基于模糊神经网络的光伏发电量短期预测

Photovoltaic power short-termprediction based on fuzzy neural network

受天气状况、辐照度、温度、湿度等气象因素的影响,光伏系统的输出具有很强的非线性和非平稳性的特点,光伏发电量预测精度较低。该文根据光伏系统的历史发电数据和实际气象数据,采用模糊识别与RBF神经网络相结合的方法,实现光伏系统发电量的短期预测。首先对影响预测结果的气象因素进行分析,然后按天气类型进行分类,对不同的天气类型分别建立模型进行训练,最后利用此模型预测未来的光伏系统发电量,并通过实验仿真验证。预测结果表明,该方法不但减少了模型所需样本数量而且提高了预测的精度,具有一定的科研价值。

Affected by weather conditions, irradiance, the influence of the meteorological factors such as temperature, humidity, the output of the photovohaic system has the characteristics of strong nonlinear and non-stationary, photovohaic power generation forecasting precision is low. Based on the history of the photovohaic system to generate electricity data and actual meteorological data. By adopting the combination of fuzzy recognition and RBF neural network method, realized the short-term prediction of a photovoltaic power system.First meteorological factors affecting the prediction results were analyzed, and then according to classify the weather types, on different weather types training model respectively, finally using this model to predict the future of a photovoltaic power system, and through the simulation experiment.Predicted results show that this method not only reduces the required sample size and improve the accuracy of the prediction model, has certain research value.