Research On The Identification Of Cutting Response Modal Parameters And The Optimization Of Main Vibration Characteristics Based On Machine Learning

The dynamic characteristics of CNC machine tools are directly related to their processing performance.The primary method to improve the machining performance of the machine tool is to improve its dynamic characteristics related to its own structure.For different processing objects and requirements,the dynamic characteristics of the same machine feedback are often different.Therefore,for the machine tool under specific processing conditions,structural optimization based on its processing response and dynamic characteristics is the most effective method to improve the processing performance.Existing modal analysis methods have large errors in the dynamic analysis of machine tools during processing,and it is impossible to accurately identify the modal parameters that need attention;At the same time,under the background of industrial big data,the existing research has a low utilization rate of data volume,and it is impossible to identify parameters efficiently and accurately;as a modal parameter of an intermediate quantity,how to use it to improve processing quality is also the key point of dynamics research.This paper combines machine learning methods and dynamics characteristics research to perform feature construction,feature extraction and parameter identification on response signals of the same working condition and different batches,providing a feasible method for real-time monitoring of production processing and subsequent processing optimization.Based on the identified parameters,the spindle-tool model is designed in the coupled state,and the vibration loss function with precise leading theory is proposed as the objective function of the model.First,this paper uses singular value decomposition to perform component decomposition and feature extraction on the processed signal data set;The feature matrix is subjected to variance filtering to extract the weighted vector;the sample component set is clustered under the projection of the weighted vector,and the clustering center is used as the feature of the reconstruction correction curve.This method only obtains dynamic parameters under processing conditions by responding to signals,while improving data utilization and considering factors such as machine tool time-varying.Secondly,the modal parameters in the machining process of the machine tool are used to model the entire machining system;Based on the machining system model,an accurate dominant mode identification method is proposed to identify the dominant mode of vibration during machining.The machine tool spindle-tool dynamic coupling system is the object of dynamic characteristics research,and the system objective function with vibration suppression as the target is designed.Finally,the genetic algorithm is used to optimize the tool parameters of the model.According to the cutting response signals of different tools,the method of this paper is used to identify the corresponding modal parameters and compare with the dominant modal parameters of the original machine tool.Comparative experiments show that this method effectively eliminates the dominant mode.In summary,this paper presents a new method to identify the dynamic parameters of the machine tool under the background of big data,taking into account factors such as machine tool time variability.This paper effectively suppresses the dominant mode of its machining system from the perspective of machine tool structure optimization and improves the machining performance of the machine tool.At the same time,the analysis method based on intelligent algorithm provides a new idea for real-time monitoring of processing status and improvement of processing quality.