基于模糊神经网络的WEDM可靠度预计  

Reliability Prediction of WEDM Based on Fuzzy Neural Network

作　　者：张宏斌[1,2] 贾志新[1] 郗安民[1] 

机构地区：[1]北京科技大学机械工程学院,北京100083 [2]陆军航空兵学院机载设备系,北京101114

出　　处：《计算机工程》2010年第8期12-14,共3页Computer Engineering

基　　金：国家自然科学基金资助项目"立式回转电火花线切割加工技术及其机理研究"(50875021)

摘　　要：为了准确预计电火花线切割机床(WEDM)的可靠度,建立基于自适应模糊神经网络的可靠度预计模型。该模型以平均无故障时间为输入,以可靠度为输出,通过神经网络自适应训练获得适合WEDM可靠度预计的平均无故障间隔时间隶属函数。仿真结果表明,该模型的预计精度较高,与应用神经网络的WEDM可靠度预计结果相比,提高了96.4%。In order to predict the reliability of Wire cut Electric Discharge Machining（WEDM） accurately,this paper establishes a reliability prediction model based on self-adaptive fuzzy neural network.This model takes the Mean Time Between Failure（MTBF） as the input and takes the reliability as the output.The membership function of the MTBF is achieved by the neural network self-adaptive training.Simulation results show that this model has high prediction precision which is improved by 96.4% compared with that of WEDM reliability by using neural network.

关 键 词：模糊神经网络 可靠度预计 电火花线切割机床 隶属函数 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]