An Enhanced Decentralized Voltage Regulation Scheme for the Reduction of Tap Changes in HV/MV

Transformers Under High DG Penetration Georgios C. Kryonidis, Natalia V. Theologou, Andreas. I. Chrysochos,

Charis S. Demoulias, and Grigoris K. Papagiannis* School of Electrical & Computer Engineering

Aristotle University of Thessaloniki Thessaloniki, Greece

*corresponding author’s e-mail: grigoris@eng.auth.gr

Abstract—In this paper, a decentralized control scheme is proposed that achieves two objectives: To mitigate overvoltages and to reduce the frequency of tap changes in the high- /medium-voltage transformer. The proposed control is based on the reactive power capability of the distributed generation units, while actions are individually taken by each unit based only on local measurements. The validity of this method is examined through time-domain and time-series simulations. The former highlights the quick response and the robustness of the proposed control, whereas in the latter the performance and effectiveness of the proposed method in the long term is presented, focusing on the reduction of the transformer tap changes.

Index Terms--Decentralized control, distributed power generation, on-load tap changers, reactive power control, voltage control.

I. INTRODUCTION Over the last decade, policies for the promotion of

distributed generation (DG) and especially of the renewable energy sources have been adopted in national and international level, in order to improve the global sustainability, flexibility, and efficiency. As a consequence, the power delivery system has been gradually changed from a downstream unidirectional to a bidirectional scheme raising a series of technical challenges considering the smooth operation of the network, such as: Protection issues, voltage rise problems, and overloading of the network equipment [1]. Among them, voltage rise is considered as the most common problem that the distribution system operators (DSOs) have to face, significantly hindering the further increase of DG penetration [2].

Focusing on medium-voltage (MV) distribution networks, an effective solution to the voltage rise problem is the participation of DG units in the voltage regulation process by incorporating reactive power control techniques [3]-[8]. Several national grid codes have already included

decentralized control schemes, which employ the reactive power capability of the DG units as a voltage regulation means [3], [4]. The application of the Q(V) droop characteristic is a typical case of decentralized control that is preferred against other centralized- or distributed-based approaches, since no communication infrastructure is required and thus robust performance of the network is ensured [5]-[8].

The on-load tap changer (OLTC) control of the high-/medium-voltage (HV/MV) transformer was traditionally employed to regulate network voltage by controlling it at the MV busbar [9]. However, this conventional OLTC operation was affected by the advent and participation of DG in the voltage regulation process [10], [11]. Specifically, under high DG penetration, the excessive reactive power consumption of the DG units, needed for the effective overvoltage mitigation, increases the frequency of the tap changes and thus reduces the lifetime of the OLTC [12], [13]. In the literature, several centralized methods have been proposed to coordinate the operation between the OLTC and the DG units [14], [15]. However, none of these have taken into account the high sensitivity of the transformer voltage drop to reactive power variations, which significantly affect the OLTC performance and life cycle [16]-[19].

In this paper, an enhanced version of the decentralized voltage control using the Q(V) droop characteristic is proposed, which, apart from regulating efficiently the network voltage, aims to reduce the frequency of tap changes of the HV/MV transformer OLTC. This is attained by coordinating specific DG units to inject reactive power, instead of demanding it by the upstream network. The performance of the proposed method is demonstrated in an extended radial MV network and is further compared to the conventional Q(V) droop characteristic. Time-domain and time-series simulations show the improved performance of the proposed control, both in terms of quick response and in the reduction of the tap changes in the long term.

The work of G. C. Kryonidis is supported by the Research Committee ofAristotle University of Thessaloniki via a merit scholarship between 2016and 2017.

2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republising this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Citation Information: DOI: 10.1109/TSTE.2016.2556009

II. PROPOSED VOLTAGE CONTROL

The application of reactive power-based methods in MV networks ensures the effective voltage regulation, since the relatively low R/X ratio of the lines introduces a strong linkage between the grid voltage and the reactive power. Therefore, active power curtailment techniques, which can be used in low-voltage (LV) grids due to their resistive nature, are avoided and thus the DG penetration can be further increased. In the next subsections, the proposed decentralized control and its features are thoroughly presented.

A. Droop Control

Decentralized voltage regulation can be generally achieved by applying droop characteristics to the DG units. In most cases, the absorbed reactive power is calculated with respect to a local measurement, e.g. the voltage magnitude at the point of common coupling (PCC) or the injected active power. The Q(P), cosφ(P), and Q(V) droop characteristics are considered as the most well-known decentralized reactive power-based methods that have already been incorporated into several national grid codes [4]. The main drawback of the first two techniques is that the reactive power depends only on the corresponding active power injection, resulting in unnecessarily high reactive power consumption since the network operational conditions are not taken into account [7]. To address this problem, the Q(V) droop characteristic shown in Fig. 1 is adopted, where the reactive power consumption is activated when the PCC voltage exceeds a certain voltage threshold (Vth), indicating a potential overvoltage limit violation.

National grid codes do not propose any specific Q(V) characteristic, leaving this definition to the local grid connection requirements [4]. In [7], a simple but effective method for the individual determination of the Q(V) characteristic in each DG unit is proposed. Nevertheless, in this approach, all DG units participate, even those with marginal contribution to the overvoltage. This results in an unnecessary increase in the reactive power consumption.

The contribution of each DG unit to the overvoltage mitigation can be quantified by using sensitivity analysis. Assuming the radial feeder of Fig. 2, the sensitivity of voltage against reactive power can be approximated, as shown in [20], by:

min ( , )min ( , )

1

1 1m km km

iik

VX X

Q V V

(1)

where Vm is the voltage at the m-th node, Qk is the reactive

Rea

ctiv

e Po

wer

Grid Voltage0

Vth Vmax

-Qmax

Fig. 1. Q(V) droop characteristic.

1 i+1 ni-1 i

MV

bus

bar

Ri+1+iXi+1Ri+iXiR2+iX2 Ri+2+iXi+2R1+iX1

DG DGDG

Ri-1+iXi-1 Rn+iXn

0

Fig. 2. Radial feeder topology consisting of DG units and loads.

power produced by the unit located at k-th node, V is the nominal network voltage, and Xi denotes the line reactance between the (i-1)-th and i-th node. According to (1), the required reactive power for the voltage regulation of the m-th node, provided by the DG unit located at k-th node, depends on the relative position of these nodes as follows:

min( , )

min( , ) 1m k k

k m k m k m

m k mk m k m k m

Q X X

Q X X

(2)

From (2), it can be concluded that the necessary reactive power is minimized when the voltage regulation process is assigned to the DG unit located at the m-th node or further downstream the feeder. Therefore, following this important outcome, the method proposed in [7] is modified by applying the Q(V) droop characteristic to specific only DG units, when overvoltages at the PCC occur, reducing, in this way, the overall reactive power consumption in the feeder.

B. Implications with the OLTC Operation

DSOs have mainly adopted two decentralized methods concerning the OLTC operation, namely the automatic voltage regulation (AVR) and the line drop compensation (LDC). In the former, the voltage at the MV busbar is kept constant to a predefined value, while in the latter, this value is calculated with respect to the current flowing through the HV/MV transformer. Under high DG penetration, the AVR method presents a more robust performance and is preferred over the LDC method, which deteriorates due to the bidirectional power flow [21]. Furthermore, DSOs must comply with the EN 50160 Standard regarding voltage operating limits, which have been defined as ±10 % of the nominal voltage [22]. In case of MV feeders, where LV distribution networks are connected, the above limits are generally decreased to ±5 % so as to guarantee a sufficient voltage regulation margin at the LV level.

The introduction of reactive power-based control schemes in the voltage regulation process leads to an additional amount of consumed reactive power that is necessary for the effective overvoltage mitigation. Although this additional reactive power is minimized, by the above-mentioned modification in the application of the Q(V) droop characteristic to specific DG units, it must be provided by the upstream network. Moreover, this amount of reactive power is highly variable due to the intermittent nature of the renewable DG units. Therefore, due to the mainly inductive nature of the HV/MV transformer, voltage drop is very sensitive to these reactive power variations, which consequently affects the performance of the AVR method by increasing the frequency of the transformer tap changes [12].

C. Enhanced Algorithm for the Reduction of Tap Changes

The main idea behind the proposed algorithm is to limit this additional amount of reactive power flowing through the HV/MV transformer. This is succeeded by exploiting the DG units not equipped with the Q(V) characteristics in order to provide part of the reactive power consumed by the DG units participating in the voltage regulation process. In this way, both the voltage drop variations of the transformer and the frequency of the tap changes are reduced.

The proposed algorithm is presented in Fig. 3 by means of a flowchart. A decentralized approach is followed in which each DG unit regulates the reactive power flowing through the line, and thus through the transformer, based only on local measurements, namely the voltage at the PCC (VPCC) and the current of the corresponding line (Iline). In case the reactive power flowing through the line (Qline), which is calculated from the acquired measurements, exceeds the predefined threshold (Qth), the DG unit starts producing reactive power (Qprov) until, either the reactive power regulation of the line is achieved or the minimum power factor (pfmin) is reached. This process is implemented by using a proportional-integral (PI) controller to eliminate the difference between Qth and Qline. However, a time delay is deliberately introduced in each DG unit to ensure the sequential activation of the reactive power regulation process among them. The most distant DG unit will instantly react to these variations, whereas the DG unit closer to the HV/MV transformer will be the last to react. The time delay between two adjacent DG units depends on the corresponding responses, varying from few hundreds of milliseconds, in case of inverter-interfaced DG units, up to several seconds.

The second part of the algorithm refers to the reverse process of reducing the injected reactive power. When the Qline falls below the Qth, the DG unit decreases the reactive power till the Qline reaches Qth or the power factor becomes unity. In this case, the allocation of the time delays among the DG units is reversed, compared to the first part of the algorithm.

In the proposed method, the determination of the Qth in each DG unit for the activation of the reactive power regulation control can be addressed by adopting one of the following methods:

The Qth in each DG unit is equal to the reactive power flowing through the corresponding line when loads absorb their rated power and there is no active DG. In this case, the proposed algorithm is activated to reduce any reactive power consumed by the DG units participating in the overvoltage mitigation process.

Introducing power flow analysis. In such a case, power flow calculations are performed using historical generation and consumption data to define the minimum value of reactive power flowing through the corresponding line of each DG unit, which invokes a tap change at the OLTC of the HV/MV transformer.

The development of an offline optimization method. Following this, optimization techniques are employed to determine the Qth that will reduce the tap changes.

Start

Acquisition of V iPCC and I iline

Calculation of Q iline

Q iline > Q ithYES NO

i-th time delay

Increase Q iprov

Q iline = Q ith|| pf i = pf i

min

End

YES

NO

i-th time delay

Q iline = Q ith|| Q iprov = 0

End

Decrease Q iprov

YES

NO

Fig. 3. Reactive power regulation flowchart of i-th DG unit.

In this paper, the second method is selected to calculate the Qth for each DG unit. Furthermore, the determination of the DG units that will participate in the proposed control is addressed by performing a sensitivity analysis in order to calculate to what extend the reactive power injection of each DG unit contributes to voltage rise. The DG units with low sensitivity to voltage changes at the PCC with respect to reactive power injections, i.e. those located close to the HV/MV transformer, are selected for the proposed scheme.

III. SIMULATION RESULTS The performance of the proposed control is investigated in

the extended MV network of Fig. 4. In this network, only one type of DG is considered, namely PV units, while the line impedance is considered equal to 0.404 + i0.386 Ω/km for all lines. The network data regarding the rated active power of PV units and loads as well as the line lengths are presented in

ΗV(1) ΜV (2)

3 4 5 6 7 8 9 10 11 12

13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

43 44 45 46 47 48 49 50 51 52 53 54 5755 56 58 59 60 61 62

63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79

80 81 82 83 84 85 86 87 88 89 90

Fig. 4. Network topology. PV units are connected to the red nodes.

TABLE I NETWORK LINE LENGTHS

Line Number km15,23,28,29,32,37,38,41,47,49,56,69,72,76,89 0.5

7,14,19,31,34,35,39,42,44,52,55,57,61,62, 66, 71,74,75,77,79,81,84,88,90 116,18,20,21,22,26,27,30,40,43,45,46,48,51,53, 54,58,60,63,64,67,68,70,

73,78,80,83,87 1.5 3,5,6,9,13,24,25,33,36,50,59,65,82,85,86 2

17 2.54,8,10,12 3

11 4

TABLE II RATED ACTIVE POWER OF PV UNITS AND LOADS

Nodes MW Nodes MW Nodes MW Nodes MW11 1.80 84 -1.40 64,87 -0.55 70,75,77 -0.3515 0.35 8,26 0.90 67,73 -0.45 71,79,90 -0.3028 -1.30 16,21 0.50 4,9,10 1.40 3,12,14,25 0.7039 2.50 34,59 2.00 5,6,7 0.40 13,18,22,27 0.6045 -0.75 46,54 1.20 17,19,23 0.80 20,24,30,50 1.0060 3.00 47,61 -0.20 29,36,49 -1.20 37,41,48,88 -0.8063 -0.90 53,86 -0.60 31,44,57 1.50 55,69,74,78 -0.4080 -1.65 58,72 -0.25 33,81,83 -0.85 32,40,52,68,85 -0.7082 -0.10 62,76 -0.15 35,43,89 -1.00 38,42,51,56,65,66 -0.50

TABLE III

PV UNITS PARTICIPATING IN THE CONTROL SCHEMES

Voltage regulation Reactive power regulation

PV units 10, 11, 12, 23, 24, 25, 26, 27, 59, 60 3, 4, 13, 14, 15, 16, 44, 46

TABLE IV

VOLTAGE SENSITIVITY OF NODE 12 AGAINST REACTIVE POWER

PV units PV3 PV4 PV5 PV6 PV7V/kVAr 0.03 0.08 0.11 0.14 0.16

(%) 8.89 21.38 29.39 37.12 40.91PV units PV8 PV9 PV10 PV11 PV12V/kVAr 0.20 0.23 0.27 0.34 0.39

(%) 52.01 59.33 70.74 86.74 100

Tables I and II, numbered according to their connection node. Following the results obtained from the power flow and the sensitivity analysis, the PV units participating in the voltage regulation process and in the proposed reactive power control scheme are presented in the two columns of Table III, respectively. The minimum power factor of the PV units in the proposed method is 0.8 leading, while the power factor of all loads is 0.95 lagging. The HV/MV transformer connection is a 150 kV/20 kV, Dy5 with a rated power of 40 MVA, short- circuit voltage of 12 %, while the full-load losses are 0.5 %. The transformer tap range is ±8 with a voltage variation of 1.25 % per tap.

First, time-domain simulation results are presented to highlight the effectiveness and the fast response of the proposed control. Then, time-series simulations are employed to demonstrate the superior performance of the proposed control, compared to the conventional Q(V) droop control, in terms of reducing the frequency of tap changes.

A. Time-domain Simulation

Time-domain simulations are conducted for the first feeder of the examined network using the PSIM software. The simulation time step is 10 μs. In this feeder, consisting of 10 PV units, the voltage at the MV busbar is considered constant and equal to 1.05 pu. According to Table ΙΙI, PV10, PV11, and PV12 are equipped with the Q(V) droop characteristic to effectively regulate the network voltages with minimum reactive power consumption. The proposed control is applied to PV3 and PV4, since they present a very small sensitivity of voltage with respect to reactive power variations, as shown in Table IV. The time delay of the proposed control is considered equal to 0.2 s. All PV units inject their rated power, except those with variable output, as shown in Fig. 5. The reactive power of the PV units that participate in the control schemes is depicted in Fig. 6, whereas the reactive power flowing through the lines at the beginning of the feeder and the voltages at the

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20

0.5

1

1.5

2

Time (s)

Act

ive

Pow

er (

MW

)

PV6 PV8 PV9 PV10 PV11

Fig. 5. Injected active power of PV units.

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2-2

-1

0

1

2

Time (s)

Rea

ctiv

e P

ower

(M

VA

r)

PV3 PV4 PV10 PV11 PV12

Fig. 6. Reactive power of PV units.

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2-7

-5

-3

-1

1

Time (s)

Rea

ctiv

e P

ower

(M

VA

r)

Line3 Line4 Line3 threshold Line4 threshold

Fig. 7. Reactive power flowing through the lines.

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 21.05

1.1

1.15

1.2

Time (s)

Vol

tage

Mag

nitu

de (

pu)

PV10 PV11 PV12 Upper limit

Fig. 8. Voltage profile at the PCC of PV units.

last nodes are presented in Fig. 7 and Fig. 8, respectively.

Initially, all PV units operate with unity power factor and inject their rated power, resulting in excessive overvoltage at the last nodes of the feeder, as shown in Fig. 8. At 0.5 s, the Q(V) droop characteristics are activated and the last three PV units start absorbing reactive power, until the voltage is regulated to the maximum permissible limit of 1.1 pu. The damped oscillations observed in the reactive power and the PCC voltage of the PV units, occur due to the abrupt activation of the associated droop control. The reactive power consumption of these PV units increases the reactive power flowing through the lines, as shown in Fig. 7. To reduce this excessive amount of reactive power, the proposed control scheme is introduced.

At 1 s, PV4 starts injecting reactive power until the minimum power factor of 0.8 leading is reached at 1.02 s. Although the reactive power flowing through the lines is reduced, as presented in Fig. 7, it still remains much higher that the corresponding limit of Qth -1245 kVA. The negative sign indicates that the reactive power flows from the upstream network to the PV units. After a time delay of 0.2 s, PV3 is activated and injects reactive power to further reduce the reactive power flowing through the line, until the minimum power factor is reached. It is worth mentioning that during this period the network voltages remain below or equal to the maximum permissible limit, as shown in Fig. 8.

At 1.5 s, the injected active power of some PV units is reduced as depicted in Fig. 5, resulting in a decrease of the overall reactive power consumption shown in Fig. 6. Due to this, the reactive power flowing through the lines is reduced below the predefined limit. Therefore, the reverse process of reducing the injected reactive power is activated. PV3 starts instantly reducing the reactive power, while PV4 is activated after 0.2 s till the reactive power regulation of the line is achieved at 1.85 s.

B. Time-series Simulation

In this section, a time-series simulation is performed by employing the simulation tool developed in [23]. The simulation period is considered one day, while the time interval is equal to 10 min. Normalized generation and load profiles, similar to those presented in Fig. 9, are randomly distributed along the examined network. The daily load or generation profile at each bus is obtained by multiplying the

0 4 8 12 16 20 240.2 0.4 0.6 0.8

1 1.2 1.4

Time (h)

a)

Act

ive

Pow

er

4 8 12 16 20 240 0.2 0.4 0.6 0.8

1

Time (h)

Act

ive

Pow

er

b)

Commercial Residential Industrial

Cloudy

Fig. 9. Example of normalized active power profiles. (a) Load profiles and

(b) generation profile.

0 4 8 12 16 20 24-8

-7

-6

-5

-4

Time (h)

Tap

Pos

itio

n

Convetional Proposed

Fig. 10. Tap variation profile.

0 4 8 12 16 20 241.03

1.04

1.05

1.06

1.07

Time (h)

Vol

tage

Mag

intu

de

(pu

)

Conventional Proposed Upper & lower limit

Fig. 11. Voltage profile at the MV busbar.

0 4 8 12 16 20 24-3

-2

-1

0

1 a)

Time (h)

Conventional Proposed Reactive power threshold

0 4 8 12 16 20 24-3

-2

-1

0

1 b)

Time (h)

Rea

ctiv

e P

ower

(M

VA

r)

Conventional Proposed Reactive power threshold

Fig. 12. Reactive power profiles. (a) Line3 and (b) Line13.

0 4 8 12 16 20 241.04

1.06

1.08 1.10

1.12 a)

Time (h)

0 4 8 12 16 20 241.04

1.06

1.08

1.10

1.12 b)

Time (h)

Vol

tage

Mag

nitu

de (

pu)

Conventional Proposed Upper limit

Conventional Proposed Upper limit

Fig. 13. Voltage profiles. (a) Node 12 and (b) Node 27.

corresponding normalized profile with the rated power of the load or PV unit. Considering the OLTC control, the AVR method is employed in which the target voltage at the MV busbar is equal to 1.05 pu with a dead-band of ±0.007 pu. The tap variation of the HV/MV transformer and the voltage profile of the MV busbar are depicted in Fig. 10 and Fig. 11, respectively. Furthermore, the reactive power absorbed by the first two feeders is shown in Fig. 12, while the voltage profiles of the corresponding last nodes are presented in Fig. 13. Prior to the activation of the proposed control, the application of the Q(V) droop characteristics to the PV units mitigates effectively the overvoltages, as verified in Fig. 13. However, the increased penetration of PV units in conjunction with their intermittent generation results in excessive reactive power consumption with high volatility, as shown in Fig. 12. Therefore, as depicted in Fig. 11, several voltage violations at the MV busbar are observed, while the number of the HV/MV transformer tap changes in Fig. 10 is also increased to maintain the voltage within the predefined limits.

The proposed reactive power control scheme reduces effectively the reactive power flowing through the lines and thus through the transformer, as shown in Fig. 12. This in turn improves the voltage profile of the MV busbar. Therefore, according to Fig. 10, the number of tap changes is decreased by 4, relieving significantly the stress of the HV/MV transformer OLTC. Furthermore, as shown in Fig. 13, network voltages remain below the maximum permissible limit for most of the time, except for specific time-instants. These small violations occur, since the Q(V) of Fig. 1 has been used without considering the adverse effect that the proposed control may have in the total reactive power regulation.

IV. CONCLUSIONS AND FUTURE WORK In this paper, the conventional Q(V) droop characteristic is

enhanced to reduce the frequency of the OLTC tap changes in HV/MV transformers. The proposed method is based on the decentralized coordination of specific DG units to provide part of the reactive power consumed by other DG units, participating in the voltage regulation process. Therefore, no communication infrastructure is needed, while the lifetime of the OLTC is significantly improved. Time-domain and time-series simulation show the validity and the improved performance of the proposed method, compared to the conventional Q(V) control scheme.

Further work will be carried out concerning the implementation of different Q(V) droop characteristics to take into account the reactive power capability of the DG units that participate in the reactive power regulation control scheme. In this way, possible small overvoltage violations will be eliminated, improving thus the proposed method.

REFERENCES [1] R. A. R. Walling, R. Saint, R. C. Dugan, J. Burke, and L. A. Kojovic,

“Summary of distributed resources impact on power delivery system,” IEEE Trans. Power Del., vol. 23, no. 3, pp. 1636–1644, Jul. 2008.

[2] C. L. Masters, “Voltage rise: The big issue when connecting embedded generation to long 11 kV overhead lines,” Power Eng. J., vol. 16, no. 1, pp. 5-12, Feb. 2002.

[3] T. Stetz, F. Marten, and M. Braun, “Improved low voltage grid-integration of photovoltaic systems in Germany,” IEEE Trans. Sustain. Energy, vol. 4, no. 2, pp. 534–542, Apr. 2013.

[4] A. Samadi, R. Eriksson, L. Söder, B. G. Rawn, and J. C. Boemer, “Coordinated active power-dependent voltage regulation in distribution grids with PV systems,” IEEE Trans. Power Del., vol. 29, no. 3, pp. 1454–1464, June 2014.

[5] V. Calderaro, V. Galdi, F. Lamberti, and A. Piccolo, “A smart strategy for voltage control ancillary service in distribution networks,” IEEE Trans. Power Syst., vol. 30, no. 1, pp. 494–502, Jan. 2015.

[6] H.-G. Yeh, D. F. Gayme, and S. H. Low, “Adaptive VAR control for distribution circuits with photovoltaic generators,” IEEE Trans. Power Syst., vol. 27, no. 3, pp. 1656–1663, Aug. 2012.

[7] A. Samadi, E. Shayesteh, R. Eriksson, B. Rawn, and L. Söder, “Multi-objective coordinated droop-based voltage regulation in distribution grids with PV systems,” Renew. Energy, vol. 71, pp. 315–323, Nov. 2014.

[8] P. Jahangiri and D. C. Aliprantis, “Distributed Volt/VAr control by PV inverters,” IEEE Trans. Power Syst., vol. 28, no. 3, pp. 3429–3439, Aug. 2013.

[9] F. Bignucolo, R. Caldon, and V. Prandoni, “Radial MV networks voltage regulation with distribution management system coordinated controller,” Elect. Power Syst. Res., vol. 78, no. 4, pp. 634–645, Apr. 2008.

[10] D. Ranamuka, A. P. Agalgaonkar, and K. M. Muttaqi, “Online voltage control in distribution systems with multiple voltage regulating devices,” IEEE Trans. Sustain. Energy, vol. 5, no. 2, pp. 617–628, Apr. 2014.

[11] M. S. El Moursi, H. H. Zeineldin, J. L. Kirtley, and K. Alobeidli, “A dynamic master/slave reactive power-management scheme for smart grids with distributed generation,” IEEE Trans. Power Del., vol. 29, no. 3, pp. 1157–1167, Jun. 2014.

[12] P. M. S. Carvalho, P. F. Correia, and L. A. F. M. Ferreira, “Distributed reactive power generation control for voltage rise mitigation in distribution networks,” IEEE Trans. Power Syst., vol. 23, no. 2, pp. 766–772, May 2008.

[13] S. N. Salih, P. Chen, and O. Carlson, “The effect of wind power integration on the frequency of tap changes of a substation transformer,” IEEE Trans. Power Syst., vol. 28, no. 4, pp. 4320–4327, Nov. 2013.

[14] Y. P. Agalgaonkar, B. C. Pal, and R. A. Jabr, “Distribution voltage control considering the impact of PV generation on tap changers and autonomous regulators,” IEEE Trans. Power Syst., vol. 29, no. 1, pp. 182–192, Jan. 2014.

[15] R. Yan, B. Marais, and T. K. Saha, “Impacts on residential photovoltaic power fluctuation on on-load tap changer operation and a solution using DSTATCOM,” Elect. Power Syst. Res., vol. 111, pp. 185–193, June 2014.

[16] K. M. Muttaqi, An D. T. Le, M. Negnevitsky, and G. Ledwich, “A coordinated voltage control approach for coordination of OLTC, voltage regulator, and DG to regulate voltage in a distribution feeder,” IEEE Trans. Ind. Appl., vol. 51, no. 2, pp. 1239–1248, Mar. 2015.

[17] H. E. Z. Farag and E. F. El-Saadany, “A novel cooperative protocol for distributed voltage control in active distribution systems,” IEEE Trans. Power Syst., vol. 28, no. 2, pp. 1645–1656, May 2013.

[18] F. A. Viawan and D. Karlsson, “Voltage and reactive power control in systems with synchronous machine-based distributed generation,” IEEE Trans. Power Del., vol. 23, no. 2, pp. 1079–1087, Apr. 2008.

[19] F. A. Viawan, A. Sannino, and J. Daalder, “Voltage control with on-load tap changers in medium voltage feeders in presence of distributed generation,” Elect. Power Syst. Res., vol. 77, no. 10, pp. 1314–1322, Aug. 2007.

[20] M. Brenna et al., “Automatic distributed voltage control algorithm in smart grids applications,” IEEE Trans. Smart Grid, vol. 4, no. 2, pp. 877–885, June 2013.

[21] H. E. Farag, E. F. El-Saadany, and R. Seethapathy, “A two ways communication-based distributed control for voltage regulation in smart distribution feeders,” IEEE Trans. Smart Grid, vol. 3, no. 1, pp. 271–281, Mar. 2012.

[22] Voltage characteristics of electricity supplied by public distribution networks, Standard EN 50160:2010.

[23] G. C. Kryonidis et al., “A simulation tool for extended distribution grids with controlled distributed generation,” in IEEE Eindhoven PowerTech, Eindhoven, Netherlands, June 29 – July 2, 2015.