Chaotic Analysis For Power Load And Prediction For Chaotic Time Series

Electric load forecasting is a quite important work in the Power System. And it is the precondition and foundation for the electric department to effectively formulate the deployment plan, reasonably arrange the generator sets’working, stopping and maintenance, and effectively ensure the power system’s steady operation. The accuracy of load forecasting directly influences power system’s security, profit and quality. Power system is a complex nonlinear system. The result that the load time series is chaotic time series can be obtained according to the analysis of the load time series. With the continuous development of the chaos theory, the analysis of time series is becoming an important research aspect that provides a scientific theory and basis for analysis and prediction of short load time series.First of all, the chaos theory, the detection of chaos and phase-space reconstruction theory are expounded. Several kinds of methods to select the parameters for phase-space reconstruction are used to determine the optimal delay time and embedding dimension. A new combined algorithm is proposed for the determination of the two parameters. Numerical simulations of several chaotic system verify that this method is applicable for determining the two parameters. And this method can recover the original phase space from the time series successfully.Then the methods to detect the chaotic time series have been studied and discussed. Power spectrum method and the largest Lyapunov exponent method are used to analyse the power load time series. Through the numerical simulation, the chaotic character of power load time series is proved.Finally, several methods to forecast the chaotic time series have been proposed and studied. The chaotic time series forecasting method based on RBF neural network are proposed. The given power load time series are the power load records of a city in China. RBF neural network forecasting method is used to forecast the load time series. Through the comparison of the forecasted load data and the actual load records, the feasibility and effectiveness of the method are verified.