Preventive Maintenance Method For Key Equipment Of Production Line Under Imperfect Maintenance

As an advanced maintenance method to ensure the smooth,efficient and reliable operation of equipment in the production process,preventive maintenance is mainly used to develop different maintenance strategies by monitoring the status information of production equipment.Due to the high complexity and functional integration of production equipment,the failure of equipment during operation is highly random,resulting in a lot of wasted time and costs,and reducing production efficiency.Therefore,assessing the condition of equipment according to the theory of preventive maintenance and formulating corresponding maintenance strategies is a method and means to ensure the efficient and reliable operation of equipment,which is of great practical significance for improving production efficiency and reducing production costs.Based on this,this project takes the injection molding equipment of the production line as the research object,applies deep learning methods and stochastic process theory,and conducts in-depth research on state-based maintenance decision-making methods under imperfect maintenance,and the specific research contents are as follows:(1)Aiming at the problem of low health status recognition accuracy caused by redundant multi-source condition monitoring data and complex data types of key equipment in the production line,a device health status assessment method based on 1DCNN-LSTM is proposed.On the basis of applying empirical mode decomposition and correlation analysis and other preprocessing methods,the multi-source monitoring information is reconstructed and dimensionalized,and the health status evaluation of equipment is realized by the deep learning model combining 1DCNN and LSTM,and the model effect is evaluated by comprehensive evaluation indicators including monotonicity,correlation and robustness.(2)Aiming at the high complexity of mechanical equipment and strong degradation randomness,resulting in the reliability of equipment operation process is difficult to accurately describe the problem,establish a equipment degradation model based on Gamma stochastic process,modify the results of health state assessment,make the degradation amount meet the increasing characteristics,remove the random fluctuations in equipment operation,and realize the parameter identification of the model through hypothesis testing and maximum likelihood estimation,on this basis,derive the equipment failure probability density function and reliability function,and realize the reliability evaluation of equipment.(3)Aiming at the problem that it is difficult to determine the decision-making variables in the maintenance process due to imperfect repair and random impact failure in the process of preventive maintenance,combined with the non-new theory of Poisson impact and repair,a preventive maintenance decision-making model under imperfect maintenance of "fixed cycle-threshold" is constructed,considering the limitation of imperfect maintenance times,Monte Carlo simulation and grid search algorithm are applied to obtain the optimal maintenance decision with the goal of minimizing the long-term expense rate,and the cost parameters of the model,The sensitivity analysis of Poisson impact parameters and maintenance number parameters was carried out to verify the accuracy of the model.