【摘要】 介绍了基于Levenberg-Marguardt(LM)算法的改进BP神经网络及其在热电偶特性建模中应用,针对其存在的问题,提出了一种基于径向基(RBF)神经网络建立热电偶特性零误差模型的方法,详细介绍了镍铬-镍硅热电偶特性0~1 370℃无差优化模型的建立过程并与BP神经网络热电偶模型进行了比较,结果表明提出的模型算法简单、精度高、稳定性好,易实现,优于BP网络,对其他传感器应用领域也有借鉴作用.更多还原

【Abstract】 BP neural network based on a Levenberg-Marguardt(LM) algorithm and its application to building model of thermocouple characteristic are introduced.In view of the existing defect,the method of building an error-free model for thermocouple characteristic using radial basis function(RBF) neural network is presented.The detailed process of building 0～1 370℃error-free model of Ni-Cr Ni-Si thermocouple characteristic and comparison with BP method are given.The results indicate that the method presented is simple,accurate steady and superior to that of BP network.更多还原