Page 245 - 《应用声学》2025年第2期

第 44 卷 第 2 期 Vol. 44, No. 2
2025 年 3 月 Journal of Applied Acoustics March, 2025

⋄ 研究论文 ⋄

基于声发射和支持向量机的插齿刀磨削

砂轮状态监测 ∗

路晨辉 刘海涛 † 王建华

(西安工业大学机电工程学院 西安 710021)

摘要：磨削加工对现代智能制造业起着至关重要的作用。砂轮的磨损直接影响到被加工工件表面质量，而主
要依靠经验判断可能会导致效率低下和成本昂贵的问题。该文提出一种基于声发射和支持向量机的插刀磨砂
轮钝化状态监测方法，首先分析了不同砂轮磨损状态下的声发射信号，声发射信号时域均方根曲线和砂轮钝
化能量的理论曲线划分砂轮钝化状态节点，对磨削插齿刀过程产生的时变非稳定声发射信号进行滤波去噪，
避免实验条件对声发射信号的影响。利用小波包分解提取声发射信号各频段有效特征，并对有效特征的选择
进行了对比分析，最终选择对多频段小波包能量系数和时域特征进行拼接特征融合，建立在小样本性能较优
的多分类模型支持向量机。最终，砂轮钝化状态识别准确率可达 91%，能够满足实际加工需求。
关键词：声发射；过程监测；支持向量机；砂轮状态监测
中图法分类号: TH165+.3 文献标识码: A 文章编号: 1000-310X(2025)02-0505-08
DOI: 10.11684/j.issn.1000-310X.2025.02.026

Condition monitoring of grinding wheels for gear shaping knives based on

acoustic emission and support vector machines

LU Chenhui, LIU Haitao and WANG Jianhua

(School of Mechatronic Engineering, Xi’an Technological University, Xi’an 710021, China)

Abstract: Grinding process plays a crucial role for modern intelligent manufacturing industry, and the wear
of grinding wheel directly affects the surface quality of the processed workpiece, while the grinding wheel
wear mainly relies on empirical judgment is likely to lead to inefficiency and costly problems. In this paper,
we propose a method for monitoring the passivation state of grinding wheels with inserted cutter based on
acoustic emission (AE) and support vector machine (SVM). Firstly, we analyze the AE signals under different
grinding wheel wear states, and the theoretical curves of the time-domain root mean square (RMS) curves of
the AE signals and the passivation energy of the wheels are divided into nodes of the passivation state of the
wheels, and we perform the filtering and denoising of time-varying and non-stationary AE signals generated by
grinding inserted cutter to avoid the impacts of the experimental conditions on the AE signals. The wavelet
packet decomposition is used to extract the effective features of each frequency band of the AE signal, and the
selection of effective features is compared and analyzed, and the final choice of the multi-band wavelet packet
energy coefficients and time-domain features are spliced feature fusion, and the multi-classification model SVM

2023-10-13 收稿; 2023-12-27 定稿
陕西省智能制造科技重大专项 (2019zdzx01-02-02)
∗
作者简介: 路晨辉 (1998– ), 男, 陕西渭南人, 硕士研究生, 研究方向: 精密测量与控制。
† 通信作者 E-mail: 1045404516@qq.com