Research On Multi-feature Decision-making And Job Optimization Scheduling Method For Complex Equipment Assembly Quality

The upgrading of manufacturing technology for complex equipment is one of the important conditions for the improvement of manufacturing industry on automation and intellectualization.Because of its multiple functions,complex assembly process and strict precision requirements,the quality of assembly process has a significant impact on the use of complex equipment and the consumer satisfaction.The task of complex equipment assembly mainly contains the assembly of components and their docking.The assembly process involves many work and constraints.Some of the work can be done simultaneously,and different work need different personnel to complete.At present,management methods of the assembly process for such complex equipment are based on experience in most of enterprises.It is difficult to predict and control the performance of the assembly products,assembly cost and assembly lead time with different operation methods,working environment,personnel and machine condition.To solve these problems,this paper proposes a method of assembly quality prediction and resource optimization of complex equipment based on digital decision-making technologies,which provides a way to realize quantification,optimization and intellectualization of complex equipment assembly process factor.The quantification for assembly data of complex equipment is the first step in digital decision-making.Based on the analysis of the status of complex equipment assembly and the quality objectives of assembly process,this paper constructs the index system of assembly quality features for marine elevator.Aiming at solving the problems of inefficiency,redundancy and contradiction in assembly quality features from multi-source,a manifold learning method is proposed for data processing.Secondly,the framework of assembly quality prediction for complex equipment based on multi-classifier combination is established.Compared with single classifier,the results through the decision-making fusion by DS evidence theory is more reliable.Finally,an assembly scheduling model is established for three quality objectives of assembly process: performance of the assembly products(precision,function and appearance),assembly cost and assembly lead time.Besides,the scheduling schemes of work and personnel are obtained with improved genetic algorithm.In the process of quantification and decision-making,this paper validates above methods with the marine elevator as the object,and it is hoped that the framework can lay a foundation for the follow-up work.