基于神经网络联想记忆模型的数控机床故障诊断  

Research on Fault Diagnosis of Numerical Control Machine Tool Based on Artificial Neural Network Associative Memory Model

作　　者：朱晓春[1] 汪木兰[1] 

机构地区：[1]南京工程学院自动化系

出　　处：《中国机械工程》2003年第15期1275-1277,1286,共4页China Mechanical Engineering

基　　金：江苏省自然科学基金资助项目 ( BK2 0 0 2 0 13 )

摘　　要：提出了将一种改进的人工神经网络联想记忆模型应用于计算机数控机床的故障诊断方法。介绍了基于联想记忆模型的诊断算法 ,总结出数控机床的故障模式及故障分析表 ,然后将该样本向量进行 HADAMARD预处理、存储记忆后 ,可据此模型进行样本和非样本的并行联想回忆 ,实现诊断功能。最后进行了数字仿真研究 。A modified associative memory model of Artificial Neural Network (ANN) is introduced into the fault diagnosis of computer numerical control machine tool. Firstly, the diagnosis algorithm is derived and the fault table is concluded. Secondly, the fault sample vectors are preprocessed by HADAMARD transform, and learned by the model. Finally, the samples and non-samples are parallel associatively recalled based on this network. Furthermore, the digital simulation results demonstrate that the novel strategy is reliable. Additionally, a hardware implementation method is presented, which uses large-scaled Field Programmable Gate Array (FPGA) circuits.

关 键 词：人工神经网络 数控机床 故障诊断 现场可编程逻辑门阵列 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置] TG659[自动化与计算机技术—控制科学与工程]