Research On Intelligent Fault Diagnosis Based On Performance Detection Platform Of Ball Screw Pair

CNC is the mother of industrial manufacturing,especially the high-speed,high-precision advanced CNC is the basis of supporting a national manufacturing industry.As an important key functional component of various CNC and automation equipment,ball screw pair directly determines the machining accuracy and performance of CNC machine tools and other equipment.In view of the instability of the overall quality of the domestic ball screw,it is of great significance to diagnose and study the fault mode of the ball screw pair to improve the reliability and intelligent manufacturing level of the ball screw pair.In this paper,the intelligent fault diagnosis method based on the performance testing platform of ball screw pair is proposed.Firstly,a large number of tests are carried out on the ball screw pair of the designated model.Through the analysis of the vibration signal of the fault state of the ball screw pair,the relationship between the different fault modes of the ball screw pair and the vibration signal is clarified,which is verified by the finite element simulation,and then the empirical mode decomposition method is used to analyze the fault state of the ball screw pair The vibration signal of the obstacle state is decomposed,and the fault eigenvector is extracted.Finally,the intelligent diagnosis model based on ELM machine learning algorithm is used to diagnose the fault state of ball screw,which provides the operator with the health state information of the ball screw pair quickly.The main research contents are as follows:(1)This paper summarizes the research and development status of ball screw performance testing at home and abroad,and summarizes the development status of intelligent detection methods for mechanical equipment faults at home and abroad.This paper summarizes and analyzes the structural form and typical failure form of the ball screw pair,and determines that the object is the wear failure of the ball screw pair.(2)The performance testing platform of ball screw is designed and built.According to the main design requirements of the ball screw performance testing platform,the structure design,control system design and testing system design are proposed,and the system hardware is selected according to the test requirements.The modal analysis and harmonic response analysis of the model are carried out by ANSYS Workbench,which are mutually verified with the actual test results,and provide an effective reference for the selection of the vibration sensor and the selection of the installation position of the detection platform.The wear failure characteristics of ball screw pair and the preparation method of simulation characteristics are described.(3)Experimental study on the relationship between wear failure and vibration signal based on EMD.The basic theory of empirical mode decomposition(EMD)is introduced,and the decomposition steps of EMD are described in detail.Then,the method is applied to decompose the vibration signal of ball screw pair in fault state and extract vectors,which verifies the effectiveness and superiority of the method.(4)Experimental research on intelligent diagnosis of wear failure based on detection platform.Firstly,the positioning accuracy of ball screw pair is measured,which provides the accuracy guarantee for the normal test.Secondly,the fault intelligent diagnosis model based on ELM machine learning algorithm is established,and its training and testing are carried out,which proves that the intelligent algorithm proposed in this paper can correctly diagnose the performance of ball screw pair.Finally,the performance comparison experiment is carried out with BP algorithm The superiority of this method is further verified.