光伏发电短期功率预测方法研究

光伏发电短期功率预测方法研究摘要新能源在环境保护、资源丰富等方面有着无可比拟的优越性，而光伏发电是一种技术相对成熟的新能源，如果能够得到有效的开发和应用，不但能够满足人民的日常生活需求，而且还能够为“双碳”目标做出贡献。然而，由于天然光能的不稳定特性，经过机组的变换后，其输出的光能也会呈现出不稳定的特性，其接入电网后，将会给电网带来极大的安全隐患，极大地加大了光能的开发与利用难度，所以，亟需对光伏发电进行有效的预测。本文将BP神经网络应用于典型的光伏发电系统，可以实现对未来某个特定天数的光伏发电功率的预测。在此基础上，分析了神经网络的隐层结构与其预测性能之间的关系，并得出了最优隐层神经元数量为6。在此基础上，通过对不同种类的气象资料，以及它们的组合，以及在不同的时间步长下观测到的光伏发电资料，研究输入资料的预测方法。实验结果表明，基于B神经网络的光伏发电预测模型，能取得较好的预测效果。关键词：光伏发电功率预测：神经网络：新能源AbstractNew energy has incomparable superiority in environmental protection andresource abundance,etc.,and photovoltaic power generation is a new energy withrelatively mature technology,which can not only satisfy people's daily life needs,butalso contribute to the goal of "double carbon"if it can be effectively developed andapplied.However,due to the unstable characteristics of natural light energy,after thetransformation of the unit,the output light energy will also show unstablecharacteristics,and its access to the power grid will bring great safety hazards to thepower grid,which greatly increases the difficulty of the development and utilizationof light energy,so there is an urgent need for effective prediction of photovoltaicpower generation.In this paper,BP neural network is applied to a typical PV power generationsystem,which can realize the prediction of PV power generation power for a specificnumber of days in the future.On this basis,the relationship between the hidden layerstructure of the neural network and its prediction performance is analyzed,and theoptimal number of hidden layers in the range of six is derived.On this basis,theprediction method of input data is investigated by looking at different kinds ofmeteorological data,and their combinations,as well as PV power data observed atdifferent time steps.The experimental results show that the PV power generationprediction model based on BP neural network can achieve better prediction results.Keywords:photovoltaic power generation;neural network;new energy目录摘要第一章绪论11.1研究背景及意义11.2国内外研究现状31.3本文研究内容4第二章光伏发电功率预测概述52.1基本算法概述2.2技术难点…......m..m.......62.3本章小结6第三章光伏发电功率预测方法分析73.1BP神经网络.73.2支持向量机93.3本章小结10第四章基于BP神经网络的光伏发电功率预测114.1光伏发电输出的影响因素分析4.2数据清洗…124.3神经网络建立134.4本章小结…14第五章实验设计和分析155.1神经网络的设计155.2基于BP神经网络的光伏发电功率预测模型165.3预测实例及误差分析165.4本章小结.…17第六章总结和展望1819参考文南歌.。致谢.20