基于3DCNN与CLSTM混合模型的短期光伏功率预测

Short-term Photovoltaic Power Prediction Based on 3DCNN and CLSTM Hybrid Model

光伏发电功率预测对于电力系统制定发电计划和协同调度至关重要。然而,由于光伏发电过程的随机性和间歇性,光伏功率预测的准确性仍有较大提升空间。为此,提出了一种基于三维卷积神经网络(three-dimensional convolutional neural network,3DCNN)和卷积长短期记忆网络(convolutional long short-term memory network,CLSTM)混合模型的光伏功率预测方法,该方法结合3DCNN和CLSTM两种神经网络模型的优势,提高了预测准确性。采用均方根误差(root mean square error,RMSE)、平均绝对误差(mean absolute error,MAE)以及平均绝对百分比误差(mean absolute percentage error,MAPE)三个指标对预测模型进行评价,通过将基于该混合模型的预测方法应用于某光伏电站的出力预测,验证了模型的适用性和正确性。结果表明,当太阳辐照强度、温度、湿度、风速等输入的时间序列相同时,基于3DCNN和CLSTM混合模型的预测效果最好,相比于单独的3DCNN模型、CLSTM模型、BP神经网络模型,混合模型的MAPE分别提高了54%、61%和64%。说明该混合模型能够更好地适应光伏发电的随机性和间歇性特点,并提高功率预测的准确性。

Photovoltaic(PV)power prediction is crucial for the formulation of power generation plans and coordinated dispatch in the power system.However,due to the stochastic and intermittent nature of PV generation,there is still significant room for improvement in the accuracy of PV power prediction.Therefore,a hybrid model based on a three-dimensional convolutional neural network(3DCNN)and convolutional long short-term memory network(CLSTM)for PV power prediction was proposed,which combines the strengths of the 3DCNN and CLSTM models to enhance prediction accuracy.The prediction model using three metrics:root mean square error(RMSE),mean absolute error(MAE),and mean absolute percentage error(MAPE)was evaluated.By applying the prediction method based on the hybrid model to the output prediction of a specific PV plant,the applicability and correctness of the model were validated.The results show that when the input time series of solar irradiance,temperature,humidity,wind speed,etc.are the same,the hybrid model based on 3DCNN and CLSTM achieves the best prediction performance.Compared to the individual 3DCNN model,CLSTM model,and backpropagation neural network(BPNN)model,the hybrid model improves MAPE by 54%,61%,and 64%,respectively.It was indicated that the hybrid model can better adapt to the stochastic and intermittent characteristics of PV generation,thus improving the accuracy of power prediction.