The Application of BP Neural Networks based on Chaotic Analysis algorithm in Short-term power Load Forecasting
Using chaotic theory, the paper analyzes the complicated time series of power load, and then concludes that the time series of power load is chaotic series. At this point a new neural network learning algorithm, chaotic analysis algorithm, is proposed in this paper. With chaotic analysis method, the paper reconfigures load series in the phase space, calculates the Branch dimension and the largest Lyapunov exponents, testes its uncertainty, and then gets its largest forecasting time scale. As a result, the number of input points in NN is determined, as well as BP neural network based on chaotic analysis algorithm is established.In this way, the algorithm overcomes the natural problem of the BP learning algorithm in feedforward NN, so that it has a good performance in convergence、 Velocity、 Error during NN training. The result shows that the BP Neural Network model based on chaotic analysis algorithm has a higher forecasting precision.
Short-term Load Forecasting (STLF) BP Neural Network chaotic analysis Phase Space Reconfiguration Theory (PSRT) Lyapunov exponent
Jia Shen Qing Zhang
Colleges of Mech.and Elec.Engineering, Agricultural University of Hebei Baoding, Hebei province, China
国际会议
杭州
英文
685-687
2006-10-12（万方平台首次上网日期，不代表论文的发表时间）