Author(s)

Mahir Faysal¹, Md. Jahirul Islam¹, Md. Mohiuddin Murad¹, Md. Imdadul Islam², M. Ruhul Amin^3*

¹Department of Electronics and Communications Engineering, East West University, Dhaka, Bangladesh.
²Department of Computer Science and Engineering, Jahangirnagar University, Savar, Bangladesh.
³Deparment of Mathematical and Physical Sciences, East West University, Dhaka, Bangladesh.

To cope with the demand and supply of electrical load of an interconnected power system of a country, we need to forecast its demand in advance. In this paper, we use a fuzzy system to forecast electrical load on short-term basis. Here, we consider temperature, humidity, seasons of a year and time segments of a day as the parameters, which govern the demand of electrical load. For each of the parameter, we use several membership functions (MFs) and then apply the Mamdani rule on MFs and the output is determined by using the centroid method. Finally, the surface plot reveals the real scenario of the load demand. The difference between actual load and the output of the fuzzy system is found as +1.65% to -13.76%. The concept of the paper can be applied in interconnected power system of Bangladesh to reduce power loss, especially when generation is higher than the demand.

Defuzzification, Fuzzy Inference, Centroid Method, Surface Plot, APE

Faysal, M. , Islam, M. , Murad, M. , Islam, M. and Amin, M. (2019) Electrical Load Forecasting Using Fuzzy System. Journal of Computer and Communications, 7, 27-37. doi: 10.4236/jcc.2019.79003.