Study on Forecasting Model of Typical Power Quality Steady State Indices

Abstract

ccurate forecasting of typical power quality (PQ) steady state indices is of great significance to optimize operation mode of power grid and improve power supply quality. According to characteristics of PQ steady state indices, such as quasi-periodic and non-periodic, this paper proposes a forecasting model of typical PQ steady state indices based on chaotic theory and the least squares support vector machine (LSSVM). Firstly, it performs phase space reconstruction based on raw PQ steady state index data to form a new data space including an attractor. Secondly, LSSVM is used to train samples under high dimensional space, and particle swarm optimization (PSO) algorithm is adopted to optimize parameters of the LSSVM model to obtain the best forecasting model. Based on actual PQ monitoring data of a distribution grid, with the proposed forecasting model, mean relative error (MRE) of typical PQ steady state indices less than 8% is achieved, better than that with conventional BP neutral network forecasting method.