摘要
对电子式三相电网电能表计量误差的预测,能够及时的发现电能表计量设备的潜在隐患。对电能表计量误差的预测,需要拟合出电子式电能表计量非线性系统,对适应度函数进行优化,完成电能表计量误差的预测。传统方法通过灰色关联分析理论,选取电能表计量误差训练样本,但忽略了对适应度函数的优化,导致预测结果偏低。提出基于神经网络的电子式三相电网电能表计量误差预测方法。研究电子式三相电网电能表计量设备综合误差的构成,建立用于描述非线性分布的组合半梯度云模型,用于预测电网传输电压和电流互感器的运行误差,利用以获取的电能表计量误差时间序列经过神经网络进行在线学习,拟合出电子式电能表计量非线性系统来预测三相电能表计量值,对适应度函数进行优化,达到电能表计量误差动态预测的目的。实验结果表明,所提方法预测误差较小,具有较强的泛化能力。
A prediction method for metering error of three - phase kilowatt - hour meters with electronic type is proposed based on neural network. Firstly, constitution of composition error of metering equipment of the meters is studied and combination semi - gradient cloud model describing nonlinear distribution is built to predict operation er- ror of transmission voltage of power grid and current transformer. Then, using the obtained time series of metering er- ror of the meter, online learning is carried out via neural network and nonlinear system of metering of the kilowatt - hour meter is fitted out to predict measuring value. Fitness function is optimized. Finally, the dynamic prediction of the metering error is achieved. Simulation results show that the proposed method has small prediction error. It has strong generalization ability.
作者
沈华
徐莹
SHEN Hua XU Ying(State Grid Shanghai Electric Power Research Institute, Shanghai 200051, China South Power Supply CompanyEPC,Shnghai 200233, China)
出处
《计算机仿真》
北大核心
2017年第10期101-104,共4页
Computer Simulation
关键词
三相电能表
计量
误差预测
Three - phase kilowatt - hour meter
Measurement
Error prediction
