基于BP神经网络的电网短期负荷预测模型研究

摘要

In order to obtain an automatic forecasting model which can reflect the power net short - term load wave,the paper combines BP neural network and Matlab logic toolbox with the weather conditions, the type of power supply dates and so on as the constraint of the BP neural network; joints the holidays, power limits and so on as the external intervention options of the adaptive control system. Through the model to obtain an adaptive learning rate automatically short - term load forecasting model of electric power system. BP neural network model improves the efficiency and accuracy of the short - term load wave forecast, which can get the accuracy data of the 24 hours load wave cure as an important research topic of the power marketing strategy.%为了获得能够反映短期负荷波动的自动预测模型,将BP神经网络运算逻辑与Matlab神经网络工具箱相结合,利用供电系统的气象条件、供电日的日期类型等函数干扰因素作为BP网络神经元的约束"权值",结合自适应控制中的节假日、供电限制等"外部干预项",通过模型对系统历史运行短期负荷数据进行自动"学习",获得一个具有自适应学习速率的电力系统短期负荷预测模型.基于BP神经网络的电力系统短期负荷预测模型能够通过模型自适应控制"学习"训练,获得供电日24小时的负荷波动曲线,通过与实际运行负荷需求数据比较,验证了该模型的正确性,为电能营销策略的制定提供了一个重要的研究课题.