基于灰色预测和神经网络的电力系统负荷预测

负荷是电力系统运行和规划的依据，准确的负荷预测有利于提高电力系统运行的经济性和可靠性。章提出了一种基于灰色预测和神经网络组合的电力系统负荷预测方法。在灰色预测中通过对历史数据作不同的取舍并经累加生成后建立不同的模型；对于灰色预测的不同结果再使用人工神经网络进行组合。具体方法是：神经网络的输入为各种灰色模型（GM）的预测，神经网络的输出为组合预测的结果。学习样本选择与预测量量近的n个已知值，学习方法使用改进的BP算法。所提方法综合了GM预测所需原始数据少、方法简单、而神经网络具有非线性的拟合能力的特点，提高了预测精度。算例表明了所提方法是可行的和有效的。

POWER SYSTEM LOAD FORECASTING BASED UPON COMBINATION OF GREY FORECAST AND ARTIFICIAL NEURAL NETWORK

Load is the foundation of power system operation and planning. Accurate load forecasting is advantageous to improving the reliability and economic effect of power system. Based on the combination of grey forecast and artificial neural network a new method for power system load forecasting is put forward. In the grey forecasting, after differently accepting or rejecting historical data and through accumulation and generation, different models are established, then the different results of grey forecasting are combined by artificial neural network. For artificial neural network, its inputs are the forecasting results of different Grey Models and its output is the result of combination forecasting. The learning samples select n known values which are most close to the forecasted values and the learning method is modified BP algorithm. The presented method synthesizes the advantages of GM forecasting method, which is simple and needs less original data, and neural network which possesses the characteristics of nonlinear fitting, therefore the forecasting accuracy is improved. Calculation examples show that the presented method is feasible and effective.