改进PSO-BP算法的短期电力负荷预测方法

Short-term power load prediction method using improved PSO-BP algorithm

针对电力负荷的周期性、随机波动性等复杂特点易造成预测精度低等问题,提出一种基于相似日分析、混沌映射优化粒子群算法(particle swarm optimization,PSO)和BP神经网络相结合的短期电力负荷预测方法。采用乘积法量化气象因素与时间因素间的综合相似度,选出综合相似度高的若干历史日作为相似日集;采用相似日集与非相似日集分别训练PSO-BP模型,相似日集的平均绝对百分比误差(mean absolute percentage error,MAPE)降低3.9%;利用Sine映射对PSO中粒子的速度和位置进行优化,增强PSO算法的全局搜索能力和寻优精度,采用2个集合分别训练SPSO-BP模型,相似日集的MAPE降低19.4%。结果表明,基于相似日分析和SPSO-BP模型的短期电力负荷预测方法可有效提高电力负荷的预测精度。

This paper proposes a short-term power load prediction method based on similar day analysis and chaotic mapping particle swarm optimization(PSO)algorithm and BP neural network to address the complex characteristics of power load,such as periodicity and random fluctuations,which can lead to low prediction accuracy.The product method is used to quantify the comprehensive similarity between meteorological and temporal factors,enabling the selection of several historical days with high comprehensive similarity as similar day set.The PSO-BP models are trained separately using the similar day set and the non-similar day set,resulting in a reduction of the mean absolute percentage error(MAPE)for the similar day set by 3.9%.Sine mapping is implemented to optimize the velocities and positions of particles in the PSO and enhance the global search ability and optimization accuracy of the PSO algorithm.The training of the SPSO-BP models for both datasets leads to a further reduction in the MAPE of the similar day set by 19.4%.The results show that the short-term power load prediction method based on similar daily analysis and the SPSO-BP model can effectively improve the accuracy of power load prediction.