ANALYSIS OF UNBALANCE ELECTRIC VEHICLE HOME CHARGING IN PEA DISTRIBUTION SYSTEM BY STOCHASTIC LOAD MODEL

Transcription

1 ANALYSIS OF UNBALANCE ELECTRIC EHICLE HOME CHARGING IN PEA DISTRIBUTION SYSTEM BY STOCHASTIC LOAD MODEL Pichai KONGTHONG Komsan HONGESOMBUT Sanchai DECHANUPAPRITTHA Provincial Electric Authority, Thailand Kasetsart University, Thailand Kasetsart University, Thailand ABSTRACT The Electrical ehicles (Es) have been continuously adoted, which not only affect to the loading but also lead to roblems on distribution system. The Es are different from general loads as they are movable and their charging eriods are quite uncertain and deend solely on drivers behavior. The voltage imact is a serious roblem of ower quality, esecially the effects of home Es charging, which is a slow charging singlehase system. This may cause the voltage unbalance of the k distribution system. This study simulates the Hua-Hin (Prachua Khiri Khan) system, by selecting 4 th feeder Hua-Hin 3 substation of PEA as the exerimental model. The Es charging behavior is modelled based on stochastic (random) variables, i.e., SOC of battery, charging time and hase connected, difference of lug-in charging, by using MATLAB, and the ower flow analysis by using DIgSILENT PowerFactory. The simulation results show not only the Es charging behaviors that cause the voltage unbalance of distribution system but also the erformance of exerimental PEA system to suort the home Es charging in the future. Finally, this study suggests the guidelines for rimary revention. INTRODUCTION The growing of consumtion in the energy from fossil fuel (such as coal, gasoline and natural gas) which is the unrenewable resource, leads to energy crisis and global warming incident. Thus, the use of renewable energy and mitigation of the greenhouse effects are the motivation and challenge asect in the world. The Electric ehicle (E) is the state-of-the-art technology that is environmental friendly and is used to increase the ower consumtion effectively. Many countries suorts in this technology including Thailand which Provincial Electricity Authority of Thailand (PEA) has initiated the study of E cars for suorting the increasing number of E cars in near future [1]. One of the most E cars on the market is single-hase household charging [], [3]. Due to the unredictability of arrival time deending on consumer s behaviour, the uncontrollable consequences may lead to serious roblems on the ower distribution system [4]. Therefore, the imact factors of the E household charging need to be analyzed, esecially the voltage unbalance roblem occurred in each duration affecting to the three-hase electrical equiment of customers and electric utilitys, such as decreasing the electric motor efficiency or reducing the transformer s aging. From the above mention, this aer investigates and analyses the E household charging esecially in the slow charging mode. The IEC : standard is used to analyse the unbalanced voltages in the ower distribution system [5]. In this study, severity of the roblem is determined by the otential to serve the E charging in PEA s system. ELECTRICAL AND E CHARGING MODEL Node Table. 1. Parameter of distribution systems , , , , , , Total to Node Total Transformer (ka) No. of Custormer (Unit) Distance 3h_185 SAC (km) 13,980 4, The Huahin 3 at the 4 th feeder distribution system in Prachua Khiri Khan Province is used to examine in this aer. This erimeter is situated in an urban area, comrising the accommodation and housing develoment of 4,731 customers, which is suitable for study the E household charging. The eak ower demand in this area is 5.31 MW at 19:15 PM. In the study system, the loads are divided into 13 grous. The conductor tye, conductor length and summation of installed transformer s rating are shown in Table 1 and Fig.1. Each grou of loads is connected with E car by the roortion of customers. To show the unbalanced voltage variation, the E enetrations are set to 10%, 0%, 30%, 40% and 50% resectively. CIRED 015 1/5

2 Arrival time and starting time The charging behavior of E is obtained by time to arrive home of E s owner. Mostly, owner romtly connects E to charge when arrives home. The robability density function of E s arrival time is reresented in normal tye as shown in Eq. (1) Where 1 ( t)/ FA ( t,, ) e ;0 t 4 (1) F A is robability of charging arrival time, t is time in a day, is mean value, is standard deviation of normal distribution. In this simulation, the mean value of arrival time is 18. The standard deviation value is varied to determine the voltage unbalance issue. This arameter is defined in (1) and () resectively. Phase connect of Es The uniform random distribution function is used in the hase connected of Es. The weighing factors of charging connected to hase A, B and C are 60%, 30% and 10%, resectively. Charging end time The battery characteristic from Nissan Leaf is used to simulate in this aer. The battery is 4 kwh caacity and is constant in charging ower at 3.4 kw [7]. When E s battery is fully charged, the maximum travel distance is almost 100 mile or 160 km. Given the remained battery SOC when connected to charge, the charging duration time can be estimated by Eq. () t SOC () Fig. 1. Distribution system and E model. Where t is charging duration time (hr), SOC is state of charge of battery (%). STOCHASTIC MODEL AND ARIABLE The simulation of charging behavior is situated on stochastic variables based contains with E s battery state-of-charge (SOC) reresents uniformly distribution, arrival time behavior of E reresents normal distribution [6]. For the behavior of connected E charging each hase of node is defined by uniformly random distribution. State of Charge (SOC) The E s SOC is remained energy in battery of E deending on customer s travel distance in each day. In this aer, the uniformly distribution remained SOC is defined in range of 0% to 80% SOC. STANDARDS AND ANALYSES The voltage unbalance in three-hase distribution system occurs when the magnitude of the voltage or the hase angle, of each hase is not equal. From requirements of the standard IEC : , the ercentage of voltage unbalance in distribution system is not exceeded % and has to consider in long-term effects that has duration more over 10 minutes. Therefore ercentage voltage unbalance in distribution systems can be calculated as following Eq. (3). n % Un 100% (3) Where %Un is ercentage of voltage unbalance, CIRED 015 /5

3 alues of Where a n is negative sequence voltage (), is ositive sequence voltage (). n and can be obtained from Eq. (4)., a 1 1 a a. b 3 1 n a a c is 110 and a is Start 1. Select Es Penetration 10%, 0%, 30%, 40% 50% of Number of customers. Uniform Random SOC 3. Normal Random Arrival time 4. Random Phase Es Penetration A=60 %, B=30%, C=10% in the system No No 6. Inut Es Profile Charing 7. Inut Customer load rofile 8. Run ower flow (Dig SILENT PowerFatory) of day (4 Hr.) Es=50% Yes Yes MATLAB Program Fig.. Flow chart of E charging simulation. (4) DigSILENT PowerFactory The E charging simulation rocedure in ower distribution system is shown in Fig.. First, MATLAB rogram is used to generate the charging rofiles of Es by using SOC stochastic variable in form of uniform random distribution, arrival time behavior in form of normal distribution and hase connected E in form of uniform distribution. When the charging rofiles is obtained, DIgSILENT Power Factory is used to analyse the ower flow followed by E s rofiles in order to determine the imact of unbalanced household slow charging mode of Es. The ositive and negative sequence voltage is used to calculate voltage unbalance incident in distribution system. For this simulation, the E charging enetration in distribution system is assigned by ercentage of customers in the system follow by 10%, 0%, 30%, 40% and 50%, resectively. In addition, standard deviation of arrival time behavior is varied for study the imact factors in the distribution system. RESULTS AND DISCUSSIONS The Es charging behavior is modeled by random variables (stochastic), consisting of random SOC of battery and hase charging by using uniform function and arrival time random by using normal function by selecting the mean value ( ) of 18 and standard deviation values ( ) of 1 and and Es enetrations of 60%, 30% and 10% at hases A, B and C, resectively. The behavior disersion of Es charging enetration affects to the ercentage of unbalance voltage, From Fig.3(a), it shows the disersion of the Es charging that is lower ( =1). The distribution system is caable for the enetration of Es charging that is only 0%. In contrast, for Fig.3(b), the disersion of the Es charging enetration is higher ( =). The ercentage of unbalance voltage exceeds % if the enetration of Es charging is u to 30%. Therefore, the coordinated charging of Es results to reduction the disersion of Es charging enetration and leads to ercentage of unbalance voltage that does not exceed the requiremnts of IEC : standard. In the total ower demand of the system, Fig.4(a) shows that the ower demand is high when the disersion of the Es charging is lower ( =1). In contrast, the disersion of the Es charging enetration that is higher ( =) as shown in Fig. 4(b) affects to the lower total ower demand. In case of voltage rofile, the disersion of the Es charging is lower ( =1). Fig.5(a) demonstrates that voltage dro at end of node is higher than that of Fig.5(b), which is the disersion of the Es charging enetration that is higher ( =). However, the voltage does not exceed from the accetable range, which is between 0.9 k to 3.1 k ( 5%) CIRED 015 3/5

4 oltage (k) oltage (k) Total Power (WM) Total Power (WM) %Unbalance oltage %Unbalance oltage 3 rd International Conference on Electricity Distribution Lyon, June Fig. 3(a). %Unbalance oltage at end node, = Fig. 3(b). %Unbalance oltage at end node, = Fig. 4(a). Total Power of the system, = Fig. 4(b). Total Power of the system, = Fig. 5(a) oltage at end node, =1 Fig. 5(b). oltage at end node, = CIRED 015 4/5

5 CONCLUSION The increasing number of electric vehicles in the near future leads to the charging of electric vehicles in residential homes that may affect load imbalance in the distribution system. Deending on the load grou of Es and behavior disersion of Es charging, the simulation study is erformed to find the occurrence of voltage unbalance, the ower caability and voltage dro in the distribution system and guidelines to solve the E s roblems of charging in the distribution system. From the simulation results, it can be found that the coordinated charging can be used to estimate the erformance of distribution system to suort the enetration of Es charging. REFERENCES [1] International Energy Association. 011, Technology Roadma: Electric and Plug-in Hybrid Electric ehicles (E/PHE), Available Source: htt:// blication/name-3851-en.html, December11, 013. [] J.Smart and S.Schey. 01, Battery vehicle driving and charging behavior observed early in the E roject SAE International ol.5 (1), 1-7. [3] International Electrotechnical Commission, 001, Electric ehicle Conductive Charging System-Part 1: General Requirements, IEC [4] R.Liu, L.Dow and E.Liu, 011, A survey of PE imacts on electric utilities, In Innovative Smart Grid Technologies (ISGT) January 011, (IEEE). Hilton Anaheim, CA, USA, 1-8. [5] International Electrotechnical Commission, 00. Electromagnetic Comatibility (EMC) Part -4: Environment Comatibility Levels in Industrial Plants for Low-Frequency Conducted Disturbance, IEC [6] J.Tan and L.Wang, 013, A stochastic model for quantifying the imact of PHEs on a residential distribution grid, In 3rd Cyber Technology in Automation, Control and Intelligent Systems (CYBER). 6-9 May 013, (IEEE). Nanjing, Chaina, [7] Power Quality Research Section, Research Division, 01, Electrical ehicle Test Reort, Provincial Electric Authority, Bangkok, Thailand, 1-34 CIRED 015 5/5