基于最优可信度的月度负荷综合最优灰色神经网络预测模型

AN OPTIMUM CREDIBILITY BASED INTEGRATED OPTIMUM GRAY NEURAL NETWORK MODEL OF MONTHLY POWER LOAD FORECASTING

月度负荷具有增长和波动二重趋势。作者首次提出以纵向历史数据为原始序列,用灰色预测模型进行增长趋势预测;以横向历史数据为原始序列,用人工神经网络模型进行波动趋势预测的方法,并在此基础上,引入最优可信度的概念,同时考虑了月度负荷的两种趋势,建立了综合最优预测模型。该模型兼顾了前两种模型的建模特点,优于只考虑单一发展趋势负荷预测的模型。对电力负荷预测应用实例的计算结果表明,该方法明显地提高了月度负荷预测的精度,也同样适用于进行周、季负荷等具有二重趋势的负荷序列的预测。

Monthly load possesses duality of increment and fluctuation. Here, it was the first time to propose a monthly load forecasting method in which taking the longitude historical data as original series the increment trend of load was forecasted by GM(1, 1) model and taking the crosswise historical data as original series the fluctuation trend of load was forecasted by artificial neural network (ANN) model. On this basis the concept of optimum credibility was led in. Taking the two trends of monthly load into account, an optimal integrated forecasting model named OGANN was built up, in which the modeling features of above-mentioned two models were considered simultaneously, and the built forecasting model is better than the load forecasting models based on single load variation trend. The case calculation results show that the proposed method can remarkably improve the accuracy of monthly load forecasting and is available for the forecasting of weekly and seasonal load series with dually trends.