Fuzzy Based Optimal Network Reconfiguration of Distribution System with Electric Vehicle Charging Stations, Distributed Generation, and Shunt Capacitors

POWER ENGINEERING AND ELECTRICAL ENGINEERING VOLUME: 21 |NUMBER: 2 |2023 |JUNE Fuzzy Based Optimal Network Reconfiguration of Distribution System with Electric Vehicle Charging Stations, Distributed Generation, and Shunt Capacitors Ajit Kumar MOHANTY , Suresh Babu PERLI Department of Electrical Engineering, National Institute of Technology, Warangal, National Institute of Technology Campus, Hanamkonda, 506004 Telengana, India [email protected], [email protected] DOI: 10.15598/aeee.v21i2.4599 Article history: Received Jun 15, 2022; Revised Sep 18, 2022; Accepted May 11, 2023; Published Jun 30, 2023. This is an open access article under the BY-CC license. Abstract. Electric Vehicles (EVs) are gaining popularity due to their low maintenance, better performance and zero carbon emission. To expand their adoption, Electric Vehicles Charging Stations (EVCS) must be integrated with the distribution system constructively to charge EVs. This study suggests an RAO-3 based on the fuzzy classification technique for the optimum EVCS, Distributed Generations (DGs), and Shunt Capacitors (SCs) sizing and positioning for 69 bus radial distribution systems with network reconfiguration. The proposed method has the following advantages (i) lower active power loss, (ii) enhanced voltage profiles, (iii) improved power factor at the substation, and (iv) optimum distribution of EVs at charging stations. Characteristic curves of Li-Ion battery charging are utilised for load flow analysis to build EV battery charging loads models. The proposed simultaneous fuzzy multi-objective study with a reconfigured network can handle the optimal number of EVs in EVCS and maintain the substation power factor at the required level, yielding an impressive distribution system performance. For example, the minimum active power loss of 18.0884 kW is achieved with a minimum voltage enhanced to 0.9905 p.u., maintaining the bus voltages at their permissible limit. The numerical results indicate that using the RAO-3 algorithm, the simultaneous technique with system reconfiguration is computationally efficient and scalable, outperforming the two-stage methodology and the method without system reconfiguration. Keywords Distributed Generators, Electric Vehicles, Electric Vehicle Charging Stations, Substation. 1. Introduction The rapid adoption of EVs in transportation in place of conventional commercial vehicles can help to minimise air pollution and fossil fuel dependency. The EVCS must be integrated with distribution networks to reduce losses and increase voltage stability [1]. In order to minimise this problem, this paper includes network reconfiguration, DGs, and SCs placement. Electrical harmonics, poor power factor, voltage instability, and imbalance are all randomly caused by the integration of EVCS into the existing grid infrastructure [2] and [3]. Many elements [4] influence the distribution system, including charging techniques, vehicle density at charging stations, etc. Experimental studies [5] and [6] are used to examine and validate the impact of EVCS on the distribution network with the objective function of initial investment cost and power quality factors. The influence of EVCS on a distributed system is investigated [7], [8] and [9] in the evolution of the power grid moving towards sustainable energy, with EVs affecting and supporting future energy growth. The impact of EVCS on the distribution system is mitigated by network reconfiguration. The distribution network’s power loss will not be minimal with variable ©2023 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 81 POWER ENGINEERING AND ELECTRICAL ENGINEERING VOLUME: 21 |NUMBER: 2 |2023 |JUNE load demand and fixed network topology. As a result, timely network reconfiguration is essential. In order to relieve overcrowded feeders and reduce real power loss in the distribution system, network reconfiguration is generally favoured [10] and [11]. On the other hand, reconfiguring the network may not meet the desired power loss reduction and power quality limitations. As a result, network reconfiguration is used in conjunction with Shunt Capacitors (SCs) or Distributed Generators (DGs) to improve voltage profile and minimise power losses and energy savings [12], [13], [14], [15] and [16]. The author of [17] developed a two-stage process for determining the best location for EVCS, DGs, and SCs. DGs and SCs are optimally situated in the initial stage. EVs are installed later in the second stage. This author does not examine the substation power factor after the installation of EVCS. Various optimisation problems can be solved using the metaheuristic technique. In [18] used a PSO method to allocate EVs in the given system. In the literature [19] and [20] JAYA and Grey Wolf Optimizer algorithms are used to integrate EVCS in the distribution systems. In this paper a robust optimisation technique known as Rao algorithms [21] is used to tackle the proposed problem. The effect of EV battery charging rates on the distribution system’s performance are investigated [22] and [23]. Most of the literature review discusses appropriate EVCS allocation in distribution systems without considering substation power factor. In this study, the electrical distribution network is reconfigured, and optimal simultaneous EVCS, DGs, and SCs are placed with a fuzzy multi-objective approach based on the RAO-3 algorithm for better distribution network performance, such as reducing active power loss, enhancing voltage profile, and keeping Substation (SN) pf at the optimal value. The performance of the RAO3 algorithm is compared with conventional algorithms like Particle Swarm Optimisation (PSO), Artificial Bee Colony (ABC), and Grey Wolf Optimiser (GWO). This work’s main points can be summed up as follows: 1. The distribution network’s EVCS, DGs, and SCs are all sized and placed optimally simultaneously with the optimal quantity of EVs. 2. In the distribution network reconfiguration, simultaneous optimal sizing and position of EVCS, DGs, and SCs with the optimal quantity of EVs. 3. In order to investigate the effect of EVCS on the distribution system’s performance, EV battery charging loads models are developed. The rest of the paper is organised in the following manner: Sec. 2. discusses the fuzzy multi-objective formulation of the problem and its constraints. Section 3. discusses the fuzzy multi-objective RAO technique. Section 4. includes the results and analyses, whereas Sec. 5. has the conclusions. 2. Problem Formulation The fuzzy-based multi-objective functions necessary for optimal deployment of EVCS, DGs, and SCs in order to improve distribution system performance are established in this section. 2.1. Substation Power Factor Membership Function: The DGs primarily run at 0.95 lagging pf; hence, the goal is to improve the Substation (SN) pf to 0.95 lagging. The following equation can be used to compute the substation power factor. pf = cos SSN kW SSN kV A ,(1) SSN kW = nbs X m=1 Pload m+Pl − ndg X n=1 PDG n,(2) SSN kV Ar = nbs X m=1 Qload m+Ql − nsc X o=1 QSC o− ndg X n=1 PDG n× ∅dg, (3) SSN kV A =qSSN kW 2+SSN kV Ar 2.(4) SSN kW and SSN kV A are the active and reactive power drawn from the substation. PDG nis the capacity of the nth DG. The total no-of DGs installations is ndg. ∅dg is the DGs units’ power factor angle. The mth node’s active power and reactive power loads are Pload m and Qload m. nbs is the total number of buses in the distribution network. Pl is the real power loss and Ql is the reactive power loss of the distribution system. The capacity rating of shunt reactive is QSC o. The total number of SCs installations is nsc. The fuzzy membership function for the SN Power-Factor (pf) [17] is depicted in Fig. (1(a)), and the mathematical expression is given in Eq. (5). ©2023 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 82 POWER ENGINEERING AND ELECTRICAL ENGINEERING VOLUME: 21 |NUMBER: 2 |2023 |JUNE δpf =                  0 for pf ≤pfmin, pf −pfmin pfs−pfmin for pfmin ≤pf ≤pfs, pfmax −pf pfmax −pfs for pfs≤pf ≤pfmax, 0 for pf ≤pfmax, (5) pfmin = 0.85, pfs= 0.95, pfmax = 1 are assumed. 2.2. DGs Penetration Membership Function: The DGP is the proportion of installed DGs to total active power load. DGP =Pndg n=1 PDG n Pnbs m=1 Pload m ,(6) The fuzzy membership function for the DGs penetration [17] is shown in Fig.1(b) and mathematical expression is given in equation Eq. (7). δDGP =                  0 for DGP ≤DGPmin, DGP −DGP min DGPs−DGP min for DGPmin ≤DGP ≤DGIs, DGPmax −DGP DGPmax −DGPs for DGPs≤DGP ≤DGPmax, 0 for DGP ≤DGPmax, (7) DGPmin = 0.4, DGPs= 0.5, DGPmax = 0.6respectively. DGPsis the desired penetration level in the distribution system. Percentage penetration is believed to be 50 %in this work. 2.3. Active Power Loss Membership Function: The following equation depicts the distribution network’s Active power Loss (AL): AL = nbs−1 X m=1 Plm,(8) Plmrepresents the branch active power loss [25], where formulated from the following equation: Plm=rm×P2 m+1 +Q2 m+1 |vm+1|2,(9) where Pm+1 is the active power load injected at the load (m+ 1) node and Qm+1 is the reactive power load. The following formula can be used to determine the active power loss index (ALX): ALX =ALDGSC ALBase ,(10) With DGs and SCs, ALDGSC denotes active power loss. ALBase denotes the real power loss in the base case. The fuzzy membership function for the real power loss [17] is depicted in Fig. 1(c) and mathematical expression is given in equation Eq. (11). ALXmax = 1. ALXminis chosen based on utility necessity so that active power loss is minimized to a desirable value. δALX =       1for ALX ≤ALXmin, ALXmax −ALX ALXmax −ALXmin for ALXmax ≤ALX ≤ALXmin, 0for ALX > ALXmax. (11) 2.4. Distribution System Voltage Membership Function: In Fig. 1(d), the fuzzy membership function of voltage [17] (δvm)of each node min the distribution system is explained, and it can be mathematically explained using Eq. (12). vld1= 0.94,vmin = 0.95,vmax = 1.05, vld2= 1.06 are assumed. The distribution system’s fuzzy voltage limit is now defined as δv= min (δvm)): δvm=                    0 for vm≤vld1, vm−vld1 vmin −vld1 for vld1< vm< vmin, 1 for vmin ≤vm≤vmax, vm−vmax vld2−vmax for vmax < vj< vld2, 0 for vm> vld2. (12) 2.5. Reconfiguration Methodology To determine the efficacy of loss reduction, the researchers’ proposed optimum network reconfiguration switching strategies must consider each feasible transition. In [11] the proposed network reconfiguration strategy is described in detail. 2.6. Optimal Allocations of EVs, DGs, and SCs Using a Multi-objective Fuzzy Function: Gzs =1 δALX +δpf +δv+δDGP .(13) The propose method is to minimise fuzzy function described in equation Eq. (13), which is exposed to various constraints: 0< PDG n≤PDG max,where n= 1,2,3,(14) 0< Qsc o≤Qsc max,where o= 1,2,3.(15) The DGs and SCs power injection at the optimal point in the distribution system are PDG nand Qsc o. ©2023 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 83 POWER ENGINEERING AND ELECTRICAL ENGINEERING VOLUME: 21 |NUMBER: 2 |2023 |JUNE 1 𝑝𝑓 𝑚𝑎𝑥 𝑝𝑓 𝑚𝑎𝑥 𝑝𝑓 s 𝑝𝑓 s 𝑝𝑓𝑚𝑖𝑛 𝑝𝑓𝑚𝑖𝑛 𝑝𝑓 δ 𝑝𝑓 δ 1 𝑝𝑓 𝑚𝑎𝑥 𝑝𝑓 s 𝑝𝑓𝑚𝑖𝑛 𝑝𝑓 δ (a) 1 𝐷𝐺𝑃 𝑚𝑎𝑥 𝐷𝐺𝑃𝑠 𝐷𝐺𝑃 𝑚𝑖𝑛 𝐷𝐺𝑃 δ 𝐷𝐺𝑃 δ 1 𝐷𝐺𝑃 𝑚𝑎𝑥 𝐷𝐺𝑃𝑠 𝐷𝐺𝑃 𝑚𝑖𝑛 𝐷𝐺𝑃 δ (b) δ 1 ALX𝑚𝑎𝑥 A𝐿𝑋𝑚𝑖𝑛 ALX δ 1 ALX𝑚𝑎𝑥 A𝐿𝑋𝑚𝑖𝑛 ALX (c) 1 𝐷𝐺𝑃 δ 𝐷𝐺𝑃 δ 1 𝑝𝑓 δ 𝑝𝑓 δ 1 𝑣 𝑚𝑖𝑛 𝑣 𝑚𝑖𝑛 𝑣𝑚𝑎𝑥 𝑣𝑚𝑎𝑥 𝑣m δ 𝑣 𝑙d1𝑙d1 𝑣 𝑙d1 𝑣 𝑙d2𝑙d2 𝑣 𝑙d2 1 𝑣 𝑚𝑖𝑛 𝑣𝑚𝑎𝑥 𝑣m δ 𝑣 𝑙d1 𝑣 𝑙d2 1 𝑣 𝑚𝑖𝑛 𝑣𝑚𝑎𝑥 𝑣m δ 𝑣 𝑙d1 𝑣 𝑙d2 1 𝐷𝐺𝑃 δ 1 𝑝𝑓 δ 1 𝑣 𝑚𝑖𝑛 𝑣𝑚𝑎𝑥 𝑣m δ 𝑣 𝑙d1 𝑣 𝑙d2 (d) Fig. 1: Fuzzy membership function. 2.7. Battery Charging Load Modelling for EVs: Equations for load flow analysis using the battery charging load model [22] can be generated from Fig. (2). Eq. (16) depicts the charging of a battery for both steady-state and transient circumstances. The following exponential equations can be used to predict the power charging properties of batteries [17]: PBEV (t) =          Pmax BEV 1−e−γ×t t2,0≤t≤t2, Pmax BEV tm−t tm−t2, t2≤t≤tm, 0, t > tm. (16) The instantaneous electric vehicle battery charging load is PBEV (t).Pmax BEV is the substation’s maximum battery charging load. δP max BEV =Pmax BEV 1−e−γ×t1 t2,(17) γ=−t2 t1ln(1 −δ),(18) t1= 0.25 h, t2= 4.5h, and tm= 5 h are in the preceding Eq. (16) and Eq. (17), respectively, taken from Fig. 2. The EV battery characteristic constants are γand δ.δis the proportion of maximum load for charging, with a value of 0.95 corresponding to 95 %of Pmax BEV at time t1. Eq. (18) (which may be derived from Eq. (17)) can be used to find the value of γ. The equation for power charging can be represented as Eq. (19). The batteries are charged from a zero-charge condition P0 BEV . PBEV (t) = Pmax BEV 1−e(−γ×t tc)+ +P0 BEV e(−γ×t tc), 0< t < tc. (19) The tcrepresents the amount of time it takes to charge a battery from its starting charging position fully. The following equation can be used to represent the status of the power charging battery. SOC(t+ 1) = SOC(t) + PBEV (t)×∆(t).(20) 3. RAO-3 Algorithm RAO-3 and RAO-1 are new optimisation algorithms [21]. It was chosen as a population-based approach for this study because of its simplicity and ease of implementation in optimization applications. It has a few control parameters. The population size is the sole control parameter that must be changed once the stop criteria are met. The proposed RAO-3 and RAO-1 algorithms make use of the worst and best solutions that can be found in Eq. (21) and Eq. (22): y′ m,p,i =ym,p,i +rand1,m.i ×(ym,b,i − |ym,w,i|) + rand2,m.i ×(|ym,p,i or ym,d,i| − (ym,d,ior ym,p,i )), (21) y′ m,p,i =ym,p,i +rand1,m.i ×(ym,b,i −ym,w,i), (22) ym,p,i is the mth variable’s value for the pth candidate in the ith iteration. The best candidate solution is denoted by ym,b,i, whereas the worst candidate solution is denoted by ym,w,i. Between exploitation and exploration, the Rao-3 algorithm can ensure optimum equilibrium. Figure 3 illustrates the suggested approach and is thoroughly explained in the steps that follow: Step 1: Read Distribution system data and run the load flow for the base case. Step 2: Initialise the algorithm parameters such as population, dimension (location and size), iteration (itr), and the maximum number of iterations (itrmax). Step 3: Randomly initialise the EVCS, DGs, and SCs location and sizes within maximum and minimum limits. Step 4: Run the load flow. Determine the fitness values for every population using Eq. (5), Eq. (6), Eq. (7), Eq. (8), Eq. (9), Eq. (10), Eq. (11), Eq. (12) and Eq. (13). Step 5: Determine the population’s best and worst solutions. Step 6: The objective function values are selected to provide the optimal solution, which is then compared to the previous one. If the ©2023 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 84 POWER ENGINEERING AND ELECTRICAL ENGINEERING VOLUME: 21 |NUMBER: 2 |2023 |JUNE 00.5 11.5 2 2.5 3 3.5 4 4.5 5 0 1 2 3 4 5 6 7 80 100 60 40 20 t (h) Charging the Battery (kW) % SOC Power ChargingPower Charging SOCSOC Power Charging SOC t1 t2 0.25 00.5 11.5 2 2.5 3 3.5 4 4.5 5 0 1 2 3 4 5 6 7 80 100 60 40 20 t (h) Charging the Battery (kW) % SOC Power Charging SOC t1 t2 0.25 Fig. 2: Li-Ion battery charging characteristics. Fig. 3: RAO-3 flow chart for the placment of EVCS, DGs, and SCs. new solution is superior to the previous one, the previous one will be replaced. Step 7: Whether the criterion is not satisfied, proceed to Step 5; otherwise, display the best optimal solution. 4. Result and Discussions In this study, a 69 bus radial distribution system [24], three bus nodes of DGs units, three bus nodes of SCs units and five EVCS bus nodes are considered. In the algorithm, parameters such as population = 100 and itrmax = 100 are assumed. Furthermore, maximum 50 EVs can be charged at each charging station are assumed. A Li-Ion battery’s maximum charging load during steady charge is 6.5 kW, according to the characteristic charging curve depicted in Fig. 2. The given system’s base values are 100 MVA and 12.66 kV. In the base case based on load flow following data, the active power demand is 3082.19 kW, the reactive power demand is 2796.77 kVAr, total real power loss is 225 kW, and the lowest voltage is 0.9092 p.u. The proposed problem is solved using the MATLAB 2022 a software installed with an Intel Core i5 8th Gen processor and 8 GB RAM. Two scenarios are studied for EVCS, DGs, and SCs in a particular distribution network to be sized and placed optimally: 4.1. Scenario 1 Fuzzy multi-objective functions described in this paper’s Eq. (13) are placed optimally using the RAO-3 method by DGs, SCs, and EVCS. The total power loss of the distribution network has decreased to 39.3467 kW and the minimum voltage was improved to 0.9762 p.u. The optimal number of EVs that can accommodate at EVCS is 193. The optimal number of EVs and EVCS are depicted in Tab. 1. Table 2 and Tab. 3 shows the DGs and SCs optimal location and size. Table 4 shows the distribution system’s performance. Figure 4 and Fig. 5 illustrate the fitness function and voltage profile curves, respectively. Figure 6 shows a single line diagram of the 69 bus radial distribution system with EVCS, DGs, and SCs from scenario 1. 010 20 30 40 50 60 70 80 90 100 0.24 0.26 0.28 0.3 0.32 0.34 0.36 0.38 0.4 0.42 JAYA RAO-3 RAO-2RAO-2 RAO-1RAO-1 TLBOTLBO JAYA RAO-3 RAO-2 RAO-1 TLBO Iteration Fitness function 010 20 30 40 50 60 70 80 90 100 0.24 0.26 0.28 0.3 0.32 0.34 0.36 0.38 0.4 0.42 JAYA RAO-3 RAO-2 RAO-1 TLBO Iteration Fitness function Fig. 4: Simultaneous placement of EVCS, DGs, and SCs fitness curve. 010 20 30 40 50 60 70 Bus 0.96 0.965 0.97 0.975 0.98 0.985 0.99 0.995 1 1.005 RAO-3 RAO-2 RAO-1 JAYA TLBO RAO-3 RAO-2 RAO-1 JAYA TLBO Voltage (p.u.) 010 20 30 40 50 60 70 Bus 0.96 0.965 0.97 0.975 0.98 0.985 0.99 0.995 1 1.005 RAO-3 RAO-2 RAO-1 JAYA TLBO Voltage (p.u.) Fig. 5: Simultaneous placement of EVCS, DGs, and SCs Voltage curve. ©2023 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 85 POWER ENGINEERING AND ELECTRICAL ENGINEERING VOLUME: 21 |NUMBER: 2 |2023 |JUNE Tab. 1: Optimum number of EVs and optimum location of EVCS. Fuzzy PSO Fuzzy ABC Fuzzy GWO Fuzzy RAO-1 Fuzzy RAO-3 Optimum node for EVs locations Optimum no-of EVs Optimum node for EVs locations Optimum no-of EVs Optimum node for EVs locations Optimum no-of EVs Optimum node for EVs locations Optimum no-of EVs Optimum node for EVs locations Optimum no-of EVs 31 24 48 34 10 31 53 38 6 38 19 38 9 41 40 35 56 37 18 45 48 32 45 33 36 36 3 33 30 35 43 31 47 36 45 36 37 50 60 30 32 42 18 33 35 43 39 30 40 45 Total no-of EVs 167 Total no-of EVs 177 Total no-of EVs 181 Total no-of EVs 188 Total no-of EVs 193 182 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 46 36 37 38 39 40 41 42 43 44 45 51 52 68 69 28 29 30 31 32 33 34 35 (72) (73) (71) (69) (70) DGDG DG SC SC SC 53 54 55 56 57 58 59 60 61 62 63 64 65 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 66 67 47 48 49 50 47 48 49 50 182 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 46 36 37 38 39 40 41 42 43 44 45 51 52 68 69 28 29 30 31 32 33 34 35 (72) (73) (71) (69) (70) DGDG DG SC SC SC 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 47 48 49 50 Fig. 6: 69 bus radial distribution system with EVCS, DGs and SCs. 4.2. Scenario 2 In this scenario, first network reconfiguration of 69 bus radial system is done.The system’s real-power loss prior to reconfiguration was 224.95 kW, and the lowest system voltage Vmin=0.9092 p.u. After network reconfiguration, the active power loss of 69 bus radial distribution network was reduced to 98.5512 kW, i.e., 56.1789 %power loss reduction, and minimum voltage is increased to Vmin = 0.94947 p.u. which is shown in Fig. 7. Table 5 represents the performances of distribution system after network reconfiguration. DGs, SCs, and EVCS are optimally positioned in the distribution system obtained from network reconfiguration. In this scenario, the distribution system’s overall power loss was decreased to 18.0884 kW and minimum voltage improves to 0.9905 p.u. The optimal number of EVs has been increased to 213 vehicles. Table 6 shows 0 10 20 30 40 50 60 70 Bus 0.9 0.91 0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.99 1 Voltage (p.u.) Before Reconfig After Reconfig Before Reconfig After Reconfig 0 10 20 30 40 50 60 70 Bus 0.9 0.91 0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.99 1 Voltage (p.u.) Before Reconfig After Reconfig Fig. 7: Voltage curve before and after reconfiguration. the performance of the distribution system, the optimal location and sizing of DGs and SCs are illustrated in Tab. 7 and Tab. 8. Table. 9 examines the distribution system’s performance. Based on the performance, RAO-3 algorithm outperforms the other conventional algorithm. Figure 8 and Fig. 9 depicts the fitness function and voltage profile curves, respectively. Figure 10 shows a single line diagram of the after-network reconfiguration of the 69 bus radial distribution system with EVCS, DGs, and SCs from scenario 2. 010 20 30 40 50 60 70 80 90 100 0.24 0.25 0.26 0.27 0.28 0.29 0.3 0.31 0.32 0.33 0.34 0.24 0.25 0.26 0.27 0.28 0.29 0.3 0.31 0.32 0.33 0.34 0.25 0.254 0.258 0.262 34 35 36 37 38 0.25 0.254 0.258 0.262 34 35 36 37 38 RAO-3 RAO-2 RAO-1 JAYA TLBO RAO-3 RAO-2 RAO-1 JAYA TLBO Iteration Fitness function 010 20 30 40 50 60 70 80 90 100 0.24 0.25 0.26 0.27 0.28 0.29 0.3 0.31 0.32 0.33 0.34 0.25 0.254 0.258 0.262 34 35 36 37 38 RAO-3 RAO-2 RAO-1 JAYA TLBO Iteration Fitness function Fig. 8: Fitness curve. According to the previous findings, scenario 2 performs better than the other scenario. Table 10 compares the outcomes of all of the scenarios. Compared to the base case, two-stage methodology [17], and scenario 1, the active power loss in scenario 2 is reduced to 91.9589 %, 55.89 %, and 54.028 %. Compared to the base case minimum voltage of 0.9092, the ©2023 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 86 POWER ENGINEERING AND ELECTRICAL ENGINEERING VOLUME: 21 |NUMBER: 2 |2023 |JUNE Tab. 2: DGs optimum location and sizing. Fuzzy PSO Fuzzy ABC Fuzzy GWO Fuzzy RAO-1 Fuzzy RAO-3 DG Node location DG sizing (kW) DG Node location DG sizing (kW) DG Node location DG sizing (kW) DG Node location DG sizing (kW) DG Node location DG sizing (kW) 63 900 19 529.4247 61 898.3838 61 806.9187 13 507.8488 23 900 61 871.2734 14 900.0000 59 591.6334 21 539.3180 6 100.7 10 500.0016 24 102.3163 19 502.1479 61 853.5330 Tab. 3: SCs optimum location and sizing. Fuzzy PSO Fuzzy ABC Fuzzy GWO Fuzzy RAO-1 Fuzzy RAO-3 SC Node location SC sizing (kVAr) SC Node location SC sizing (kVAr) SC Node location SC sizing (kVAr) SC Node location SC sizing (kVAr) SC Node location SC sizing (kVAr) 38 540.1420 32 457.3101 38 575.9351 50 511.6601 64 595.8269 58 582.9905 62 472.1094 62 519.4239 40 241.6474 69 372.0686 69 241.5578 28 455.2351 69 266.6933 12 635.2496 22 392.8228 Tab. 4: 69 bus system performance comparison. 69 bus Base case Fuzzy PSO Fuzzy ABC Fuzzy GWO Fuzzy RAO-1 Fuzzy RAO-3 SN real power (kW) 4027.19 2240.19 2223.76 2218.69 2173.09 2131.49 SN Reactive power (kVAr) 2796.77 733.7 731.01 719.2853 710.0001 702.53 SN pf 0.8214 0.95 lag 0.95 lag 0.95 lag 0.95 lag 0.95 lag DGs Penetration - 1900.7 1900.7 1900.7 1900.7 1900.7 Real Power loss (kW) 224.95 45.1027 43.9481 43.7336 41.688 39.3467 Voltage minimum (p.u.) 0.9092 0.96507 0.96605 0.96681 0.96948 0.9762 Tab. 5: Performances of distribution system after network reconfiguration. Entity Base case Fuzzy PSO Fuzzy ABC Fuzzy GWO Fuzzy RAO-1 Fuzzy RAO-3 Tie Switches 69, 70, 71, 72, 73 17, 55, 61, 69, 71 9, 17, 56, 63, 71 14, 57, 61, 69, 70 14, 57, 61, 69, 70 14, 55, 61, 69, 70 Active Power loss (kW) 224.95 115.7826 112.8772 98.6046 98.6046 98.5512 Vmin(p.u.)0.9092 0.94831 0.94831 0.94947 0.94947 0.94947 Tab. 6: Performance comparison of 69 bus system. 69 bus Base case Fuzzy PSO Fuzzy ABC Fuzzy GWO Fuzzy RAO-1 Fuzzy RAO-3 SN Active power (kW) 4027.19 3053.60 3154.76 3155.33 3257.9 3303.3 SN Reactive power (kVAr) 2796.77 1003.673 1036.919 1037.106 1070.8 1085.71 SN pf 0.8214 0.95 lag 0.95 lag 0.95 lag 0.95 lag 0.95 lag DGs Penetration - 1900.4177 1900.5 1900.6997 1900.70727 1901.69996 Real Power loss (kW) 224.95 20.06 19.64 19.09 18.2453 18.0884 Voltage minimum (p.u.) 0.9092 0.9811 0.9812 0.9852 0.9868 0.9905 Tab. 7: Optimum number of EVs and optimum location of EVCS. Fuzzy PSO Fuzzy ABC Fuzzy GWO Fuzzy RAO-1 Fuzzy RAO-3 Optimum node for EVs locations Optimum no-of EVs Optimum node for EVs locations Optimum no-of EVs Optimum node for EVs locations Optimum no-of EVs Optimum node for EVs locations Optimum no-of EVs Optimum node for EVs locations Optimum no-of EVs 19 36 28 36 31 38 39 39 42 46 42 35 44 39 38 39 38 41 27 38 36 44 30 36 28 34 37 44 32 45 38 40 47 39 37 44 28 39 7 48 48 35 25 40 35 35 36 43 58 36 Total no-of EVs 190 Total no-of EVs 190 Total no-of EVs 190 Total no-of EVs 206 Total no-of EVs 213 ©2023 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 87 POWER ENGINEERING AND ELECTRICAL ENGINEERING VOLUME: 21 |NUMBER: 2 |2023 |JUNE Tab. 8: DGs optimum location and sizing. Fuzzy PSO Fuzzy ABC Fuzzy GWO Fuzzy RAO-1 Fuzzy RAO-3 DG Node location DG sizing (kW) DG Node location DG sizing (kW) DG Node location DG sizing (kW) DG Node location DG sizing (kW) DG Node location DG sizing (kW) 61 812.6740 23 591.1429 21 535.7574 57 373.9585 61 885.1275 27 574.4756 61 850.1275 60 513.8592 61 797.9275 25 638.8911 45 513.2681 27 459.2271 61 851.0831 21 728.8167 46 377.6810 Tab. 9: SCs optimum location and sizing. Fuzzy PSO Fuzzy ABC Fuzzy GWO Fuzzy RAO-1 Fuzzy RAO-3 SC Node location SC sizing (kVAr) SC Node location SC sizing (kVAr) SC Node location SC sizing (kVAr) SC Node location SC sizing (kVAr) SC Node location SC sizing (kVAr) 50 267.0198 48 493.5362 43 388.2457 61 441.8827 27 287.9169 61 351.9917 61 353.1125 58 363.0878 57 122.4775 60 514.4220 11 414.3575 51 206.4988 55 295.7983 28 449.6673 16 195.6782 0 10 20 30 40 50 60 70 0.98 0.985 0.99 0.995 1 1.005 RAO-3 RAO-2 RAO-1 JAYA TLBO RAO-3 RAO-2 RAO-1 JAYA TLBO Bus Voltage (p.u.) 0 10 20 30 40 50 60 70 0.98 0.985 0.99 0.995 1 1.005 RAO-3 RAO-2 RAO-1 JAYA TLBO Bus Voltage (p.u.) Fig. 9: Voltage curve. 182 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 46 36 37 38 39 40 41 42 43 44 45 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 47 48 49 50 28 29 30 31 32 33 34 35 (70) (69) (14) (55) (61) DG DG DG SC SC SC Bus Sectionalize Switch Tie Switch Bus Sectionalize Switch Tie Switch 182 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 46 36 37 38 39 40 41 42 43 44 45 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 47 48 49 50 28 29 30 31 32 33 34 35 (70) (69) (14) (55) (61) DG DG DG SC SC SC Bus Sectionalize Switch Tie Switch Fig. 10: EVCS, DGs, and SCs allocation in the optimal network structure after reconfiguration. 0 1 2 3 4 5 Charge time of the battery (h) 0.99 0.992 0.994 0.996 0.998 1 1.002 1.004 1 2 3 0.99994 0.99996 0.99998 Voltage (p.u.) Node 2, EVCS 1 Node 18, EVCS 2 Node 28, EVCS 3 Node 30, EVCS 4 Node 43, EVCS 5 Node 2, EVCS 1 Node 18, EVCS 2 Node 28, EVCS 3 Node 30, EVCS 4 Node 43, EVCS 5 0 1 2 3 4 5 Charge time of the battery (h) 0.99 0.992 0.994 0.996 0.998 1 1.002 1.004 1 2 3 0.99994 0.99996 0.99998 Voltage (p.u.) Node 2, EVCS 1 Node 18, EVCS 2 Node 28, EVCS 3 Node 30, EVCS 4 Node 43, EVCS 5 Fig. 11: EVCS Voltage transients. bus’s minimum voltage is enhanced to 0.9905 p.u. and 0.9762 p.u. in scenarios 2 and 1. Compared to the two-stage methodology [17] and scenario 1, the optimal number of EVs in scenario 2 increases to 12.1 % and 10.362 %. Figure 11 depicts the impact of EVs on EVCS node voltages. It is also worth noting that, even with EV charging demand, the voltage quality may be maintained at kept at deservedly high levels due to the availability of the complete DGs capacity and SCs installations. ©2023 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 88 POWER ENGINEERING AND ELECTRICAL ENGINEERING VOLUME: 21 |NUMBER: 2 |2023 |JUNE Tab. 10: Comparison results. Cases Real Power loss (kW) Minimum Voltage (p.u.) Total number of EVs Scenario 2 18.0884 0.9905 213 Scenario 1 39.3467 0.9762 193 Two-stage Methodology [17] 41.01 0.9461 190 Base Case 224.56 0.9092 - 5. Conclusion This paper suggests an RAO-3 algorithm for 69 bus radial distribution systems with network reconfiguration based on the fuzzy classification technique for the simultaneous optimum size and positioning of EVCS, DGs, and SCs to simultaneously supply the peak of the distribution system and the EV charging load. EV battery charging Pand Qload models are built with Li-Ion characteristic curves. The proposed technique achieves its primary goal of (a) reducing active power loss, (b) enhancing the substation power factor, (c) boosting the distribution system’s voltage profile, and (d) deploying the optimum number of EVs to EVCS. The influence of transient battery charging load impacts node voltages at the EVCS, and with the help of DGs and SCs, node voltages are kept at acceptable levels during steady-state charging. The existing work can be enhanced with vehicle-to-grid technologies. Author Contributions A.K.M. performed writing, original draft, methodology and conceptualisation. S.B.P. performed review and editing. References [1] KALAMBE, S. and G. AGNIHOTRI. Loss minimization techniques used in distribution network: bibliographical survey. Renewable and Sustainable Energy Reviews. 2014, vol. 29, iss. 1, pp. 184–200. ISSN 1879-0690. DOI: 10.1016/j.rser.2013.08.075. [2] AHMADI, A., A. TAVAKOLI, P. JAMBORSALAMATI, N. REZAEI, M. R. MIVEH, F. H. GANDOMAN, A. HEIDARI and A. E. NEZHAD. Power quality improvement in smart grids using electric vehicles: a review. IET Electrical Systems in Transportation. 2019, vol. 9, iss. 1, pp. 53–64. ISSN 2042-9746. DOI: 10.1049/iet-est.2018.5023. [3] KARMAKER, A. K., S. ROY and M. R. AHMED. Analysis of the Impact of Electric Vehicle Charging Station on Power Quality Issues. In: 2019 International Conference on Electrical, Computer and Communication Engineering (ECCE). CoxsBazar: IEEE, 2019. pp. 1–6. ISBN 978-1-53869111-3. DOI: 10.1109/ECACE.2019.8679164. [4] DEB, S., K. KARUNA and P. MAHANTA. Impact of electric vehicle charging stations on reliability of distribution network In: International Conference on Technological Advancements in Power and Energy (TAP Energy). Kollam: IEEE, 2017. pp. 1–6. ISBN 978-1-5386-4021-0. DOI: 10.1109/TAPENERGY.2017.8397272. [5] PAWELEK, R., K. PAWEL and I. WASIAK. Experimental analysis of DC electric vehicles charging station operation and its impact on the supplying grid. In: IEEE International Electric Vehicle Conference (IEVC). Florence: IEEE, 2014. pp. 1–5. ISBN 978-1-4799-6075-0. DOI: 10.1109/IEVC.2014.7056152. [6] CHEN, L., C. XU, H. SONG and K. JERMSITTIPARSERT. Optimal sizing and sitting of EVCS in the distribution system using metaheuristics: A case study. Energy Reports. 2021, vol. 7, iss. 1, pp. 208–217. ISSN 2352-4847. DOI: 10.1016/j.egyr.2020.12.032. [7] LIU, Z, F. WEN and G. LEDWICH. Optimal Planning of Electric-Vehicle Charging Stations in Distribution Systems. IEEE Transactions on Power Delivery. 2013, vol. 28, iss. 1, pp. 102–110. ISSN 1937-4208. DOI: 10.1109/TPWRD.2012.2223489. [8] AWASTHI, A., K. VENKITUSAMY, S. PADMANABAN, R. SELVAMUTHUKUMARAN, F. BLAABJERG, A. K. SINGH. Optimal planning of electric vehicle charging station at the distribution system using hybrid optimization algorithm. Energy. 2017, vol. 133, iss. 1, pp. 70–80. ISSN 1873-6785. DOI: 10.1016/j.energy.2017.05.094. ©2023 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 89