电价与负荷的相关性分析及其在电价预测中的应用

为解决传统电价预测模型需要对周末等电价波动较大预测日单独建模,以及模型不加区分地引入负荷因素影响预测精度的问题,提出了利用电价与负荷的相关系数判定是否将负荷因素引入粒子群-BP神经网络模型的新方法,将相关系数作为输入样本的阈值,判定是否在模型输入样本中引入负荷因素.在电价变化平稳、电价与负荷相关性较弱时,在电价预测模型中不引入负荷因素,解决了粒子群-BP神经网络模型由于非关联输入样本过多而影响学习效率、导致预测精度降低的问题.仿真结果表明,新的预测模型对电价相对平稳和波动较大的预测日预测精度明显提高,可用于电力市场的短期电价预测.

Correlation analysis of electricity price and load and its application to forecasting electricity price

To resolve the traditional electricity price forecasting model,which needs the weekend forecast more price volatility on separate modeling,and model without distinction introduction of the load factor affecting the forecasting accuracy problems,a new method based on PSO-BP neural network model is proposed.The method uses the correlation coefficient as threshold of the input stylebook to decide whether to put the load into the input stylebook.When the price is stable and no much relevant to the load,the load factor are not imported to the forecast model.The problem is improved based on BP neural network of price forecast,because much non-related input exist,which learning efficient is affected and forecast precision is reduced.The simulation results show that the forecast accuracy is significantly improved by using the new electricity price forecasting model on relatively stable and volatile.The method can be effectively used in the short-term electricity price forecast in the market.