毕业设计论文风电场功率短期预测方法研究.doc

PAGE 风电场功率短期预测方法研究 摘要 风力发电作为一种清洁的可再生的发电形式，近几年来发展迅猛。但是由于风力发电的输出功率的不稳定性，使得其大量并网的难度增大，安全性和可靠性得不到保障，从而制约了风电的大面积推广应用。因此对风电的输出功率的预测显得尤为重要。研究出一种准确、有效的风电功率预测系统可以将风力发电的应用提升到一个新的高度，具有重要的意义和深远的影响。在国内，风力发电处于一个飞速发展的阶段，相对应的预测方法研究却尚处于起步阶段，因此风电功率的预测研究在国内需求更加迫切。 因此，本次的本科毕业设计课题我选择了研究对于风电功率进行短期预测的一些方法，在了解了各种方法的优劣之后，我选择采用BP神经网络的方法对短期的风电功率进行预测，同时分析误差，提出可行性建议。主要的工作包括了：课题研究背景的了解，国内外的研究现状了解，预测方法的了解及筛选，程序的编写，预测结果的分析及可行性建议的提出。 本次论文采用BP人工神经网络的方法，根据从马科基谷电力合作社获得的某个月的相关风速、风向、温度数据作为输入变量，建立一个输入层有四个节点数，隐含层八个节点数，输出层一个节点数的BP人工神经网络结构模型，对四小时内的风电功率进行短期预测，得出结果并分析误差。通过分析得知，BP人工神经网络模型预测短时间内的风电功率时具有一定的准确性和可行性，得出的预测结果可作为并网及调度的参考数据。 关键词：风电；功率；短期预测；BP神经网络 The Research Methods of Wind Power Short-term Prediction Abstract Wind power, as a kind of clean and renewable power, has developed rapidly in recent years. But the instable property of output power has made the popularization of wind power more difficult and safety and reliability can not be guaranteed to restrict the wide application of wind power. So the prediction of output power of wind power is particularly important. A precise, efficient wind power prediction system can lift the wind power to a new height, which has a great importance and far-reaching influence. In our country, wind power is in a rapid development stage, the prediction method of corresponding studies just starts. Therefore the demand of wind power prediction research in domestic demand is much urgent. Therefore, about the undergraduate course graduation design paper, I chose to study some short—term forecasting methods of wind power .I will choose the method by using BP neural network for short-term prediction of wind power, at the same time analyzing the error and giving some proposes suggestions. The main work includes: knowing the background of the research, finding out the research situation at home and abroad, understanding and screening of prediction methods, writing program, analyzing the results and proposing suggestions. According to the input data variables of wind direct