【摘要】 根据电力负荷序列的混沌特性,提出混沌理论和蚁群优化支持向量机结合的电力系统短期负荷预测新方法,以相空间重构理论确定支持向量机的输入量个数;训练样本集由对应预测相点的最近邻相点集构成,且是按预测相点步进动态相轨迹生成;采用蚁群优化算法对支持向量机敏感参数进行优化,从而可增强预测模型对混沌动力学的联想和泛化推理能力,提高负荷预测的精度和提高预测稳定性。对某地区负荷系统日、周预测仿真测试,证明其可获得稳定的较高预测精度。更多还原

【Abstract】 A new approach of STLF (short-term load forecasting) in power systems using chaos theory and OSVM (optimal support vector machine) based on ant colony algorithm is presented according to the chaotic character of the load series.The input number of OSVM is decided by PSRT (phase space reconstruction theory).The training samples are formed by the nearest neighbor forecasting phase points,and they are generated by means of the stepping dynamic space track of the forecasting phase point.The sensitive parameters of OSVM are optimized by the ACOA (ant colony optimization algorithm),and it can enhance associative memory and generalization ability to chaotic dynamics of forecasting model,and it can improve effectively and stably the precision of STLF.The load system simulation results show the proposed model can acquire good and stable forecasting precision.更多还原