利用BP神经网络实现开关柜温度预测

摘要

电力设备在线监测是智能电网的一个重要应用领域，中压配电开关柜的运行温度监测就属于这一范畴。本文在分析了开关柜温度场的特性之后，提出了利用BP神经网络进行温度预测的解决方案。文中采用图表结合的方式，阐述了神经网络的结构，对于关键性函数，给出了具体的公式。最终，本文详细阐述了利用BP神经网络实现开关柜温度预测的实施方案，并利用一组实验数据对方案进行了验证。%Electrical equipment status on-line monitoring is an essential domain of smart power grid, including monitoring operation temperature of medium voltage distribution cabinet. Firstly, the article analyzes the character of temperature field in a cabinet, and then it points out that using nerve net to forecast temperature of a cabinet is an effective way. The article shows the common structure of a nerve net, and then it demonstrates the scheme using BP nerve net to forecast temperature of a cabinet. In order to give clear principle demonstration, necessary figures and tables are given in the paper, so are some formulations. In the end, the paper uses a series of testing data to validate the scheme.