Optimal Electricity Charging for Smart Grid

1 Advanced Science and Technology Letters (AST 2017), Optimal Electricity Charging for Smart grid Kihoon Baek , Euna Jang, Daeho Han, Yungcheol Byun, and Hoon Kwon*. Jeju National University, Dept. of Computer Engineering, Jeju, Korea {masinogns, jej07275, dhhan, ycb, Abstract. The amount of energy consumption has been increasing over the years and the peak usage of energy is approaching to limited capacity of a power grid system. To prevent large scale black-out power outage in worst case, the consumer can control their own energy consumption, and electric vehicles are good means to control the demand of Electricity .}

2 The time when the users charge the electric vehicles is one of the important factors to avoid peak usage of energy. Even though the electric rates are decided depending on the demand which is decided according to the weather, temperature, and a variety of regional factors, there are regular patterns in the past and current rates, which means the current electric rates can be decided by referring the past data. Therefore, we proposed a method to optimally charge electric vehicles based on past real-time price information with Genetic Algorithm. For this, the past date to use and the time zone to charge electric vehicles are encoded as a binary string in a gene.

3 With a real data set collected by an Electricity power company, we tested and evaluated the proposed approach. We could effectively charge electric vehicles at a low cost with Genetic Algorithm, which means we can effectively avoid the peak usage of energy. Keywords: Electricity Charging , Optimization, Smart grid 1 Introduction The total amount of energy consumption has been increasing over the past years owing to economic development and large scale industrialization. As the demand of limited energy resources such as oil and gas grows and a total of load has been rapidly increasing, peak usage of energy is approaching to the limited capacity of the power grid system [1][2].

4 Because it can cause large scale black-out in the worst case, we can find a solution by building a number of power plants to raise the capacity limitation of the power system, which would cost much. Therefore, technologies and policies to effectively use energy as a different solution have been gaining a lot of attention. The Electricity consumer is one of the core elements consisting of Smart grid , and there is a solution for the consumer to control their own energy consumption in conjunction with energy markets, which is one of the issues nowadays.

5 Especially, the electric vehicles are good means to control the demand of Electricity from the view * Corresponding Author ISSN: 2287-1233 ASTL. Copyright 2017 SERSC. Advanced Science and Technology Letters (AST 2017). point of a consumer [3][4][5]. The time when users charge their own electric vehicles is one of the important factors to decrease energy consumption. Therefore, we propose a method to optimally charge electric vehicles in real-time electric rate plan environment based on real-time price information in the past with Genetic Algorithm (GA).

6 This paper is organized as follows. In Section 2, we briefly discuss Smart grid , electric rates plans, and Charging electric vehicles as related works. We present Optimal Electricity Charging with GA in Session 2. In Section 3, we implement the proposed approach and present experimental results. Finally, we summarize the research in Section 4. 2 Optimal Electricity Charging with GA. The current and past rates have patterns according to the seasons and dates. Also, the rates in a day tend to be similar to that of previous day during weekdays. So, a variety of hourly electric rates in real markets occur according to date and week.

7 We can find power consumption patterns or rules in the past electric rate data. For example, the electric rates are generally high in the afternoon during summer season owing to the use of air conditioners at home and in industry. On the contrary, the rates during autumn, winter, and spring season becomes low because power consumption with air conditioners decreases. Therefore, we tried to find Optimal Charging time zone with the past electric rate data based on the patterns. To charge electric vehicles relatively at a low price, Genetic Algorithm is used to find the time zone to charge [6][7].

8 The rate data to refer for optimization is automatically selected by Genetic Algorithm. According to the rate patterns discussed above, the algorithm selects the data in the same season, week, day of the week to optimize by itself, and the rate data of the previous week is used to evaluate the fitness of a gene. 3 Experimental Results A 220 volts battery charger for family use was used for the test and verification. The capacity is which is 80% of the total capacity. We assumed that four hours Charging time is required to fully charge the battery.

9 As an official data, the real-time electric rates information collected by an electric power company named ComEd Co. Ltd. at Illinois in was used as an experimental data, which was obtained every hour from March to April, 2011. As the number of generation increases, the genes evolved and the error fell down. After 7000 generations the error converged and we could find the time zone for Optimal Charging . In this experiment, off-peak time Charging , random Charging , Charging with the previous rates, and Optimal Charging with GA were tested. The electric rates to charge depending on the four methods.

10 With the proposed method, we could save 5%, , of electric rates compared to off-peak time Charging , random Charging , Charging with previous rates, respectively. Copyright 2017 SERSC 87. Advanced Science and Technology Letters (AST 2017). 4 Conclusion The electric rates are decided depending on the demand, and the demand changes according to the weather, temperature, and etc. Even though there are many kinds of environmental factors impacting on the demand, we could find regular patterns in the demand and rates from the past electric rates. In this paper, we proposed a method to optimally charge electric vehicles at a low cost based on the patterns, and we used Genetic Algorithm to search the patterns.