Research On Product Reuse Oriented Methodology Of Intelligent Maintenance Decision-Making

Nowadays, natural resource is being over exploited due to highly developed manufacturing, which has brought serious adverse effect on human beings'life, it is hence crucial to find approaches to increase resource utilization rate. Reuse of products has received much attention due to the significant profits it may create and the huge amount of raw resources it may save. Intelligent maintenance technique, which provides users with useful information on the degradation and even failure of products so that they can be maintained timely, can support the reuse of products. In order to reduce maintenance cost and guarantee the reuse of facilities, the maintenance decision-making of facilities in complex production systems based on intelligent maintenance technique has great practical value and realistic significance.A mathematical model for maintenance decision-making of series-parallel production system is built in this paper. The performance degradation process is modeled by reliability function based on Weibull distribution. The definitions, triggers and effects of four types of maintenance actions, namely minor maintenance, medium maintenance, overhaul, and replacement, are presented. The maintenance effects of these maintenance actions on the performances of facilities are modeled by improving Malik's Proportional Age Reduction Model. The topology of production system is built by assigning facilities with facility IDs and production line IDs according to the process sequence of the system.A group maintenance policy is proposed. Static groups are formed by referring to the topology of production system and layout of workshop. Dynamic groups are formed referring to the types of facilities. According to the group maintenance policy, facilities are maintained by groups if they are included in either a static group or a dynamic group, and otherwise maintained separately. Maintaining facilities as a group according to their similarity in aspects of process, position, and type can reduce the complexity of maintenance decision-making problems of series-parallel production systems and increase the efficiency and economic savings of maintenance.Based on the model established above, the constraints and objective function of the maintenance decision-making problem are presented. Limited maintenance resources and capacity of work-in-process buffers are set as the constraints; minimum maintenance cost for one maintenance activity is set as the short-term objective; and minimum maintenance cost rate in a certain period is set as the long-term objective. The calculation method of objective function according to maintenance action information and Gantt chart information is presented. Genetic Algorithm combined with Simulated Annealing, denoted as SA-GA, and Tabu Search (TS) are employed to solve the maintenance decision-making problem. The flow charts of the two algorithms are presented. The utilization of SA-GA in solving maintenance decision-making problems is presented in terms of population generation, fitness function, annealing temperature, selection operator, crossover operator, mutation operator, and parameter settings. The utilization of TS in solving maintenance decision-making problems is presented in terms of encoding, neighbourhood generation, tabu list, aspiration level, and parameter settings.A numerical simulation case study on a bearing production system is presented to introduce the application of the proposed methodology. The maintenance decision-making problem of the bearing production system is solved under group maintenance policy and non-group maintenance policy. The optimal solutions are imported into Flexsim software to carry out production simulation. The results of this case study are analyzed and discussed in detail.