神经元网络钢丝绳断丝信号定量识别技术  

Application of Improved B-P Algorithm of Neural Network to Wire Rope Quantitative Detecting Technology

作　　者：黄锐[1] 卢文祥[1] 杨叔子[1] 

机构地区：[1]华中理工大学

出　　处：《中国机械工程》1994年第1期1-3,共3页China Mechanical Engineering

基　　金：国家自然科学基金

摘　　要：在分析了钢丝绳电磁无损探伤技术的发展对检测系统提出的新问题之后，尝试了采用多层神经元网络改进Ｂ－Ｐ算法设计钢丝绳检测网络。讨论了加快网络收敛的主要措施。实验结果表明，采用神经元网络可以较好地解决钢丝绳断丝信号的定性、定量分析。In this paper, two problems involved in wire rope electromagnetic non-destructive detecting systern are first analysed. Then, improved B-P algorithm which is based on multilayer neural network is designed for broken wire recognition,main measures for accelerating the B-P network convergence are discussed. Last,experimental results show that qualitative and quantitative analysis of broken wire in wire rope can be well done by the neural network.

关 键 词：钢丝绳 断丝 模式识别 神经网络 

分 类 号：TD532[矿业工程—矿山机电]