LSSVM Based on PSO Algorithm to Short-Term Load Forecasting Model Research

Abstract

短期负荷预测的精度直接影响电力系统运行的可靠性和供电质量。提出一种基于粒子群优化算法的最小二乘支持向量机短期负荷预测的模型和算法，对最小二乘支持向量机的参数寻优，再以测试集误差作为判决依据，对模型参数的进行优化选择，从而提高预测精度，避免最小二乘支持向量机对经验的依赖以及预测过程中对模型参数的盲目选择。利用该模型对某电网进行负荷预测，证明该模型有较好的收敛性、较高的预测精度和较快的训练速度。 Short-term load forecasting accuracy directly affects the reliability of power system operation and power supply quality. Least squares support vector machine short-term load forecasting model based on model particle swarm optimization algorithm is proposed. The model optimizes the parameter of least squares support vector machines, with the test set error as the basis of judgment for optimal selection of the model parameters so as to improve prediction accuracy, avoid blind choice of model parameters in the forecasting process and prevent dependence on least squares support vector machine experience. We use this model to predict the loads on the grid and prove that the model has better convergence, higher accuracy and faster training speed.