Modeling And Analysis Of Multiproduct-Multi-Stage Manufacturing Systems For Quality Improvement

Market requirements of products are changing and developing constantly.To meet the changing needs,on the one hand,products should be various.On the other hand,they are becoming more and more complicated which need to be manufactured by several related processes.This paper,aiming at such multi-product-multi-stage manufacturing systems,developed a Markov model to evaluate the quality performance of the system.It considered both product variety and quality propagation along the production line and did some analysis about quality properties including monotonicity analysis,sensitivity analysis,product sequence analysis and quality bottleneck identification.The paper consists of the following five parts.(1)A basic Markov model for single-product-two-stage systems has been proposed.In the system,upstream stage and downstream stage are closely related.The quality of the final product is not only related to the state of the current stage,but also has something to do with the quality of coming parts fromupstream stage.In this case,two adjacent stages should be seen as a whole when applying Markov model and the model for two-stage systems is basic.(2)An extended Markov model for single-product-multi-stage systems has been developed.By applying iterative method,the model for two-stage systems could be extended to the model for single-product-multi-stage systems.Numerical experience demonstrates the accuracy of this iterative method.(3)A Markov model for multi-product –two-stage systems has been developed.Producing different products in the system means different internal and external transition probabilities.Based on this fact,we could extend the model for single-product-two-stage systems to Markov model for multi-product-two-stage systems.(4)A general Markov model for multi-product-multi-stage systems has been derived.Combining the single-product-multi-stage model and the multi-product-two-stage model,the general Markov model for multi-product-multi-stage systems could be derived successfully.(5)Analysis of quality properties based on the model has been conducted.The quality analysis consists of production strategies,product sequences as well as internal and external quality bottleneck identification under Bernoulli and Bernoulli-relaxation conditions.Finally the Markov models proposed in this paper are applied in the manufacturing system of astronautical valves and the quality properties including quality changes along the production line,sensitivities of different parameters,best product sequence and quality bottleneck identification are analyzed.These cases show that the models developed in this paper are practical and effective and have high-level accuracy.