一种光伏功率预测的RBF神经网络校正模型

A correction model to forecast output power of photovoltaic system based on RBF neural network

提出一种基于灰色系统校正的RBF神经网络光伏功率预测模型。从提高预测精度的角度出发,本文采用具有较强拟合能力的RBF神经网络,对非理想条件下光伏出力进行预测。为了进一步提高预测精度,通过以相邻日数据为样本,构建的灰色系统模型,对光伏出力预测结果进行校正,确定最终的光伏功率预测值。通过对预测结果的比较分析,验证所提算法的准确性,减小了单独使用RBF神经网络进行预测所产生的误差。

A RBF neural network prediction model of PV output power is proposed based on correction of gray system model.To improve the prediction precision, the paper uses the RBF neural network which based on the similar day algorithm and has a great fitting ability to predict the PV output power under non ideal conditions. In order to get further improvement of the prediction precision, the gray system model is constructed with the sample of adjacent day data and the prediction results of PV output power are corrected to determine the final prediction data of PV generation. Finally, a practical example proves the method is feasible and effective. and decrease the forecast deviation when using RBF artificial neural network lonely.