基于小波神经网络的钢丝绳断丝定性和定量检测  

Qualitative Classification and Quantitative Inspection for Broken Wires in Wire Ropes Based on Wavelet Neural Network

作　　者：张东来[1] 徐殿国[1] 

机构地区：[1]哈尔滨工业大学,哈尔滨150001

出　　处：《仪器仪表学报》2002年第z2期486-488,491,共4页Chinese Journal of Scientific Instrument

基　　金：哈尔滨工业大学跨学科交叉性研究基金资助(No HIT.MD2001.21)项目。

摘　　要：提出了钢丝绳断丝定性和定量分级检测的方案,并根据二者的特点给出了两种小波神经网络模型和权值学习算法。对定性检测,输入层和隐含层之间用小波函数作为权系数,两层之间无非线性;对定量检测,应用小波非线性,神经网的输入是特征向量和小波的内积。前者适于定性分类,后者适于特征与断丝程度之间定量关系的逼近。实验结果表明:两种小波神经网络较一般的BP网络收敛速度快,外推能力强,识别精度好,这种方法成功地区分了内、外部断丝,极大地提高了断丝定量检测的准确度。This paper proposes two wavelet neural network models and weights study algorithms for qualitative classification and quantitative inspection of broken wires in steel wire ropes respectively. Wavelet functions act as weights but not using wavelet nonlinearity between input and hidden layer in qualitative classification, while the input layer of neural network is inner product of feature vector and wavelet using wavelet nonlinearity in quantitative inspection. The former is suitable for signal classification, the latter suitable for description of quantitative relationship between features and damage ratio. The experiment results show the two models have faster convergence speed for network training, more generalization capacity and accurate inspection than the general BP-Net-work. The method can differentiate internal and outer broken wires, especially improve accuracy of quantitative inspection greatly.

关 键 词：小波神经网络 钢丝绳 断 丝 无损检测 

分 类 号：TP2[自动化与计算机技术—检测技术与自动化装置]