基于最小二乘支持向量机和裸骨烟花算法的电力负荷短期预测

Short-term Power Load Forecasting Based on Least Square Support Vector Machine and Bare-Bone Fireworks Algorithm

作为电力系统调度控制的先行工作,短期负荷预测在电力系统中尤为重要。本文针对目前短期负荷预测工作的短板,即预测所需时间和预测精度难以兼顾的问题,提出了一种基于裸骨烟火算法(BBFWA)优化的最小二乘支持向量机(LSSVM)算法。所提算法采用BBFWA对LSSVM参数进行优化,然后基于优化后的LSSVM建立预测模型。最后算例结果表明所提算法相对其他算法,能更精确地进行电力系统短期负荷预测,且优化过程所需时间更少。

As the advance work of power system dispatching and controlling,short-term load forecasting is particularly important in power system.In this paper,aiming at the shortcoming of current short-term load forecasting,namely the difficulty in balancing the time required for forecasting and the prediction accuracy,a least squares support vector machine (LSSVM) algorithm based on bare-bone pyrotechnic algorithm (BBFWA) optimization is proposed.The proposed algorithm adopts BBFWA to optimize LSSVM parameters,and then builds a prediction model based on the optimized LSSVM.Compared with other algorithms,the proposed algorithm can predict power system short-term load more accurately and takes less time to optimize.