.80% relativevoltage change (perceptibility standard)

Nu

mb

er

of vio

latio

ns

A B C A B C

0

10

20

Without PV With PV

(c) Daily violations over 3.70% relativevoltage change (irritability standard)

Nu

mb

er

of vio

latio

ns

A B C A B C0

20

40

60

Without PV With PV

(d) Daily violations over 2.40% relativevoltage change (irritability standard)

Nu

mb

er

of vio

latio

ns

Limit



18

From all the tables and figures it has been exhibited that though the average

number of daily fluctuations may be within standard limits, there can still be a

significant amount of days when daily fluctuations exceed limits due to large-scale PV

output variations.

4. An advanced management solution to excessive voltage violations and tap

changes

The impacts of PV power fluctuations on network voltage and tap changers have

been examined via yearlong statistical analysis in the previous section, and concerns of

excessive voltage violation and tap change operation have arisen through the

investigation. Therefore, this section will focus on the solutions that can effectively

rectify these concerns.

4.1 Approved power factor droop control

The studied UQ Gatton PV system (3.3MWp) is equally divided into five

660kWp PV blocks, which are equipped with five SMA SC720CP inverters (720kVA

rating with maximum 346kVAr reactive power capacity) [44]. Such configuration has

left substantial inverter space for reactive power support to grid voltage. Therefore,

based on Australia National Electricity Rules [45], a form of reactive power

compensation strategy – power factor control (PFC) has been approved by both the

utility and UQ in the connection agreement [46]. The provisional control scheme is

illustrated in Fig. 7 with an allowable power factor range from 0.9 lagging to 0.9 leading.

A demonstration of network performance with the PFC strategy for a sample day

is shown in Fig. 8. By comparing with the grid profile without this control presented in

Fig. 5, it has been observed that by implementing the PFC strategy, the voltage variation

range has shrunk from over 11% to around 8%, and consequently voltage violation is

significantly decreased. Furthermore, the daily tap changes are also reduced from 90 to



19

48 in total; however, the amount of changes are still substantially higher than those of

the scenario without PV integration, which is again a serious concern for the operation

of voltage regulators.

Fig. 7. Power factor droop control curve (power factor vs voltage).

Fig. 8. Daylong network performance with the PFC strategy

4.2 Proposed advanced management solution using communication

Section 4.1 shows the PFC strategy alone may not be sufficient to resolve the

issue of excessive tap changes; therefore, an advanced management solution (AMS) is

0.95 1 1.05

0.90 lead

0.95 lead

1

0.95 lag

0.90 lag

Voltage (pu)

Po

we

r fa

cto

r

0.985 pu

Capacitive

Inductive

1.015 pu


0.95

1

1.05


Time of day

Vo

lta

ge

(p

u)

4

6


Ta

p p

ositio

n

0

0.5

1


Po

we

r (M

W)

A

B

C

Limit

AB

BC



20

developed in this research to reduce unnecessary tap movement. For example, when PV

power rapidly drops to a low level due to cloud coverage, the network voltage will

accordingly decline. If the voltage profile after such an event is still at a decent level

within the limits, tap changes will not be required. Normally the overall load variation is

very slow and the already decreased PV power cannot drop much lower, therefore, the

feeder voltages are at very little risk of falling any further to violate the voltage limit.

However, the conventional LDC scheme of an SVR does not have the essential

information of PV generation and system load. As a result, traditionally tap position is

changed as needed to regulate feeder voltage to the target with a large margin to the

limits. Based on this idea, the proposed AMS is designed with inclusion of a modern

cost-effective power line communication (PLC) system [47, 48] to provide PV and load

data to the SVRs for making decisions on tap changes. If the estimated further potential

voltage variation is less than the leftover voltage margin, then the tap will be held at its

current position regardless any tap changing instruction from the LDC. This proposed

control strategy is illustrated in Fig. 9.

The proposed control strategy includes two characteristic curves (Fig. 9), which

assess the voltage magnitude and potential voltage variation in the distribution network

from measured load (Pload), photovoltaic (PV) output, and voltage at the end of the 11kV

feeder (V8). These curves are derived from load flow results for all-possible PV and

load profiles based on the historical data, so they can be utilized for voltage estimation

at any time of a day. Firstly, the voltage at the low-voltage end can be evaluated from

the voltage drop obtained in the curve of ‘Est1’ (Fig. 9) and the measured upstream

voltage. In addition, the present PV power output can be gauged to determine the

potential maximum PV power fluctuations (∆Ppotential variation) for the next time interval,

and then the corresponding induced maximum voltage variations (∆Vpotential variation) can



21

be calculated from the curve in ‘Est2’ (Fig. 9). As these values are estimations, a

voltage safety margin (∆Vsafety margin) has been considered to create a safety buffer from

the absolute voltage limits and consequently reduce voltage violations. In this research,

1% voltage safety margin is selected after verifying in a yearlong analysis. Finally,

based on the two voltage estimations and the pre-defined voltage safety margin a

decision will be made on tap change requirement for a better voltage regulation. This

entire strategy constitutes the AMS, which reduces the amount of tap changes and the

risk of voltage violations.

Does the SVR

estimate to

change tap?

Estimate ΔVpotential variation

Get Ppv and estimate ΔPpotential variation

Get Pload and V8

Estimate ΔVleftover margin

ΔVpotential variation >ΔVleftover margin - ΔVsafety margin?

Change tap position as required

Do not change tap position

Yes

ΔVpotential variation

ΔPpotential variationPload ,V8

No

Start AMS

Yes

ΔVleftover margin

No

0 0.3 0.6 0.90

1.5

3

4.5

Ppotential

variation

(MW)

V

po

tential v

ariatio

n (

%)

0.4 0.6 0.8 1

-2

0

2

4

Pload

(MW)

Vd

rop (

%)

Est

2

Est

1

Fig. 9. Flow chart of the proposed AMS



22

The network performance with both the PFC and AMS is presented in Fig. 10.

In contrast to Fig. 8, the voltage range is similar and voltage profile becomes slightly

smoother. However, the major improvement has been the total number of tap changes

through the day – 8, which is much less than 48 with the pure PFC strategy.

Fig. 10. Daylong network performance with the PFC and AMS strategy

4.3 Comparison of yearlong performance

The yearlong analyses of voltage violations, tap changes and voltage

fluctuations are summarized in Table 7, Table 8, and Fig. 11 respectively. Different

scenarios have been listed for comparison, including no PV case, PV only case, PV with

PFC case, and PV with PFC-AMS case. The overall tendency shows that the PV only

case can cause the worst performance with respect to voltage and tap operation, while

the situation is significantly improved by introducing the PFC strategy. However, this

strategy may not effectively rectify the excessive tap changing issue. In contrast, the

proposed PFC-AMS strategy can successfully solve this issue, and overall it can

improve the voltage and tap performance to a similar level, which is comparable to that

of the no PV case.


0.95

1

1.05


Time of day

Vo

lta

ge

(p

u)

4

5

6


Ta

p p

ositio

n

0

0.5

1


Po

we

r (M

W)

A

B

C

Limit

AB

BC



23

Table 7. Annual voltage violation comparison between different scenarios

Average annually Maximum in a week

- 100min/week

Allowable Limit 528.7 28


limit without PV system (minutes) 2935.9 220


limit with PV system (minutes) 550.9 44


limit with PV and PFC (minutes) 550.9 44

Table 8. Annual tap change analysis in different investigations

Annual Total Maximum in a day

Step voltage regulator connection AB BC AB BC

Number of tap changes without PV system 1924 1612 6 6

Number of tap changes with PV system 9998 9013 80 77

Number of tap changes with PV and PFC 4734 3998 40 36

Number of tap changes with PV and PFC-

AMS 2107 1897 23 24

Fig. 11. Daily voltage violation comparison between PV only case and PV with PFC-

AMS case over a year

A B C A B C

0

10

20

With PV PFC-AMS

(c) Daily violations over 3.70% relativevoltage change (irritability standard)

Nu

mb

er

of vio

latio

ns

A B C A B C0

20

40

60

With PV PFC-AMS

(d) Daily violations over 2.40% relativevoltage change (irritability standard)

Nu

mb

er

of vio

latio

ns

A B C A B C

0

50

100

With PV PFC-AMS

(a) Daily violations over 1.30% relativevoltage change (perceptibility standard)

Nu

mb

er

of vio

latio

ns

A B C A B C

0

100

200

With PV PFC-AMS

(b) Daily violations over 0.80% relativevoltage change (perceptibility standard)

Nu

mb

er

of vio

latio

ns

Limit



24

5. Conclusions

This paper investigates the interaction between SVR and PV for a real world

unbalanced three-phase four-wire distribution network. A long-term quasi-static time

series analysis has been conducted to examine the repercussions on tap operations,

voltage magnitude variations, and voltage fluctuations. These effects have been studied

considering open-delta SVR in different target settings with cloud induced high PV

fluctuations. Finally, countermeasures have been introduced to remedy the adverse

effects.

The study has demonstrated scenarios where voltage violations can become

more frequent owed to reduced leftover margin of the acceptable voltage variation range

caused by PV power fluctuation and imprecise SVR voltage target from historical

settings. This is potentially harmful for the consumers’ appliances. In addition,

substantial rise in tap operations and numbers of voltage fluctuations beyond the

standards of perceptibility and irritability were observed due to PV output variability.

These increased tap operations may adversely affect the durability of SVR and

significantly augment maintenance cost. At the same time, high voltage fluctuations can

cause perceptible disturbances to consumers. To resolve these issues, an advanced

management solution has been proposed which can improve the concerning voltage and

tap performances to the acceptable levels. This work will contribute to both academics

and engineers by providing the aspects that need to be considered when evaluating

large-scale PV integration.

6. Acknowledgments

This work was performed in part or in full using equipment and infrastructure

funded by the Australian Federal Government’s Department of Education AGL Solar

PV Education Investment Fund Research Infrastructure Project. The University of



25

Queensland is the Lead Research Organisation in partnership with AGL, First Solar, and

the University of New South Wales.

7. References

[1] International Energy Agency, ‘About us - FAQs: Renewable Energy’,

http://www.iea.org/aboutus/faqs/renewableenergy/, accessed February 2015.

[2] Australian PV Institute, ‘Australian PV market since April 2001’, http://pv-

map.apvi.org.au/analyses, accessed February 2015.

[3] International Energy Agency, ‘High Penetration of PV in Local Distribution Grids’,

http://iea-pvps.org/index.php?id=295&eID=dam_frontend_push&docID=2210,

accessed February 2015.

[4] Nijhuis M., Rawn B., Gibescu M.: ‘Prediction of power fluctuation classes for

photovoltaic installations and potential benefits of dynamic reserve allocation’, IET

Journal on Renewable Power Generation, 2014, 8, (3).

[5] Gilmore J., Vanderwaal, B., Rose, I., Riesz, J., ‘Integration of solar generation into

electricity markets: an Australian National Electricity Market case study’, IET

Renewable Power Generation, 2014, 9, (1)

[6] Miñambres-Marcos V., Guerrero-Martínez M.A., Romero-Cadaval, E., González-

Castrillo, P,: ‘Grid-connected photovoltaic power plants for helping node voltage

regulation’, IET Renewable Power Generation, 2015, 9, (3)

[7] Tonkoski R., Turcotte D., El-Fouly T.H.M.: ‘Impact of high PV penetration on

voltage profiles in residential neighborhoods’, IEEE Trans. Sustainable Energy,

2012, 3,.(3)

[8] Yan R., Saha T.K.: ‘Investigation of Voltage Stability for Residential Customers

Due to High Photovoltaic Penetrations’, IEEE Trans. Power Systems, 2012, 27,.(2)

http://pv-map.apvi.org.au/analyseshttp://pv-map.apvi.org.au/analyses



26

[9] Yan R., Saha T.K.: ‘Voltage Variation Sensitivity Analysis for Unbalanced

Distribution Networks Due to Photovoltaic Power Fluctuations’, IEEE Trans. Smart

Grid, 2012, 27, (2)

[10] Alam M. J. E., Muttaqi K. M., Sutanto D.: ‘A three-phase power flow approach for

integrated 3-wire MV and 4-wire multigrounded LV networks with rooftop solar

PV’, IEEE Trans. Power Systems, 2013, 28, (2)

[11] Alam M. J. E., Muttaqi K. M., Sutanto D.: ‘A SAX-based advanced computation

tool for assessment of clustered rooftop solar PV impacts on LV and MV networks

in smart grid’, IEEE Trans. Smart Grid, 2013, 4, (1)

[12] Ranamuka D., Agalgaonkar A.P., Muttaqi K. M.: ‘Dynamic adjustment of OLTC

parameters using voltage sensitivity while utilizing DG for volt/VAr support’, IEEE

PES General Meeting, July 2014.

[13] Daratha N., Das B. and Sharma J.:’ Coordination between OLTC and SVC for

voltage regulation in unbalanced distribution system distributed generation’, IEEE

Trans. Power Systems, 2014, 29, (1)

[14] Yan R., Marais B., Saha T.K.: ‘Impacts of residential photovoltaic power fluctuation

on on-load tap changer operation and a solution using DSTATCOM’, Electric

Power Systems Research, 2014, 111

[15] Ranamuka D., Agalgaonkar A.P., Muttaqi K. M.: ‘Online voltage control in

distribution systems with multiple voltage regulating devices’, IEEE Trans.

Sustainable Energy, 2014, 5, (2)

[16] Gomez-Gonzalez M., Ruiz-Rodriguez F.J., Jurado F.: ‘A binary SFLA for

probabilistic three-phase load flow in unbalanced distribution systems with technical

constraints’, International Journal of Electrical Power and Energy Systems, 2013, 48,

(1)



27

[17] Ruiz-Rodriguez F.J., Hernández J.C., Jurado F.: ‘Probabilistic load flow for radial

distribution networks with photovoltaic generators’, IET Renewable Power

Generation, 2012, 6, (2)

[18] Ruiz-Rodriguez F.J., Hernández J.C., Jurado F.: ‘Voltage unbalance assessment in

secondary radial distribution networks with single-phase photovoltaic systems.’

International Journal of Electrical Power & Energy Systems. 2015, 64

[19] Ruiz-Rodriguez F.J., Gomez-Gonzalez M., Jurado F.: ‘Binary particle swarm

optimization for optimization of photovoltaic generators in radial distribution

systems using probabilistic load flow’, Electric Power Components and Systems,

2011, 39, (15)

[20] Hernández J.C., Medina A., Jurado F.: ‘Impact comparison of PV system integration

into rural and urban feeders.’ Energy Conversion and Management, 2008, 49, (6)

[21] Mayer J., Lillywhite J., Walker G., Mengede A., Goodwin S., Dennis G., Ingham S.,

Meredith P. and Froome C.: ‘Provisioning a 3.3 MW PhotoVoltaic Generation

System onto a Distribution Authority’s 11 kV rural feeder’, Proceedings of the 52nd

Annual Conference, Australian Solar Energy Society (Australian Solar Council)

Melbourne May 2014.

[22] ABB Group, ‘XLPE Land Cable Systems, User Guide’,

http://www02.abb.com/global/gad/gad02181.nsf/0/a8a42f36692365dcc1257a62003

101ce/$file/XLPE+Land+Cable+Systems+-+User%C2%B4s+Guide.pdf, accessed

February 2015.

[23] General Cable, ‘Electric Utility Cables – MV XLPE’,

http://www.generalcable.com.au/getattachment/c2b07e79-b6b8-442b-abf5-

0201b18237fe, accessed February 2015.

[24] General Cable, ‘ENERGY CABLES ACSR Aerial - ACSR 100mm² & below’,



28

http://www.generalcable.com.au/getattachment/82663836-89fc-4fc7-b8fd-

562b357394aa, accessed February 2015.

[25] Nexans Olex, ‘Aerial Overhead Conductors Descriptions’,

http://www.olex.com.au/Australasia/2012/OLC12641_AerialCat.pdf, accessed

February 2015.

[26] Cooper Industries, ‘Eaton Electrical Sector, Cooper Industries – Content – Library’,

http://www.cooperindustries.com/content/dam/public/powersystems/resources/librar

y/225_VoltageRegulators/S225111.pdf, accessed February 2015.

[27] Penido D.R.R., Araujo L.R.D., Carneiro Jr S., Periera J.L.R., Garcia P.A.N.: ‘Three-

Phase Power Flow Based on Four-Conductor Current Injection Method for

Unbalanced Distribution Networks’, IEEE Trans. Power Systems, 2008, 23, (2)

[28] MathWorks, ‘MATLAB’, http://au.mathworks.com/products/matlab/, accessed

February 2015.

[29] Energex, ‘Technical document updates notification service’,

https://apps.energex.com.au/service_providers/technical_docs/asp/technical_docum

ents.asp, accessed February 2015.

[30] Kersting W.H.: ‘Distribution system modeling and analysis’ (The electric power

engineering series, 2007, 2nd ed.)

[31] R. Yan, T. K. Saha, "Investigation of Voltage Imbalance Due to Distribution

Network Unbalanced Line Configurations and Load Levels", IEEE Trans. Power

Systems, 2013, 28, (2)

[32] Chen T.H., Chen M.S., Inoue T., Kotas P., Chebil E.A.: ‘Three-Phase cogenerator

and transformer modes for distribution system analysis’, IEEE Trans. Power

Delivery, 1991, 6, (4)

[33] R. Yan and T. K. Saha, “Analysis of Unbalanced Distribution Lines with Mutual



29

Coupling across Different Voltage Levels and the Corresponding Impact on

Network Voltage”, IET Generation, Transmission & Distribution, (paper in press).

DOI: 10.1049/iet-gtd.2015.0178.

[34] Chen T.H., Chang J.D.: ‘Open wye-open delta and open delta-open delta

transformer models for rigorous distribution system analysis’, IEE Proceedings-C,

1992, 139, (3)

[35] Internal documents, The University of Queensland load records

[36] Energex, ‘Power Quality Strategic Plan, Appendix 29 (2015-20)’,

https://www.aer.gov.au/sites/default/files/Energex%20-

%2029.%20Power%20Quality%20Strategic%20Plan%20-%20October%202014.pdf,


[37] The University of Queensland (UQ), ‘UQ Solar Photovoltaic Data’,

http://solar.uq.edu.au/user/reportEnergy.php, accessed February 2015.

[38] Queensland Government, ‘Electricity Regulation 2006, Electricity Act 1994,

Queensland, Australia’,

https://www.legislation.qld.gov.au/LEGISLTN/CURRENT/E/ElectricR06.pdf,


[39] Energex, ‘Energex Distribution Annual Planning Report (2014 / 2015 to 2018 /

2019)’, https://www.energex.com.au/__data/assets/pdf_file/0003/219720/DAPR-

2014-15-to-2018-19-Volume-1.pdf, accessed February 2015.

[40] IEEE Std 1159™: ‘IEEE Recommended Practice for Monitoring Electric Power

Quality’, 2009.

[41] AS 2279.4: ‘Australian Standard – Disturbances in mains supply networks –

Limitation of voltage fluctuations caused by industrial equipment’, 1991.

[42] Aurecon, ‘11 kV Connection Feasibility Investigation’ (Internal Consultation Report,



30

4 June, 2013).

[43] Energex, ‘Power Quality Strategic Plan 2015-20’ (Asset Management Division, 25

September, 2014)

[44] SUNNY CENTRAL 500CP XT / 630CP XT / 720CP XT / 760CP XT,

http://www.sma.de/en/products/solarinverters/sunny-central-500cp-xt-630cp-xt-

720cp-xt-760cp-xt.html#Downloads-14553, accessed May 2015.

[45] Australian Energy Market Commission, “National Electricity Rules”, Version 71,

9th April, 2015.

[46] Vince Garrone: ‘Technical Requirements for the Connection of a MW-scale PV

Array with Battery Storage to an 11kV Feeder in Queensland’, Asia-Pacific Solar

Research Conference, 2014

[47] X. Lu and O. Monnier. Developing Robust Power Line Communications (PLC)

with G3, Texas Instruments, 2012. http://www.ti.com/lit/wp/slay021/slay021.pdf,

accessed May 2015.

[48] Texas Instrument, Power Line Communications Kit for CENELEC Frequency Band

(TMDSPLCKITV4-CEN). http://www.ti.com/tool/TMDSPLCKITV4-

CEN?DCMP=g3plc&HQS=gma-indu-gsg-sgplc-g3plc-pr-tf1-en#buy, accessed May

2015.

Documents

Investigation of the interaction between step voltage regulators …381149/Ruifeng... · 2019. 10. 11. · This paper is published in IET Renewable Power Generation, Vol. 10, 2016.This