相空间重构和混沌神经网络融合的短期负荷预测研究

A NEW MODEL OF STLF BASED ON THE FUSION OF PSRT AND CHAOTIC NEURAL NETWORKS

该文首次提出基于PSRT和ICNN融合的电力系统STLF模型,所构造的ICNN预测模型对负荷初值和混沌轨迹的游动性有很强的敏感性,可表征复杂的动力学行为和具有全局寻优的性能,以PSRT确定ICNN输入维数,训练样本集按预测相点步进动态相轨迹和最近邻点集原理形成的,可增强预测模型对混沌动力学的联想和泛化推理能力;文中用遗传算法作为ICNN的学习算法,对两类不同负荷系统日、周预测仿真测试,证实所研究的预测模型能有效、稳定的提高预测精度,且有较高的适应能力,为将基于PSRT和ICNN融合的电力系统STLF方法用于实际运行系统在理论上取得了有效的进展。

A new model of Short-Term Load Forecasting (STLF) based on the fusion of Phase Space Reconstruction Theory(PSRT) and Improved Chaotic Neural Networks(ICNN) is presented in this paper. The ICNN model has strong sensitivity to the initial load value and to the walking of the whole chaotic track. And it can characterize the complicated dynamics behaviors and has global searching optimal ability. The input dimension of ICNN is decided by PSRT, and the training samples are formed by means of the stepping dynamic space track and nearest neighbor point set of the forecasting phase points. So it can enhance associative memory and generalization ability of forecasting model. The learning algorithm of ICNN adopts genetic algorithm. Two kinds of load systems are used to simulate, and the testing results show that the proposed model can improve effectively and stably the precision of STLF and possesses a good adaptability in the weekday and weekend. This research acquires the effective theoretic progress in the practical prediction engineering.