电力系统负荷预测是日常生活中电力系统调度部门的一项重要工作,预测精度的高低直接影响到电力系统的安全性、经济性和供电质量.混沌理论中,负荷预测模型的建立通常由单变量时间序列的相空间重构来实现,但实际过程中往往难以确定是否包含了重构动力系统的全部信息,特别是在有限时间序列存在噪声时.因此,将单变量时间序列方法拓展到多变量时间序列中,进行多变量时间序列的相空间重构,计算了各时间序列的延迟时间和嵌入维数,建立了预测模型.研究结果表明多变量时间序列的预测效果有较大提高.%Electric short-term load forecasting is an important and integral component in the operation of any electric utility whose accuracy directly influences power system's security, profit and quality. In the theory of the chaos, the model of the load forecasting is based on the realization for the space reconstruction of single argument time series, but it is usually hard to confirm that whether the model contains all of the information in the system of dynamic reconstruction in the fact process, especially when it has unpitched sound in the limit of the time series. So the method of the single argument time series expands to the multivariable time series for the space reconstruction of multivariable chaotic time series, and the delay time and the embedding dimension in different time series are calculated, the model of the load forecasting is set up. Simulaiton results show that the prediction effect of the variables for multivariable time series is improved greatly.
展开▼
