基于KPCA和NSGAⅡ优化CNN参数的电动汽车充电站短期负荷预测

Short-Term Load Forecasting of Electric Vehicle Charging Station Based on KPCA and CNN Parameters Optimized by NSGAⅡ

为提升电动汽车充电站短期负荷预测的效率和精度,提出了基于核主成分分析(kernel principal component analysis,KPCA)和非劣排序遗传算法II(non-dominated sorting genetic algorithm II,NSGAII)优化卷积神经网络(convolutional neural network,CNN)的充电站短期负荷预测方法。应用KPCA对模型输入变量进行降噪处理,简化了网络结构,加快了预测速度;通过多次负荷预测测试比较误差的方式确定卷积神经网络模型中卷积层和子采样层的最佳神经元个数,保证了预测方法的准确性;利用NSGAII对卷积神经网络的参数进行优化,提高了预测方法的运算速度和预测精度。通过算例分析以及和其他方法的对比,验证了文中方法具有较高的效率和精度。

In order to improve the short-term load forecasting efficiency and precision of electric vehicle charging station,this paper proposes a short-term load forecasting method for charging station based on kernel principal component analysis（ KPCA） and non-dominated sorting genetic algorithm Ⅱ（ NSGAⅡ）. The KPCA is used to reduce the noise of the model input variables,which simplifies the network structure and accelerates the prediction speed. Through the comparison of the load forecasting error to define the convolutional neural network（ CNN） model in convolution layers and sub sampling the top layer neurons number,the accuracy of the model is guaranteed. By using the NSGAⅡ to optimize the parameters of the CNN,the operation speed and precision of the prediction method are improved. Through example analysis and comparison with other methods,it is proved that the method has high efficiency and precision.