摘要
微细铣削过程中的颤振是一种加工不稳定现象,会导致加工表面恶化、刀具快速磨损甚至刀具破损。本文提出了一种基于梅尔倒谱系数—隐马尔可夫模型改进的声音信号机器学习模型,更适用于加工过程的状态识别。开展Ti-6Al-4V钛合金微细铣削在不同加工状态下的声音信号采集试验,用于训练机器学习模型并获得模式库。通过对铣削过程中的不同声音信号与模式库进行比较,验证了所提出的机器学习模型的准确性。研究表明,基于合理的特征选取和模型参数优化,所提出的机器学习模型对加工状态的识别准确率达到82%。本研究可为改进微细铣削过程中的在线监测技术提供指导。
Chatter in micro-milling process is an unstable vibration that could cause deteriorated surface,rapid tool wear and even tool breakage.In this paper,an improved machine learning model for sound signals based on the Mel frequency cepstrum coefficient(MFCC)-Hidden Markov model is proposed,which is more suitable for process recognition.Experiments are carried out to capture the sound signal during different processing states,which are used to train the machine learning model and obtain a pattern library.The accuracy of the proposed machine learning model is verified by comparing the pattern library with different sound signals in micro milling of Ti-6Al-4V titanium alloy.The results show that based on reasonable feature selection and model parameter optimization,the proposed machine learning model achieves an accuracy of 82%for the recognition of processing states.This work will provide guidance for improving the online monitoring technology during micro milling.
作者
宋吉超
赵国龙
李亮
年智文
何宁
Song Jichao;Zhao Guolong;Li Liang;Nian Zhiwen;He Ning(不详;College of Mechanical and Electrical Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China)
出处
《工具技术》
北大核心
2023年第12期135-139,共5页
Tool Engineering
基金
国家自然科学基金(52075255)
中央高校基本科研业务费专项资金(NT2021020)。
关键词
颤振预测
微细铣削
状态识别
梅尔倒谱系数
隐马尔科夫链
chatter prediction
micro milling
state identification
mel frequency cepstrum coefficient(MFCC)
hidden markov model(HMM)