Study Of Manufacturing Execution System Oriented To Automobile Dies Enterprise

Automobile die making is a key sector of automobile industry and the die making capacity thereof may reflect to some extent the automobile manufacturing strength of a country. In recent decade, China witnessed a rapid growth of automobile die manufacturers prompted by industrial informatization. However, the sector in China lags far behind that in developed countries in that the production control at workshop level needs further improvement. The advent of manufacturing execution system (MES) provides the automobile die manufactures in China with great opportunities for improving the workshop management. Due to the dynamic variation of workshop conditions in this sector, the implementation of MES becomes rather difficult. Thus, the objective of the study is confined to those key issues for building a MES to be used by potential automobile die manufacturers.Taking the specific features of automobile die manufacturing into consideration, it is significant to propose a rational control structure when building a MES requiring adaptivity and reconfigurability. The advantages and disadvantages of certain control structures that are currently used, and the fractal feature of agile manufacturing system are analyzed. Based on the analysis, a fractal control structure using for building a MES is proposed. The control structure is a reasonable compromise between hierarchical control and distributed control. Accordingly, the proposed structure may provide the independence and reconfigurity that are typically owned by distributed control structure, and the coordinated-control for the behaviors and objectives within a specific system as well. To obtain the fractal control structure, fractal control unit is proposed. In particular, multi-agent technology is used to describe the internal structure of the fractal control unit. The functionality of fractal control unit is investigated, and the implementation method of system reconfigurity is exemplified. Dynamic scheduling is the core function to be provided by MES of automobile die manufacturers. To cope with the variation of shop conditions of the manufactures, the MES shall respond any real-time events possibly occurred at workshops. In addition, the production scheduling of automobile dies is typically a single-piece workshop scheduling problem where the priority of dies plays an important role in factors affecting the scheduling process. Therefore, the priority constraint of dies shall be considered during the dynamic scheduling. In this study, the JSS problem using rolling scheduling strategy is formulated by mathematical programming method. The determination of sliding windows, a major procedure in rolling scheduling, is investigated. More specifically, the amount of work-pieces to enter into sliding windows will be obtained on the resource constraint basis. Consequently, two constraints including the priority and the due time of work-pieces are handled by neural network. The neural network based algorithm for selecting work-pieces to be included in sliding windows is devised. Based on the variation of shop conditions in automobile die manufacturing enterprises, a mixed rolling scheduling strategy, the preconditions for initiating rolling scheduling, and the algorithm framework thereof are presented accordingly.To deal with the weakness of neural network (dependency on sample set) used for devising work-piece selection algorithm, the sample set is optimized with rough set theory. This is necessary because the neural network may slow down or even stagnate for unsatisfactory sample set. Based on the optimized decision table, decision rules are proposed. A 5-layer rough neural network is built and a rough-set-based algorithm for selecting work-pieces into sliding windows is devised. With the rule conflict extracted from the optimized decision table, a rough-set-based solution is presented. The due time, priority of work-pieces, and the weight for judging whether a specific work-piece will be selected into sliding windows are determined. By using the weight as the probability of rule selection, the rule conflicts are solved and the efficiency of the proposed algorithm is improved accordingly.Finally, a manufacturing execution system (MES) in the service of automobile die making enterprises is developed. Major core functional modules of the MES are implemented. The MES has been applied to Weifang Die Plant and is playing important role in solving the workshop control problems.