基于数据挖掘和遗传小波神经网络的光伏电站发电量预测

为了解决光伏发电预测不确定性问题,进一步提高光伏电站发电量的预测精度.提出一种基于数据挖掘和遗传小波神经网络的光伏电站发电混合预测模型,利用K均值聚类算法对历史数据进行分类,并对传统BP神经网络进行改进.以BP神经网络为基础,引入小波分析构建小波神经网络,同时利用遗传算法对网络的初始参数进行全局寻优得到最优参数,利用交...

PHOTOVOLTAIC POWER GENERATION PREDICTION BASED ON DATA MINING AND GENETIC WAVELET NEURAL NETWORK

In order to solve the prediction uncertainty and further improve the prediction accuracy for photovoltaic(PV)power stations.The paper proposes a hybrid prediction model for forecasting the generation of PV power stations based on data mining and genetic wavelet neural network. The model utilizes the K mean clustering algorithm to classify historical data and constructs a genetic wavelet neural network based on BP neural network by using wavelet analysis. Furthermore,the initial parameters of the network can be globally optimized with a genetic algorithm and the learning rules of the network are improved by a cross-entropy function. The proposed network has not only good characteristics with local time and frequency domain owned by wavelet analysis,but also has global search ability,which increases the possibility of jumping out of local optimum and has faster convergence and better stability. The experimental results shown the effectiveness of the proposed method.