Research On Condition Monitoring Of Car Engine Cylinder Head Machining Center Feed System Ball Screw Based On Vibration Signal

Machining center is the core equipment of automobile engine cylinder head production,which affects the accuracy retention and machining reliability of engine cylinder head production line.As a key moving parts of machining center feed system,ball screw’s performance will gradually decline with the increase of running time,affecting the machining precision and quality of machining center.Therefore,it is extremely crucial to study the condition monitoring of feed system ball screw.In order to effectively monitor of feed system ball screw,this paper analyzed and processed vibration signals of ball screw to extract characteristic frequency of vibration excitation source and established a nonlinear mapping model from vibration signal to state index,achieving the research on condition monitoring of feed system ball screw based on the vibration signal.The main research contents are as follows:1)Vibration characteristic analysis of ball screw.First of all,the vibration signal fed into the system’s three-way sensor in ball screw was collected,processed by EMD,wavelet packet transformation and envelope spectrum to extract vibration characteristic frequency of ball screw drive source.At the same time,the method was applied to five machining center in the car engine cylinder head production line for experimental verification.The results show it can effectively identify the screw waviness of excitation frequency and the frequency between balls and returner,providing a strong support for the vibration characteristic analysis of ball screw.2)Feed system ball screw condition monitoring based on wavelet packet and AE-MD.Through the state monitoring method based on wavelet packet and AE-MD,the collected vibration signal was processed by data preprocessing,wavelet packet energy feature extraction,autoencoder for feature dimensionality reduction and state evaluation,realizing the nonlinear mapping from the vibration signal to state monitoring.And applying this method to the machining center under no-load state,load state and moving parts wear for analysis and verification.It shows: the condition monitoring method based on wavelet packet and AE-MD not only performs online real-time condition monitoring and wear monitoring of feed system ball screw,but also has good stability and effectiveness.3)Application of multi-data fusion in condition monitoring of feed system ball screw.In order to make full use and integration of vibration signal of ball screw in all directions and reduce the impact of vibration influencing factors on the condition monitoring,the credibility fusion was used in the characteristic layer and decision layer,the separability index was used for evaluation.The fusion result shows: feature layer fusion c an effectively integrate the differences between vibration signal effective features in various directions,and decision layer fusion can balance the fluctuations of state index in each direction,the both have good stability and feasibility.