当前位置:主页 > 科技论文 > 自动化论文 >

基于云理论与LS-SVM的刀具磨损识别方法

发布时间:2018-03-03 23:00

  本文选题:刀具磨损 切入点:状态识别 出处:《振动.测试与诊断》2017年05期  论文类型:期刊论文


【摘要】:针对刀具磨损过程中产生声发射信号的不确定性以及神经网络学习算法收敛速度慢、易陷入局部极小值、对特征要求较高等问题,提出了基于云理论和最小二乘支持向量机的刀具磨损状态识别方法。首先,对声发射信号进行小波包分解与重构,滤除干扰频段对求取特征参数的影响;其次,对重构后的信号利用逆向云算法提取云特征参数:期望、熵、超熵,分析刀具磨损声发射信号的云特性及磨损状态与云特征参数之间的关系;最后,将云特征参数组成特征向量送入最小二乘支持向量机进行识别。研究结果表明:所提取的特征可以很好地反映刀具的磨损状态,云-支持向量机方法可以有效地实现刀具磨损状态的识别,与传统神经网络识别方法相比具有更高的识别率,识别率达到96.67%。
[Abstract]:Aiming at the uncertainty of acoustic emission signal produced in tool wear process and the slow convergence speed of neural network learning algorithm, it is easy to fall into local minimum value, and the characteristic requirement is high. A tool wear recognition method based on cloud theory and least squares support vector machine (LS-SVM) is proposed. Firstly, the acoustic emission signal is decomposed and reconstructed by wavelet packet, and the influence of interference band on obtaining characteristic parameters is filtered. Using reverse cloud algorithm to extract cloud characteristic parameters: expectation, entropy, excess entropy, analyze the cloud characteristics of tool wear acoustic emission signal and the relationship between wear state and cloud characteristic parameters. The cloud feature parameters are input into the least squares support vector machine for recognition. The results show that the extracted features can well reflect the tool wear state. The cloud-support vector machine method can effectively realize tool wear recognition. Compared with the traditional neural network recognition method, it has a higher recognition rate and the recognition rate reaches 96.67.
【作者单位】: 东北电力大学机械工程学院;
【基金】:吉林省科技厅科技公关计划资助项目(20140204004SF) 吉林省教育厅“十二五”科学技术研究资助项目(20150249)
【分类号】:TG71;TP18


本文编号:1563045

资料下载
论文发表

本文链接:https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/1563045.html


Copyright(c)文论论文网All Rights Reserved | 网站地图 |

版权申明:资料由用户ecbd5***提供,本站仅收录摘要或目录,作者需要删除请E-mail邮箱bigeng88@qq.com