A Hybrid GA-Modified Harvey Model for Short-term Forecasting of Day-ahead Electricity Price and Electricity Load

Subject Areas : International Journal of Smart Electrical Engineering

Mehdi  Abroon ¹ (Department of Electrical Engineering, Hamedan Branch, Islamic Azad University, Hamedan, Iran)
Alireza  Jahangiri ² (3Department of Electrical Engineering, Hamedan Branch, Islamic Azad University, Hamedan, Iran)
Ahmad  Ghaderi Shamim ³ (3Department of Electrical Engineering, Hamedan Branch, Islamic Azad University, Hamedan, Iran)

Received: 2020-02-19 Accepted : 2020-06-06 Published : 2019-06-01

Keywords: Genetic Algorithm, Harvey Model, Electricity Price Forecasting, Electricity Load Forecasting,

Abstract :

The ability of different Harvey models has been proven for long term forecasting of time series. In this paper a new approach based on modified Harvey model tuned by genetic algorithm is proposed for short term forecasting of electricity price and electricity load. To consider the fluctuate nature of electricity price and electricity consumption, the model consists of some nonlinear terms of forecasts, which the optimal order of the nonlinear terms is determined based on T test and RMSE factor. The optimal order for hourly electricity price and hourly electricity consumption is 3 and 2 nonlinear terms, respectively. The proposed model is applied to the hourly electricity consumption and power market hourly price data for Iran from 22/12/2014-19/02/2015 using statistical analysis software EViews 5. The comparison revealed that the modified Harvey model is a very appropriate candidate for day ahead simultaneous forecasting of hourly electricity price and hourly electricity consumption