Preventive Maintenance Decision Based On Dynamic Hazardous Prediction

With the improvement of production rate and the diversification of customers' demand,the traditional mass production mode under single-batch in large-scale is becoming more and more uneconomic for enterprises.And today's demand-oriented production mode under multi-variety in small batch is gradually becoming mainstream.This production mode put forward higher requirements for more flexible production planning,lower production,operation and maintenance costs,equipment operate more reliably and more safely.To this end,the study of how to achieve the highest level of equipment reliability,how to make scientific preventive maintenance cycle decision to achieve the most economical equilibrium strategy has become an important research field of reliability.Firstly,the dynamic risk prediction model of repairable system is studied in this paper.Based on the equipment's fault tree structure,the equipment topology-Bayesian network is established.When the different parts are maintained at different levels,the life and reliability of the equipment are updated dynamically.Thus,the real-time overall risk of the system could be assessed with the components' life and the failure rate distribution parameters.And Numerical illustration is done by some examples to further verify and analyze the theory.Secondly,this paper studies the PM(preventive maintenance)decisionoptimization problem of complex equipment under imperfect repair context aiming at the deficiencies of residual life prediction and cost estimation.Based on the virtual age theory,we built the relationship between maintenance efficiency and equipment age.Furthermore,three different methods are put forward to calculate the relation between checking intervals and maintenance period.The quality fluctuation and variable maintenance cost caused by fault are considered to build a more concise overall cost estimation model.A multi-objective decision-making model for preventive maintenance is established to ensure high availability of equipment and low total cost.Numerical verification is performed using the production and maintenance history data of a machining center.The verification intends to compare the different overall cost rate and equipment availability under static,dynamic and failure limits calculation methods and various maintenance levels.In the same way,the applicability of the three methods is discussed.The influence of the quality cost on the optimal strategy is analyzed to prove the advanced nature of the method.