电力短期负荷时间序列混沌特性分析的技术改进

Technical improvement of chaotic property analysis to the power short-term load time series

电力短期负荷具有混沌特性。通过深入研究其混沌吸引子的形状、关联维数及最大Lyapunov指数,发现这些混沌特征数对电力短期负荷预测具有重要意义。在相空间重构的基础上,经研究发现电力短期负荷二维相图两坐标间延迟时间为8时,其混沌吸引子能得到最好的再现;对计算关联维数的G-P算法中关键参数的选取提出了简单易行的方法,使该算法更易于操作;对计算最大Lyapunov指数的小数据量方法做进一步改进,提出了一套固定参数,并用Henon系统数据验证了该方法的有效性。用上述改进方法进一步分析了电力短期负荷的混沌特性。

Power short-term load has chaotic property. Chaotic attractor and its correlation dimension and maximum Lyapunov exponent have important influence on power short-term load forecasting. Based on phase-space reconstruction,this paper shows two-dimensional figure of power load data, and points out that chaotic attractor can be reconstructed best when the delay time is 8 hours. An estimate method of the key parameter of G-P method is presented which makes G-P method easy to handle. The calculation method of maximum Lyapunov exponent for small data sets is improved. Furthermore, this paper puts forward a set of fixed parameters of the improved method that are tested to be valid by Henon system. With the above-mentioned methods, this paper analyses deeply the chaotic property of power system load.