Indonesian J our nal of Electrical Engineering and Computer Science V ol. 25, No. 1, January 2022, pp. 25 34 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v25.i1.pp25-34 25 Optimizing the effect of char ging electric v ehicles on distrib ution transf ormer using demand side management Swapna Ganapaneni, Srini v asa V arma Pinni Department of Electrical and Electronics Engineering, K oneru Lakshmaiah Education F oundation, Guntur , India Article Inf o Article history: Recei v ed Mar 14, 2021 Re vised Oct 27, 2021 Accepted No v 26, 2021 K eyw ords: Char ging Demand side management Distrib ution transformer Electric v ehicle Ov er loading Scheduling ABSTRA CT This paper mainly aims to present the demand side ma nagement (DSM) of electric v e- hicles (EVs) by considering distrib ution transformer capacity at a residential area. The objecti v e f unctions are formulate d to obtain char ging schedule for indi vidual o wner by i) minimizing the load v ariance and ii) price indicated char ging mechanism. Both the objecti v e functions prot the o wner in the follo wing w ays: i) fullling their needs, ii) minimizing o v erall char ging cost, iii) lessening the peak load, and i v) a v oiding the o v erloading of distrib ution trans former . The proposed objecti v e functions were w ork ed on a residential area with 8 houses each house with an EV connected to a 20 kV A dist rib ution transformer . The formulations were tested in LINGO platform- optimization modeling softw are for linear , nonlinear , and inte ger programming. The results obtained were compa red which gi v es good insight of EV load scheduling with- out actual price prediction. This is an open access article under the CC BY -SA license . Corresponding A uthor: Sw apna Ganapaneni Department of Electrical and Electronics Engineering, K oneru Lakshmaiah Education F oundation Green Fields, V addesw aram, Guntur District, Andhra Pradesh, 522502, India Email: sw apna@kluni v ersity .in 1. INTR ODUCTION Electric v ehicle (EV) is the one can satis fy the need of future t ransportation due to lack of enough fossil fuels which tops the demand for electrical ener gy . In such a situation if char ging and dischar ging of EVs are not handled properly will o v erloads the grid. In this re g ard char ging the EV at residential places is a k e y issue resulting in se v eral technical problems at the le v el of distrib ution transformer , demand side management (DSM) is possibly a good solution. An o v ervie w on the literature of DSM techniques is presented here. Rapid increase in day to day el ec- tricity demand, DSM helps to a v oid utilities b uilding e xtra capacity of the generation by means of decreasing the peak demand through shifting and adjusting customers electricity consumption. DSM mainly in v olv es three programs lik e ef cient ener gy management (EM), demand response (DR), ef fecti v e load management (ELM) by the customers [1]. Figure 1 represents the detailed classication of DSM. Ener gy management (EM) mainly aims to reduce ener gy consumption which automatically minim izes the ener gy cost. A good scope of ener gy management can be easily achie v ed in v arious sectors lik e industrial, commercial, agricultural and e v en in households if ener gy sa ving tips are follo wed. Proper m aintenance of boilers, steam systems, compressed air systems, motor and dri v e systems, and lightening aspects, will lo wers the ener gy usage. T ime to time audit of ener gy; a w areness and training programs; and good metering and billing systems, are the important aspects through which ef cient ener gy management is obtained [1]. Demand response (DR) is one important polic y of DSM which concentrates mainly on the pricing J ournal homepage: http://ijeecs.iaescor e .com Evaluation Warning : The document was created with Spire.PDF for Python.
26 ISSN: 2502-4752 system to manage the peak l oad. DR denotes “changes in electric usage by end-use customers from their normal consumption patterns in response to changes in the price of electricity o v er time, or to incenti v e payments designed to induce lo wer electricity use at times of high wholesale mark et prices or when system reliability is jeopardized” [2]. Figure 1. Classication of demand side management techniques T w o major classication are done in the DR programs based on the pricing system are t ime-based pricing system (TBPS) and incenti v e-based pricing system (IBPS) [3]. Direct load control (DLC), interrupt- ible/curtailable service (I/CS), emer genc y demand response program (EDRP), capacity management (CM), demand bidding (DB), ancillary s ervice mark et (ASM) are classied under IBPS whereas time of use (T oU), real time pricing (R TP), critical peak pricing (CPP) are cate gorized under TBPS. Direct load control (DLC): DLC is one approach of DR where customer’ s loads are shuts do wn by the utilities remotely on short notice for reliability problems. This is mainly e x ecuted on small consumers lik e residential and small commercial customer [4]. Interruptible/curtailable service (I/CS): It is the program where customers on curtail of their equipment gets a discount or credited on their bill when the system is under contingenc y and if the y do not agree to curtail, the y are penalized. Emer genc y demand response program (EDRP): When an e v ent occurs, emer genc y DR is a usual program to implement. In the case of reliability e v ents EDRP of fers incenti v es to customers for reducing their loads and cannot be penalized for not curtailing their load because the prices are pre-specied [5]. Capacity ma n a gement (CM): I t is a demand side resource, during contingencies it commits to reduce pre specied amount of load and penalizes the participants if the y do not c u r tail the load on instructions. Customers obliging the instructions are guaranteed to recei v e payments in e xchange. T oU: Prices are set in adv ance b ut dif fers depending on the ti mes of the day and will not reect an y adjustments to the actual conditions of the system. Consumers will not ha v e an y incenti v es for reduced consumption in electricity during peak periods and hourly metering is not required. R TP: It is also termed as dynamic pricing as prices v aries with real time conditions and reects the actual phenomenon of the system by pro viding best information about the po wer a v ailable at a location. Ener gy consumption should be measured on hourly basis as it is char ged appropriately , and customers are of fered with incenti v es for their reduced consumption of ener gy during peak periods. CPP: This is a dynamic pricing scheme where fe w peak hours are cha r ged with high prices to reduce peak demand and other time periods are char ged with normal prices, there by permits the customers to minimize their o v erall ener gy bill. Indonesian J Elec Eng & Comp Sci, V ol. 25, No. 1, January 2022: 25–34 Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 27 Ef fecti v e load management (ELM) techniques are the one where utility tries to reduce the peak con- sumption by the subsequent approaches lik e peak clipping, v alle y lling, load shifting, strate gic conserv ation, strate gic load gro wth, e xible load shape [6]. Peak clipping: It is the w ay where load is reduced during peak time by means of DLC lik e shutting do wn the equipment of the consumer . This method will not greatly inuence the entire load curv e b ut reects in reduction of load during peak period. V alle y lling: It encourages ener gy consum ption during of f peak hours by the customers o v er of fering in- centi v es lik e allo wing them to pay lo w tarif f for changing their schedule to of f peak hours, and discounts. Load shifting: This method aims to shift the load during peak hours to period where load is lessened ho we v er o v erall demand remains constant in this phenomenon. Strate gic conserv ation: It mainly dri v es to bring do wn seasonal ener gy consumption by encouraging consumers to w ards the use of ef cient de vices and appliances, decreasing w astage of ener gy . It also of fers incenti v es to consumers who adopts technological changes in their usage. Strate gic load gro wth: It mainly tries to control seasonal increase in ener gy consumption. The dealership emplo ys intelligent systems, ef fecti v e de vices and more viable sources of ener gy to reach their goals. Fle xible load shape: It includes set of acti vities and inte grated planning between concessionary and the customer render ing to the requirement of the moment. Consumer loads are modeled with the help of load limiting equipment such that there will not be much de viation in the actual load and will not disturb security issues. Upgrading the distrib ution transformer with penetration of EV in the distrib ution system is a cost e x- pensi v e and unplanned char ging of these EVs may cause the grid o v erloading. Therefore, a strate gy DR is applied to a v oid o v erloading of transformers by considering the priority of each indi vidual home and con v e- nience preference setting. Ho we v er , impact of v arying price signals is not considered in applying DR [7]. Real time optimal scheduling of a battery ener gy storage system is proposed to reduce peak load of a b uilding as an DSM technique to reduce cost of electrical ener gy in [8] and i n t e gration of r ene w able ener gy sources are not considered. Microgrid grid resources were inte grated to the Indian distrib ution system and ha v e been scheduled to reduce dependenc y on main grid and on the other hand peak loads were managed by means of e xible load shaping which is a tool of DSM minimizes the customer’ s dissatisf action e v en diminished the operation cost of micro grid [9]. PEVs char ging control is done based on a ne w distrib uted random-access approach which does not need centralized control and can be e x ecuted in real time. The w ork dif fers from the e xisting methodologies as it considers the historical data to coordinate smart agents rather than R TP [10]. Multi objecti v e formulations were done in [11] to minimize total ener gy generation and cost associated for implementing DSM such that ener gy planning w as done in a decentralized manner , PEVs char ging is shifted result ed in reduction of total emissions and sa vings in cost. Scheduling of PEVs at a b uilding g arage to reduce the peak load and ener gy cost is achie v ed in [12] by formulating an optimization model which minimizes the square of the Euclidean distance. Similarly , in a decentralized system, non-cooperati v e g ame approach is follo wed for obtaining char ging and dischar ging schedules of the batteries and a distrib uted algorithm w as de v eloped where the total ener gy char ging cost of a PEV is minimized. Ho we v er , a pricing mechanism for v ehicle to b uilding model is not proposed, impact of dis- char ging process of battery on its life is not e v aluated, rene w able ener gy sources inte gration is not considered. Dynamic pricing mechanism is one of the possible solutions to achie v e DSM w as w ork ed out in [13] e v aluated the ef fects of herding unusual participation of customers, laziness of customers, and dif ferent usage group of customers. Minimum size of the ener gy storage system is proposed in Plug in electric v ehicle char ging station supported by rene w able ener gy sources [14]. Quadratic programming (QP) and m ulti agent system (MAS) approaches were discussed and com- pared by Mets et al. in [15] reduced the peak load and v ariability in the load of a distrib ution grid. MAS pro v ed to be the best ho we v er QP results gi v e more optimal solutions. Ho we v er , EVs char ged at of f peak time can be helped to dischar ge their ener gy during peak periods back to the grid and v ehicles arri ving randomly to char ge at the w orkplace are not considered. A micro grid consis ting of wind, photo v oltaic generation, utilized the stationary plug in h ybrid electric v ehicles by de v eloping a n optimal schedule for char ging them to support the dynamic nature of rene w able resources is proposed in [16]. A no v el algorithm is proposed to char ge lar ge EV eet by predicting their load day a head satisfying grid constraints, the indi vidual requirements of the customer Optimizing the ef fect of c har ging electric vehicles on distrib ution tr ansformer using ... (Swapna Ganapaneni) Evaluation Warning : The document was created with Spire.PDF for Python.
28 ISSN: 2502-4752 lik e arri v al and departure times, by minimizing their o v erall char ging cost, de v eloping their indi vidual plans in [17]. P article sw arm optimization is implemented to optimally schedule the EVs in a coordinated manner and minimized the acti v e po wer losses compared to uncoordinated char ging of EVs for an IEEE-33 b us radial system [18]. A meta heuristic algorithm used as optimization algorithm in [19] to optimize demand side of en- hance time of use (ET OU) pricing for a commercial load demand and signicantly analyzed that the technique shifted the maximum demand from peak time to of f peak time which mi nimized the cost of electricity . Impacts of EV technology and ho w the y help the w orld’ s gro wing demand for ener gy is demonstrated in [20]. Increase in number of EVs gro ws demand for electricity and to a v oid interruptions in the grid, PV in- te gration with EV char ging station is presented thoroughly in [21]. Ener gy controller for micro grid is designed in [22] to de v elop char ging and dischar ging schedules of EVs by absorbing o v er produced electricity with the inte gration of rene w able ener gy and sha v es the peak load of the micro grid. This paper introduced an objecti v e function to minimize the load v ariance and price indicated char ging mechanism for controlled scheduling of EVs during v alle y hours. Finally results obtained in both the methods were compared and highlighted the bes t solution. The paper is outlined as follo ws: i) Section 2 consists of problem formulation, modelling of EV load; ii) Section 3 consists of tw o objecti v e functions formulations; and iii) Section 4 summarizes the results by comparing both the scheduling schemes. 2. PR OBLEM FORMULA TION The residential area under consideration is serv ed by a 20 kV A distrib ution transformer from the grid. It is ha ving 8 houses and each house with an EV as sho wn in Figure 2. Each household load is the po wer consumed for lighting, air conditioner , w ashing machine, and w ater hea ter . and not including the EV load. The household load prole is adopted from [23] which is in f act considered from the website of electric reliability council of T e xas (ERCO T), a South-Central T e xas residential area. Basic household load prole of a day is sho wn in Figure 3. 2.1. Stochastic EV load modelling As EVs char ging adds e xtra load to the distrib ution transformer there is a need to kno w about their daily tra v elling distances, esti mating initial state of char ge (SoC), starting time of char ging to balance the ener gy and to a v oid upgrading the e xisting transformer . T o account uncertainties in the beha vior of EV load, probabilistic distrib ution functions are used to estimate arri v al time of the v ehicle, distance tra v elled, initial SoC and time required to char ge its battery . As p e r national househol d tra v el surv e y (NHTS) 2009 report which pro vides complete details of transportation in US, daily distance tra v elled by an EV follo ws Lognormal distrib ution and arri v al time of the v ehicle follo ws Gaussian distrib ution functions. Distance tra v elled by most of the people is around 20-25 miles a day and more than half of the people tra v el less than 30 miles/day [23]. The tra v elled dist ance in miles per day can be approximated by Lognormal distrib ution gi v en by (1) with mean ( µ ) of 3.37 and standard de viation ( σ ) of 0.5 and it is sho wn in Figure 4. F dist ( d ) = 1 2 π e ( l nd µ ) / 2 σ 2 for d> 0 (1) % of soc j = [1 ( E C d j ) /C bat ] 100 (2) Based on the distance tra v elled initial state of char ge (SoC) of all EVs can be estimated as follo wing from (2). Where soc j is the initial soc of j th v ehicle. d j is the distance tra v elled in miles by j th v ehicle. E c is the ener gy consumed in kWh/ miles. C bat is the battery capacity in kWh. The EV model considered in this study is Niss an Leaf 2016 model, a car which solely runs on electricity with 24 kWh battery capacity and ha ving 0.28 kWh/miles consumption. Ener gy still required and time needed to char ge the battery can be obtained as from (3) and (4). E r eq j = soc f soc j η C bat (3) Where E r eq j ener gy required to ll j th v ehicle’ s battery in kWh, η is the ef cienc y of the char ger which is considered as 0.95, S oC f is the nal SoC to be attained by the end of the char ging and (4), C time = E r eq j /P (4) Indonesian J Elec Eng & Comp Sci, V ol. 25, No. 1, January 2022: 25–34 Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 29 where C time is the char ging time required to char ge battery in Hours, P is the output po wer of the char ger in kW . Since it is assumed char ging EVs is done at residential area from [24], [25], le v el 1 char ging output po wer is 1.44 kW (120 V ,12 A) and for le v el 2 it is 3.3 kW (240 V ,14 A) are considered. Most of the household o wners char ge their EVs after the y return home from w ork at 16:00 to 21:00 according to NHTS 2009 report, the randomness in connecting EVs to start char ging follo ws Gaussian distrib ution as gi v en in (6) with a mean ( η ) of 17:00 and standard de viation ( σ ) of 2.28. The distrib ution function is described as follo ws and the distrib ution curv e for arri v al time of the EV is sho wn in Figure 5. F A ( t ) = 1 σ 2 π e ( t µ ) / 2 σ 2 , for 0 <t< 24 (5) Figure 2. Residential area serv ed by distrib ution transformer under study Figure 3. Household load prole of a day T o come up with randomness in the arri v al time and distance tra v elled by the EVs random function is applied to their probability distrib ution function such that arri v al ti me and distance tra v elled by each EV are obtained. The uncontrolled load curv e which include EVs along with household load, household load and EV load are sho wn in Figure 6 and Figure 7 respecti v ely for le v el 1 and le v el 2 char ging. The uncontrolled load results from household load and EVs load when e v ery household o wner connects their EV immediately the y arri v e home. It is observ ed that from Figure 5 and Figure 6 due to uncontrolled char ging of EVs distrib ution transformer is o v erloaded for 4 hours about 25% in le v el 1 whereas for almost 2 hours to 50% in le v el 2 char ging. Optimizing the ef fect of c har ging electric vehicles on distrib ution tr ansformer using ... (Swapna Ganapaneni) Evaluation Warning : The document was created with Spire.PDF for Python.
30 ISSN: 2502-4752 Figure 4. Distrib ution of distance tra v elled in miles Figure 5. Distrib ution of arri v al time of EVs Figure 6. Uncontrolled char ging of EVs o v erloading distrib ution transformer in le v el 1 char ging Figure 7. Uncontrolled char ging of EVs o v erloading distrib ution transformer in le v el 2 char ging 3. OPTIMIZA TION PR OBLEMS FOR CONTR OLLED CHARGING OF EVS In this section tw o objecti v e function are designed based on DSM methodologies to control the char g- ing of EVs such that optimal schedule of EVs are obtained which allo ws distrib ution transformer to operate within its capacity limits. 3.1. Method 1: Minimizing load v ariance The idea behind this objecti v e function is to reduce load during peak hours, atten the load curv e which minimizes the v ariance of ener gy required in a day . Minimizing the dif ference of load between of f peak hours and peak hours helps the distrib ution transformer to function at high ef cienc y . Appropriate Scheduling of EVs achie v es this objecti v e more easily . The objecti v e formulation of DSM can be e xpressed as follo ws and EV ij is the optimization v ariable. M inimiz e τ 24 X i =1 8 X j =1 ( H ij + E V ij ) 2 (6) Where τ is the total load prole to be met in a day in kW . H ij is the j th home load prole at i th time period in kW . E V ij is the char ging po wer of the j th v ehicle at i th time period in kW . and τ = 8 X j =1   24 X i =1 H ij + E r eq j ! Indonesian J Elec Eng & Comp Sci, V ol. 25, No. 1, January 2022: 25–34 Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 31 Subjected to follo wing constraints, j soc j + td X i = ta E V ij η E r eq j i and j 0 E V ij E max j i 8 X j =1 ( H ij + E V ij ) P tr ans where S oc j is the e xisting soc of the j th v ehicle before it connects for char ging in kW . η is the ef cienc y of the char ger which is considered as 0.95. E r eq j char ging po wer needed to ll j th v ehicle’ s battery in kW . E max j is the po wer rating of the char ger in kW . P tr ans transformer load in kW . ta,td are arri v al and departure times respecti v ely . 3.2. Method 2: Price indicated char ging mechanism The char ging price of the EVs mainly depends on the uctuations in the load. Prior information about electricity price helps consumers to shift their load from peak period to of f peak period. The main aim of this objecti v e is to shift load from peak hours to of f peak hours with the help of a price indicator . Thus, this objecti v e function schedules the char ging of EVs more optimally without kno wledge of real time prices such that it minimizes the char ging cost of the EVs and pre v ents distrib ution transformer getting o v erloaded. Therefore, the objecti v e function is (7), M I N 24 X i =1 E i + 24 X i =1 8 X j =1 P ij C i (7) where C i = δ i T and E i is the total household load prole at i th hour in kW . P ij is the char ging po wer of the j th v ehicle at i th time period in kW . C i is the price indicator which is the ratio of household load at i th period to the a v erage household load of the day . δ i is the household load at i th period in kW . δ T is the a v erage household load of the day in kW . 4. SIMULA TION RESUL TS This section presents the results of the proposed optimization problems in tw o dif ferent char ging le v els. The proposed objecti v es are e v aluated in LINGO platform. The aim is to achie v e char ging plan of each v ehicle while satisfying the capacity of the distrib ution transformer capacity , minimizing the total cost of char ging EVs and lling the batteries to 90% of the SoC. The optimization is performed on a residential area with household load, EV load obtained from stochastic modelling and the total load on distrib ution transformer based on uncontrolled char ging of EVs. Here total load represents the sum of household load and EVs load. Figure 6 and Figure 7 presents the 20 kV A distrib ution transformer o v erloading condition for le v el 1 and le v el 2 respecti v ely for uncontroll ed char ging of EVs. Figure 8 and Figure 9 for le v el 1 and le v el 2 respecti v ely sho ws the same after implementing the optimization for method 1 alle viating the problem of distrib ution transformer o v erload condition while satisfying the EV o wner requirements lik e arri v al time, departure time and ensuring that the battery reaches 90% of SoC. The optimization problem in method 2 i.e. Price indicated char ging mechanism is analysed on le v el 2 char ging as more po wer is required to char ge EVs in less duration when compared to le v el 1 and resulted in signicant peaks as sho wn in Figure 7. The optimization results sho w that EVs load is scheduled such that the total load curv e is well belo w within the rated capacity of distrib ution transformer which minimises the o v erall cost of the electricity . Ho we v er of f-peak hours from 1:00 to 8:00 are turned out to be peak hours which is treated as price indicated uncontrolled char ging as sho wn in Figure 10. From the household load prole of a day sho wn in Figure 3 of f peak hours are from 1:00 to 10:00. Ov erall char ging cost of EVs are further minimised by slightly adjusting the departure time of v ehicles up to 10:00 am. Figure 11 sho ws the price indicated controlled char ging of EVs. Optimizing the ef fect of c har ging electric vehicles on distrib ution tr ansformer using ... (Swapna Ganapaneni) Evaluation Warning : The document was created with Spire.PDF for Python.
32 ISSN: 2502-4752 Figure 8. Controlled char ging of EVs in le v el-1 alle viating o v erload condition of distrib ution transformer after minimising load v ariance Figure 9. Controlled char ging of EVs in le v el-2 alle viating o v erload condition of distrib ution transformer after minimising load v ariance Figure 10. Price indicated uncontrolled char ging mechanism for le v el 2 char ging of EVs Figure 11. Price indicated controlled char ging mechanism for le v el 2 char ging of EVs 4.1. Comparison of pr oposed optimization methods f or le v el 2 char ging The household load, EV load and total load of the distrib ution transforme r when EVs are connected in le v el 2 char ging mode based on uncontrolled char ging, minimising load v ariance and price indicated char ging mechanisms are presented in the Figure 12 and Figure 13 respecti v ely . After performing the proposed opti- mization methods for DSM, without disturbing non EV load i .e household load by proper management of EVs, distrib ution transformer o v erloading problem is solv ed. From the Figure 12 it can be observ ed that alone considering household load is well belo w the l imits of the transformer b ut the total load of transformer along with household load when EVs are connected in un- controlled manner o v erloaded the distrib ution transformer from 100 to 150% of its capacity for almost 3 hours during peak hours in the night. So, to minimize the load uctuations and peak load the abo v e implemented optimizations resulted in total load of controlled char ging of EVs, total load of price indicated uncontrolled char ging of EVs, total load of price indicated controlled char ging of EVs. Out of which price indicated con- trolled char ging of EVs balanced the system v ery well in terms of the total load on the transformer and in minimizing the char ging price of the EVs. In price indicated uncontrolled char ging though the total load on distrib ution transformer is well up to the capacity of transformer , of f peak hours are loaded to full e xtent of the transformer capacity which may result in increase of char ging prices of EVs compared to price indicated controlled char ging. So, e v en though there is a slight violation in the departure times of v ehicles price indicated controlled char ging seems to be best when compared to controlled and price indicated uncontrolled char ging. Indonesian J Elec Eng & Comp Sci, V ol. 25, No. 1, January 2022: 25–34 Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 33 Ho we v er , without ha ving the kno wledge of price uctuations, not disturbing the household load while satisfy- ing EV o wner’ s requirements EVs load is scheduled in an optimal manner by adopting the transformer limits in minimizing the load v ariance optimization i.e method 1. Therefore, out of the three cases if price is not a constraint controlled char ging of EVs, Price indicated uncontrolled char ging yields the good solution to the residential area as well as to the distrib ution grid follo wing all their requirements e v en optimizing the cost of char ging EVs to most possible e xtent. Similarly , if there is minor e xibility considered in departure time of EVs abo v e all price indicated controlled char ging gi v es the best solution obe ying the limits on the transformer as well as minimizing o v erall char ging cost of EVs further proceeded to satisfy the EV o wner requirements. Figure 12. Household load, comparison of EV load of uncontrolled, controlled, price indicated uncontrolled, price indicated controlled for le v el 2 char ging of EVs Figure 13. Household load, comparison of total load of uncontrolled, controlled, price indicated uncontrolled, price indicated controlled for le v el 2 char ging of EVs 5. CONCLUSION This paper rstly presented the methodology to model the stochastic EV load and total load at a res- idential area including EV load is calculated. As uncontrolled char ging of EVs resulted in o v erloading of distrib ution transformer , demand side management techniques are implemented. The proposed optimization methods alle viated the problem of o v erloading transformer . T otal load on the distrib ution transformer is com- pared in all the approaches. It is observ ed that method one minimized the load uctuations by shifting EV load from peak hours to of f peak hours and method tw o is implemented where EVs load is scheduled during lo w price hours and no kno wledge on uctuations in real time price is needed. Both methods satised the constraint on the transformer capacity , a v oided peak load on the system, minimized the char ging cost of EVs and sched- uled them within the gi v en time limit. W e also presented the price indicated controlled char ging mechanism which further optimized the char ging price of EVs with slight de viation in departure times. REFERENCES [1] C. W . Gellings, “Ev olving practice of demand-side management, J ournal of modern power systems and clean ener gy , v ol. 5, no. 1, pp. 1–9, 2017, doi: 10.1007/s40565-016-0252-1. [2] M. C. F alv o, G. Graditi, and P . Siano, “Electric v ehicles inte gration in demand response programs, in 2014 International Symposium on P ower Electr onics, Electrical Drives, A utomation and Motion , 2014, pp. 548-553, doi: 10.1109/SPEED AM.2014.6872126. [3] P . Khaja vi, H. Abniki, and A. Arani, “The role of incenti v e based demand response programs in smart grid, in 2011 10th Interna- tional Confer ence on En vir onment and Electrical Engineering , 2011, pp. 1-4, doi: 10.1109/EEEIC.2011.5874702. [4] J. Aghaei and M.-I. Alizadeh, “Demand response in smart electricity grids equipped with rene w able ener gy sources: A re vie w , Rene wable and Sustainable Ener gy Re vie ws , v ol. 18, pp. 64–72, 2013, doi: 10.1016/j.rser .2012.09.019. [5] R. Aazami, S. Daniar , and V . T alaeizadeh, “Emer genc y demand response program modeling on po wer system reliability e v aluation, J ournal of Electrical Engineering , v ol. 23, no. 4, pp. 151–157, 2016, doi: 10.6329/CIEE.2016.4.03. [6] G. Gaur , N. Mehta, R. Khanna, and S. Kaur , “Demand side management in a smart grid en vironment, in 2017 IEEE International Confer ence on Smart Grid and Smart Cities (ICSGSC) , 2017, pp. 227-231, doi: 10.1109/ICSGSC.2017.8038581. [7] S. Shao, M. Pipattanasomporn, and S. Rahman, “Demand response as a load shaping tool in an intelligent grid with electric v ehicles, IEEE T r ansactions on Smart Grid , v ol. 2, no. 4, pp. 624–631, 2011, doi: 10.1109/TSG.2011.2164583. Optimizing the ef fect of c har ging electric vehicles on distrib ution tr ansformer using ... (Swapna Ganapaneni) Evaluation Warning : The document was created with Spire.PDF for Python.
34 ISSN: 2502-4752 [8] M . Amoasi Acquah, D. K odai ra, and S. Han, “Real-time demand side management algorithm using stochastic optimization, Ener - gies , v ol. 11, no. 5, p. 1166, 2018, doi: 10.3390/en11051166. [9] S. P aul, A. T amrakar , and N. P . P adh y , “Demand side management based optimal scheduling portfolio of a microgrid in linear pro- gramming platform, in 2018 20th National P ower Systems Confer ence (NPSC) , 2018, pp. 1-6, doi: 10.1109/NPSC.2018.8771711. [10] K. Zhou and L. Cai, A decentralized access control algorithm for PHEV char ging in smart grid, Ener gy Systems , v ol. 5, no. 4, pp. 607-626, 2014, doi: 10.1007/s12667-013-0092-2. [11] B . Ramachandran and A. Ramanathan, “Decentralized demand side management and control of pe vs connected to a smart grid, in 2015 Clemson Univer sity P ower Systems Confer ence (PSC) , 2015, pp. 1-7, doi: 10.1109/PSC.2015.7101679. [12] H. K. Nguyen and J. B. Song, “Optimal char ging and dischar ging for multiple phe vs with demand side management in v ehicle-to- b uilding, J ournal of Communications and networks , v ol. 14, no. 6, pp. 662-671, Dec. 2012, doi: 10.1109/JCN.2012.00032. [13] T . K. W ijaya, T . G. P apaioannou, X. Liu, and K. Aberer , “Ef fecti v e consumption scheduling for demand-side management in the smart grid using non-uniform participation rate, in 2013 Sustainable Internet and ICT for Sustainability (Sust ainIT) , 2013, pp. 1-8, doi: 10.1109/SustainIT .2013.6685188. [14] M. Shaaban, A. Eajal, and E. El-Saadan y , “Coordinated char ging of plug-in h ybrid electric v ehicles in s mart h ybrid A C/DC distri- b ution systems, Rene wable Ener gy , v ol. 82, pp. 92-99, 2015, doi: 10.1016/j.renene.2014.08.012. [15] K. Mets, R. D’hulst, and C. De v elder , “Comparison of intelligent char ging algorithms for electric v ehicles t o reduce peak load and demand v ariability in a distrib ution grid, J ournal of Communications and Networks , v ol. 14, no. 6, pp. 672-681, Dec. 2012, doi: 10.1109/JCN.2012.00033. [16] K. V ijay K umar and T . Bharath K umar , “Optimal scheduling of micro grid for plug-in electrical v ehicle, International J ournal of Engineering and T ec hnolo gy (U AE) , v ol. 7, no. 2.7, pp. 558-564, 2018, doi: 10.14419/ijet.v7i2.7.10882. [17] O. Sundstr ¨ om and C. Binding, “Planning electric-dri v e v ehicle char ging under constrained grid conditions, in 2010 International Confer ence on P ower System T ec hnolo gy , 2010, pp. 1-6, doi: 10.1109/PO WERCON.2010.5666620. [18] W . I. F . W . Zalnidzam, H. Mohamad, N. A. Salim, H. Mokhlis, and Z. M. Y asin, “Optimal char ging schedule coordination of electric v ehicles in smart grid, Indonesian J ournal of Electrical Engineering and Computer Science (IJEEC S) , v ol. 11, no. 1, pp. 82-89, 2018, doi: 10.11591/ijeecs.v11.i1.pp82-89. [19] M. Sulaima, N. Dahlan, Z. Y asin, N. Asari, and Z. Bohari, “Optimum enhance time of use (ET OU) for demand side electricity pricing in re gulated mark et: An implementation using e v olutionary algorithm, Indonesian J ournal of Electrical Engineering and Computer Science (IJEECS) , v ol. 8, no. 1, pp. 253-261, 2017, doi: 10.11591/ijeecs.v8.i1.pp253-261. [20] W . A. Salah et al. , “Electric v ehicle technology impacts on ener gy , International J ournal of P ower Electr onics and Drive Systems (IJPEDS) , v ol. 10, no. 1, pp. 1-9, 2019, doi: 10.11591/ijpeds.v10.i1.pp1-9. [21] A. R. Bhatti, Z. Salam, M. J. B. Abdul, and K. P . Y ee, A comprehensi v e o v ervie w of electric v ehicle char ging using rene w- able ener gy , International J ournal of P ower Electr onics and Drive Systems (IJPEDS) , v ol. 7, no. 1, pp. 114-123, 2016, doi: 10.11591/ijpeds.v7.i1.pp114-123. [22] J. Lee and G.-L. P ark, “Inte grated coordination of electric v ehicle operations and rene w able ener gy generation in a microgrid, Inter - national J ournal of Electrical and Computer Engineering (IJECE) , v ol. 7, no. 2, pp. 706-712, 2017, doi : 10.11591/ijece.v7i2.pp706- 712. [23] C. M. Af fonso and M. K ezuno vic, “Probabilistic assessment of electric v ehicle char ging demand impact on residential distrib ution transformer aging, in 2018 IEEE International Confer ence on Pr obabilistic Methods Applied to P ower Systems (PMAPS) , 2018, pp. 1-6, doi: 10.1109/PMAPS.2018.8440211. [24] C. P ang, P . Dutta, S. Kim, M. K ezuno vic, and I. Damnjano vic, “Phe vs as dynamically congurable dispersed ener gy storage for v2b uses in the smart grid, 2010, doi: 10.1049/cp.2010.0903. [25] S. S. Ragha v an and A. Khaligh, “Impact of plug-in h ybrid electric v ehicle char ging on a distrib ution netw ork in a smart grid en vironment, in 2012 IEEE PES Inno vative Smart Grid T ec hnolo gies (ISGT) , 2012, pp. 1-7, doi: 10.1109/ISGT .2012.6175632. BIOGRAPHIES OF A UTHORS Mrs. Ganapaneni Swapna w orking a s Assistant professor in the department of Electrical and Electronics Engineering at K L Deeme d to be uni v ersity(KLEF) in the area of Po wer systems control and Automation. She obtained her B.T ech and M.T ech de gree from JNTU Kakinada and currently pursuing PhD in KLEF . She is ha ving nine years of teaching e xperience. Her research interests include po wer systems dere gulation and optimal char ging of Electric v ehicles i n smart grid. She published 7 a rticles in SCI and Scopus inde x ed journals. She is af liated wit h IEEE as student member and A CM as professional member . She can be contacted at email: sw apna@kluni v ersity .in. Dr . Pinni Srini v asa V arma completed his M. T ech. from JNTU Hyderabad. He has completed his Ph.D. from JNTU Anantapur . His areas of research are Po wer System Dere gulation and Po wer Syste m Reliability . He has published 50 research papers in v arious international journals. He has written a te xtbook on Po wer Syste m Dere gulation and is publ ished by Lambe rt publishers. He has published 2 patents in the area of Po wer Systems. No w , he is w orking as Associate Professor in EEE Dept., K L Uni v ersity , Guntur , Andhra Pradesh, India. At present Dr . P S V arma is serving as Associate Dean RD, KLEF . He can be contacted at email: pinni v arma@kluni v ersity .in. Indonesian J Elec Eng & Comp Sci, V ol. 25, No. 1, January 2022: 25–34 Evaluation Warning : The document was created with Spire.PDF for Python.