以金刚石砂轮为研究对象,提出利用隐马尔可夫模型进行砂轮磨损状态识别,根据磨削过程声发射信号来识别隐藏的砂轮磨损状态,实现砂轮磨削性能退化评估.基于上述理论,搭建实验平台,采集砂轮全寿命周期磨削过程声发射信号,利用线性判别分析降维算法对声发射频域信号进行特征降维,将降维后的特征作为模型观测序列,分别建立砂轮磨损状态的连续隐马尔可夫模型.实验数据表明,连续隐马尔可夫模型可以清晰地反映砂轮从稳定磨损到急剧磨损的状态转移,砂轮磨损阶段状态间相似度能够达到95%以上.
In this study,we take diamond grinding wheels as the research object,and propose to use the hidden Markov model to identify the grinding wheel wear status.According to the acoustic emission signal of the grinding process,the hidden grinding wheel wear status is identified to realize the grinding wheel performance degradation assessment.Based on this theory,we build an experimental platform to collect the acoustic emission signals of the grinding wheel during the full life cycle grinding,and use linear discriminant analysis(LDA)to reduce the dimensionality of the acoustic emission frequency domain signal.In addition,we have used the post-dimensionality reduction characteristics as model observation sequences to establish continuous hidden Markov model of grinding wheel wear state.Experimental data show that the continuaus hidden Markov model can clearly reflect the state transition of the grinding wheel from stable wear to sharp wear,and the similarity between the grinding wheel wear stages can reach more highly than 95%.Experimental results verify the effectiveness of the evaluation of grinding wheel performance degradation based on continuous hidden Markov model.
