主轴系统切削振动的ANFIS预警研究  

The Cutting Vibration Model of the Spindle Using ANFIS

作　　者：李延磊[1] 胡友民[1] 吴波[1] 

机构地区：[1]华中科技大学机械科学与工程学院,湖北武汉430070

出　　处：《机械设计与制造》2015年第9期128-130,134,共4页Machinery Design & Manufacture

基　　金：973项目"难加工航空零件的数字化制造基础研究"(2011CB706803)

摘　　要：对主轴系统切削振动状态的提前识别,可以避免切削剧烈振动的发生,对提高加工质量和切削效率,具有十分重要的理论和实际价值。基于小波包能量信息熵和高阶累计量谱理论提出了一种识别切削振动状态的特征识别矢量,并依据此建立了基于ANFIS(自适应模糊神经推理系统)的切削状态识别模型。利用小波包理论对切削振动信号的能量分布进行了研究,以信息熵的形式量化了能量分布情况。同时,利用高阶累积量谱理论中的双谱分析统计了切削振动信号的极大值分布情况,以概率的离散系数进行了量化分析。使用量化后的两种特征识别量对ANFIS模型进行了训练和验证,实验证明了模型的有效性。It has very important theoretical and practical value for identifying the vibration status of the spindle system, and avoids the occurrence of severe vibration cutting, thus improves the machining quality and cutting efficiency, It proposed a identification model based on ANFIS （adaptive neuro-fuzzy inference system）, and the model used wavelet packet energy entropy and higher-order accumulation spectrum theory. It studied the cutting vibration signal with wavelet packet energy distribution, and quantified the enersy distribution in the form of information entropy. At the same time, it estimated the maximum value distribution of the cutting vibration signals with higher-order cumulated spectrum of bi-spectrum analysis, then performed the quantitative analysis of probability using discrete coefficient. It trained and tested our proposed ANFIS model, and experimental results show the effectiveness of the model.

关 键 词：ANFIS 小波包 高阶累积量谱 切削振动 

分 类 号：TH16[机械工程—机械制造及自动化] TP29[自动化与计算机技术—检测技术与自动化装置]