短期负荷预测最大Lyapunov指数预报模式预测值的判定  

Determination of Forecasted Values of Short-Term Load Forecasting Model Based on Largest Lyapunov Exponent

作　　者：杜杰[1] 陆金桂[1] 曹一家[2] 

机构地区：[1]南京工业大学自动化学院,江苏省南京市210009 [2]浙江大学电气工程学院,浙江省杭州市310027

出　　处：《电网技术》2006年第20期20-24,共5页Power System Technology

基　　金：国家自然科学基金资助项目(60225006);高等学校博士学科点专项科研基金资助项目(20030335003)~~

摘　　要：首先分析了相空间中混沌吸引子邻近轨道间的平行、交叉、折叠3种拓扑关系,根据负荷吸引子的特点提出了负荷时序最大Lyapunov指数预报模式预测值的判定依据,并探讨了相空间中临近点(轨道)的选择方法。仿真试验表明改进的负荷预测建模策略避免了原有Lyapunov指数预报模式预测值选择的盲目性,吸引子临近点的选择符合电力负荷数据的特点,所建立的短期负荷预测模型提高了预测精度并达到了预期效果。The authors fistly analyze three kinds of topological properties of the adjacent orbits including approximate parallel, overlap and fold in the phase space in this paper; then according to the feature of load attractor and based on load series, a criterion for the choice of the forecasted results from the forecasting mode using the largest Lyapunov exponent is proposed. In addition, a selection method of adjacent orbit in the phase space of load attractor is also researched. Simulation results show that the improved load forecasting modeling strategy can avoid the blindness during the choice of forecasted value in original Lyapunov exponent forecasting mode, the selection of adjacent point of the load attractor is in accordance with the feature of power load series, and the established short-term load forecasting model can improve the accuracy of forecasting.

关 键 词：短期负荷预测 最大Lyapunov指数预报模式 混沌吸引子 预测值判定 

分 类 号：TM715[电气工程—电力系统及自动化]