Intelligent Forecasting Model for Regional Power Grid With Distributed Generation

Abstract

With the rapid growth of energy demand, environmental and energy security issues have gain more and more attention recently. Distributed generation (DG), as one of the innovation energy power modes, has been rapidly developing in recent years. Thus, accurate prediction of distributed power load is significant for the balance of power supply and demand. This paper presents an intelligent model with a focus on the prediction of distributed power generation and the analysis of regional power supply structure. First, support vector machine is implemented to predict the utilization coefficient of DG, with fruit fly and immune algorithm (FOA-IA) optimizing its parameters. In addition, combined with the power capacity, the generating capacity of DG can be calculated. Then, in order to analyze the load reduction effect of regional power with DG, a novel stage combination forecasting model with neural network and polynomial regression is adopted to forecast power supply of the main grid. It can effectively solve the defect of nonconvergence of neural network and improve the prediction accuracy. Finally, based on the load forecasting results of both DG and main grid, the peak clipping effect of DG on power grid load has been carried. A certain regional power grid with distributed wind power generation in China is taken as case study to verify the practicability and effectiveness of the proposed method. The results demonstrate that the proposed model has the characteristics of high accuracy and strong generalization.