基于遗传算法优化小波神经网络的短期光伏功率预测研究

Short-term Photovoltaic Power Prediction Based on Wavelet Neural Network Optimized by Genetic Algorithm

针对西北地区对光伏功率预测准确性考核要求的提升,提出了一种基于遗传算法优化小波神经网络的短期光伏功率预测模型,利用遗传算法优化小波神经网络,提高模型的预测精度。首先,对原始数据做数据预处理,得到比较理想的数据源;其次,利用遗传算法对小波神经网络进行最佳适应优化赋值,从而避免神经网络陷入局部最小值的问题;最后,对模型进行仿真,并与传统的小波神经网络、BP神经网络的预测结果进行对比。结果表明,优化后的小波神经网络,具有适应度好、预测精度高、鲁棒性能强的能力,且能满足西北能源局实行的新细则要求。

A short-term photovoltaic power prediction model based on wavelet neural network optimized by genetic algorithm was proposed to improve the accuracy of photovoltaic power prediction in northwest China. Firstly, the original data is cleaned, and the ideal data source is obtained.Secondly, the wavelet neural network is optimized by genetic algorithm to avoid the problem of local minimum.Finally, the model of wavelet neural network optimized by genetic algorithm is simulated and compared with the prediction results of traditional wavelet neural network and BP neural network.The results show that the optimized wavelet neural network has good adaptability, high prediction accuracy and strong robustness, and can meet the requirements of the new rules implemented by the Northwest Energy Bureau.