Research On Prediction Methods Of Grinding Chatter Based On Time Series

With the continuous improvement of the requirements for precision,reliability and longevity of mechanical products in the industry,grinding processing,as the most effective cutting processing method for precision machining in the modern manufacturing field,has an increasingly prominent position and role.In the process of grinding,there are inevitably many problems,in which grinding chatter is one of the main manifestations of the failure of the grinding machine.The grinding chatter not only reduces the quality of the machined surface,accelerates tool wear,generates excessive noise,reduces machining efficiency and accelerates machine tool damage,but also in severe cases can lead to unsustainable machining.Grinding machine chatter signals are usually non-stationary and non-linear,and they are more difficult to predict in strong background signals and noise.Aiming at the hazards caused by grinding chatter,it is of great theoretical and practical significance to seek a method to predict the vibration chatter in a timely and effective manner.The prediction method based on fuzzy time series has been applied in many fields,but in the actual modeling process of fuzzy time series,only the relationship between two adjacent sample data is considered,and the possible existed contact between non-adjacent samples is neglected,which makes the predictions not necessarily ideal.In this paper,the AR(p)type high-order fuzzy time series prediction method is adopted,and the situation of multiple early moments is considered,which improves the correlation between sample data and makes the model have higher prediction performance.The prediction method of AR(p)type high-order fuzzy time series is used for the prediction of grinding chatter,which is the application of this method in a new field,and the method proposed in this paper does not need to consider the non-stationary and non-stationary features of the grinding chatter signal,and has wider applicability.The main research ideas of this paper are as follows:Firstly,based on the in-depth study of the grinding chatter mechanism in the grinding process,the virtual instrument program of LabVIEW software is used to simulate the occurrence of the grinding chatter signal,and the feature quantity of the simulated chatter signal is extracted,The real-time variance feature time series of simulated chatter signals are selected as the predicted sample data,which are applied to two kind of grinding chatter prediction methods which are respectively based on the existing ARIMA model and AR(p)type high-order fuzzy time proposed in this paper.By comparing the prediction results of the two prediction methods,the feasibility of the proposed prediction method is verified.Secondly,the KD4020X16 type of numerical control moving beam gantry guide grindingmachine is taken as the research object,a vibration signal acquisition experimental platform is built.And a detailed experimental scheme is developed to obtain vibration signals of multiple sets of different working conditions,then each group of experimental signals is collated,screened and analyzed,which lays a data foundation for the subsequent processing of experimental signals and verification of prediction methods.Finally,according to the flow of the prediction method,the experimental chatter signal is applied to the proposed method based on AR(p)type high-order fuzzy time series,then the original signal variance time series and predicted value is compared.At the same time,the predictive evaluation index is calculated to verify the predictive effectiveness of the proposed method.In order to further demonstrate the superiority of the proposed method,the same chatter signal is applied to the ARIMA prediction model,and the prediction evaluation index is calculated and compared with the prediction evaluation index value of the previous method,which verifies the accuracy of the proposed method in this paper.