Study Of The Distributed Power System Load Forecasting Based On Chaos Theory

The power load forecasting plays a vital role in the whole operation and control in electrical power system. Economic dispatch, hydro-thermal power coordination and power generation plan are all based on the load forecasting data. In recent years, with the arising of wind power generation, micro turbine generating and other new generating methods, distributed generators has greatly changed the traditional generation mode and transmission mode of power grid. The structure of power network becomes more and more complex and the scale continues to expand, which constantly raises the demand of power system stability. The accurate power load forecasting data can provide a solid protection for grid planning, scheduling and controlling. Because there are many influencing factors, the power load time series are complexity, nonlinearity and uncertainty. Because of these features, the load sequence is not accurately predicted by the conventional methods.Chaos is considered as an effective method of solving inherent randomness solution which in the deterministic nonlinear system. The strange attractors of chaos time series contain abundant dynamics information, and recover the chaos attractors with dynamic characteristic in the limited time series. Therefore, the chaos method is known as the best way of researching dynamics characteristic of the nonlinearity time series. Proceed from the phase space reconstruction of time sequence, this test makes use of the MATLAB, adopts power spectrum and maximum Lyapunov exponent to research the chaos of electric power system load time series which containing distributed generators, and discusses embedded dimensions of determining phase space reconstruction and different ways for delaying time parameters. Moreover, the modified C-C algorithm is introduced to the parameter selection of phase space reconstruction's power load time series for the first time. From the quantitative angle, when adopting different ways of time series forecasting predict the same time series, it will bring out different predicting precision; on the contrary, when adopting the same way of time series forecasting predicts different time series, it will bring out differences in the predicting precision. According to the calculations, the prediction accuracy of load predicting by the chaos method is satisfactory.