基于径向基函数神经网络拟合水泵特性曲线  

Fitting the characteristic curve of water pump based on radial base function 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,江西南昌330027 [2]江西农业大学理学院,江西南昌330045

出　　处：《机械设计》2008年第6期14-16,共3页Journal of Machine Design

基　　金：江西省重点新产品研究资助项目(200310707300)

摘　　要：通过试验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。在分析水泵性能试验及其特性曲线和阐明径向基函数神经网络的方法理论及其特性的基础上,建立了径向基函数神经网络拟合水泵特性曲线的模型并加以应用。结果表明:效果显著,具有工程应用价值。The characteristic curve of water pump protracted by the experimental method is the base for judging the water pump whether or not satisfying the designing requirements. The fitting precision of characteristic curve decides the accuracy of judgement. The artificial neural network possesses rather strong non-linear processing ability and approaching capability. On the basis of analyzing the performance test of water pump and its characteristic curve and of expounding the methodology theory of radial base functioned neural network and its characteristics, the model of water pump characteristic curve fitted by the radial base functional neural network was established and being adopted. The result showed a remarkable effect and the model possesses values of engineering application.

关 键 词：径向基函数 神经网络 水泵 特性曲线 拟合 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TH31[自动化与计算机技术—控制科学与工程]