基于小波-支持向量机的出清电价预测仿真

Simulation of clear price of electricity prediction based on wavelet analysis and supporting vector machine

为了提高电力市场出清电价的预测精度,在分析其形成机理的基础上,根据加州电力市场实际运行的历史出清电价数据对其变化特点进行研究,得出不同时段的出清电价时间序列特性不同,具有不同的变化趋势.提出分时段预测出清电价的思路.采用小波分析和支持向量机工具对出清电价分时段进行预测,利用Matlab进行仿真,结果表明,该模型对系统出清电价的预测精度较高,证明此方法对加州电力市场的电价预测有很好的能力.

In order to improve the prediction accuracy of clear price in electricity market, the variation character of the data of electric power market clear price was investigated on the basis of its formation mechanism and according to the data of actual historic market clear price in California State. It was found that the time sequence feature of electric clear price would be different for different period. The change of market clear price would have different trend. So, an idea of time-stepwise prediction of market clear price was put forward. The forecast model was set up by using wavelet analysis and supporting vector machine toolbox and by using Matlab, its simulation was carried out. It was showed by its result that the prediction accuracy of clear price would be higher with this model, verifying that this method would have a good abil- ity to predict the electricity price on the electric market in California State.