基于BP神经网络拟合水泵特性曲线  

Characteristics Curve Fitting Performance Test of Water Pumps Based on Back Propagation Neural Network

作　　者：刘超[1] 白玲[2] 

机构地区：[1]江西师范大学计算机信息工程学院,南昌330022 [2]江西农业大学理学院,南昌330045

出　　处：《农机化研究》2009年第6期202-205,共4页Journal of Agricultural Mechanization Research

基　　金：江西省重点新产品研究项目(200310707300)

摘　　要：通过试验的方法绘制出水泵的特性曲线是判断水泵是否满足设计要求的基础。特性曲线拟合的精度决定判断的准确性。人工神经网络具有较强的非线性处理能力和逼近能力。为此,在分析水泵性能试验及其特性曲线、阐明BP神经网络的方法理论及其特性的基础上,建立了BP神经网络拟合水泵特性曲线的模型,并加以应用。结果表明,其效果显著,具有工程应用价值。Characteristics curve of water pumps drawing by test method is the basis to use for judging water pumps whether meet its design demand. Fit precision of characteristics curve decided judging veracity . artificial neural network （ANN） has strong capability to approach function and treat non - linear problem. On the basis of analyse performance test of water pumps and its characteristics curve and summarizing the theory and methodology of the back propagation （BP） neural network, this paper establishes BP neural network model fitting characteristics curve of water pumps as well as carried out its application, the results demonstrates its satisfactory effect and values of engineering application.

关 键 词：水泵 特性曲线 BP神经网络 拟合 

分 类 号：TH31[机械工程—机械制造及自动化] TP183[自动化与计算机技术—控制理论与控制工程]