Research On Service Condition Assessment Of Electric Spindles Based On Multi-source Information Fusion

As the central component of high-grade CNC machine tools,the spindle’s service performance directly affects the machine tool processing efficiency and workpiece processing accuracy.To evaluate the service status of electric spindles and study the trend of their performance changes,it is possible to identify potential threats of failure of electric spindles in time,so that maintenance decisions can be made to achieve pre-failure maintenance and extend the service life of machine tools.In addition,with the increasing demand for intelligentmachinery,the development of intelligent spindle unit has become the technical development direction of the next generation spindle,and the electric spindle condition evaluation technology is of great significance in guaranteeing the high-speed operation of intelligent spindle.To this end,this paper investigates the state change process of electric spindles in normal service based on the state parameters of electric spindles in service,coupled with single-source processing algorithms and multi-source information fusion algorithms,through the analysis of the location of common failures of electric spindles and the generation mechanism,as follows:1.In this paper,based on the understanding of the internal structure and working principle of the electric spindle,the location of common faults and the mechanism of their generation are analyzed,and the vibration signal and motor current signal during the service of the electric spindle are determined as the data base for condition evaluation.Meanwhile,the multi-domain feature extraction method is used to extract features from a single information source,and a state feature processing method based on PCA and improved DBSCAN is constructed for calculating the feature indicators of electric spindle operation state.Finally,the method was validated using the Case Western Reserve University(CWRU)bearing experimental dataset.2.In order to further analyze the trend of electric spindle state change,this paper analyzes the fusion level and fusion algorithm on the basis of multi-source information fusion technology;By comparing the advantages and disadvantages between different fusion levels,the fusion method with decision level fusion as the direction is determined.Also,the information fusion algorithm based on the classical D-S evidence theory is determined by combining the practical application background and the characteristics of the information fusion algorithm,and the principle and defects of the algorithm are analyzed and discussed.The algorithm is validated using the experimental dataset of Paderborn University(PU)bearings.3.To address the shortcomings of the classical D-S evidence theory in practical applications,this paper proposes a fusion algorithm based on the entropy weight method to improve the D-S evidence theory.The weights of multiple evidence sources are determined by calculating the entropy values among them,and the initial evidence sources are corrected so that the evidence sources that are more sensitive to changes in the service state of the spindle are enhanced.To compare the fusion results of classical D-S evidence theory as a way to verify the superiority of the improved algorithm.To establish the framework of electric spindle service condition assessment and further validate the above method in conjunction with the electric spindle experimental bench.4.Integrate the electric spindle condition assessment methods studied in this paper,and design and implement the development of the electric spindle service condition assessment system.Relying on Qt,C++,MySql database and other tools,it completes the integrated development of interface login,data collection and condition monitoring,condition assessment and database management to realize the practical application of electric spindle service condition assessment.