Life Prediction And Maintenance Optimization Of Welding Automatic Production Line Based On Condition Monitoring

As the devices and components of complex large-scale system are very complex, which leads to more fault characteristics and larger repair tasks. A reasonable maintenance schedule can prolong equipment’s life, improve productivity and reduce production costs, so the research on maintenance is particularly important. It is worth noting that the maintenance methods experience three stages: passive maintenance, preventive maintenance and predictive maintenance. At present, the method of predictive maintenance schedule for complex large-scale system, which combined with the life prediction and maintenance schedule, needs to be further optimized and discussed.In order to solve these problems, this thesis researched on the equipment’s maintenance management with an objective to minimize the cost, and finally constructed the kind of complex large-scale system’s predictive maintenance optimization model. Firstly, according to the reverse comprehensive analysis method, the fault tree was established, the key events were listed through the fault tree analysis method, and the fault feature parameters which need to be monitored and the corresponding failure threshold were identified. Secondly, based on the Weibull Proportional Hazards Model and considering the recession characteristics of the monitoring signal, the change point was given to improve the traditional life prediction model. Thirdly, the dynamically updated maintenance sequential planning of multi-components system was summarized so as to optimize the maintenance sequential planning model with an objective to minimize the cost. Based on the above studies, the predictive maintenance optimization method and strategy of multi-devices hybrid system were proposed under the relationship of physical structure, and then the predictive maintenance optimization model was established, too. Finally, the splice assembly production line of automatic welding production line was regarded as an example to verify the effectiveness of both life prediction optimization and maintenance sequential planning optimization model. Meanwhile, it also demonstrated that the multi-device hybrid system’s predictive maintenance optimization model was reliable and economical.This thesis makes full use of the condition monitoring information, and provides a new way for the equipment maintenance management. At the same time, the predictive maintenance optimization method also makes up the shortcomings of the traditional one, which makes it more satisfied with the requirement of practical engineering and achieves a higher economic significance. In addition, the research on improving the predictive maintenance sequential planning can guide the equipment management and design works, which will contribute to the positive social value.