摘要
为解决车辆减振器工业生产中产品性能在线检测完全依赖于人工判断等问题,将BP神经网络理论应用到减振器示功图的自动判别中。在Matlab平台上建立了一个用于减振器示功图识别的3层BP神经网络模型,对示功图的复原和压缩曲线分别进行离散处理,构建了复原和压缩中心距两个特征向量,其中包括复原和压缩饱满度两个特征值;通过训练和分析,选取Scaled共轭梯度算法作为BP神经网络的训练算法;进行了示功图自动判别实验,对示功图复原和压缩阻力作出了评判。研究结果表明,经过训练的BP神经网络能够对减振器示功图进行判别,其判别结果与本领域技术人员判断结果基本一致。
In order to solve the problems of the artificial judgment in the on-line detection of product performance in the vehicle shock absorber industry,the theory of BP neural network was applied to identification of the indicator diagrams of shock absorber. A three-layer BP neural network used to identify the indicator diagrams was established by means of Matlab. Based on discretizing the tensile and compression parts in the indicator diagram curve, the two eigenvectors of the central distance of tensile and compression, including the two eigenvalues of the satiation degree of tensile and compression, were given. Through training and analysis, the Scaled algorithm of conjugate grads was chosen as the training one for BP neural network. The experimental results indicate that the recognition effect based on the trained BP neural network generally coincides with the judgment of technicians.
出处
《机电工程》
CAS
2012年第8期929-931,共3页
Journal of Mechanical & Electrical Engineering
关键词
减振器
示功图
BP神经网络
识别方法
shock absorber
indicator diagram
BP neural network
identification method
