基于CEEMDAN分解的GA-BP神经网络电动汽车配电网短期负荷预测

Short-term load forecasting for electric vehicles in the existing distribution network based on GA-BP neural network of CEEMDAN decomposition

短期电动汽车并网的随机性、复杂性和易受外界因素干扰,针对通过单一的负荷预测模型很难做出准确的分析和预测,提出一种基于CEEMDAN分解的GA-BP神经网络短期电动汽车配电网负荷预测的方法.方法可以充分利用分解出来的各个分解出来的IMF分量不同的特点分别对数据进行预测最后叠加.利用电动汽车配电网所给的相似日的历史数据作为输入参数进行训练所建立的模型来预测次日的发电量.该方法适用于短期电动汽车配电网负荷预测,能有效减小误差,具有一定的参考价值.

Because of the randomness,complexity and vulnerability to disturbance by external factors in the existing distribution network,it is difficult to make an accurate analysis and prediction through a single load forecasting model for the short-term load forecasting for electric vehicles. In this paper,a method of short-term load forecasting for electric vehicles in the existing distribution network based on GA-BP neural network of CEEMDAN decomposition is proposed,which can make full use of the different characteristics of each IMF component by decomposition to predict the final superposition of the data. Using the historical data of similar days given by the electric-vehicle distribution network as input parameters,the model established by training is used to predict the power generation of the next day. This method is suitable for the short-term load forecasting in the existing distribution network for the electric vehicles,because it effectively reduces errors and has certain reference value.