Human Reliability Analysis Based On Multi-state Theory In Mixed Model Assembly Systems

With a wide range of products varieties, assembly systems moved from a massproduction scheme to a mass customization scheme, where the mixed model assemblysystems become necessity. The product quality can’t be assurance with more humanfactors errors increased for the change of product work task. So studies the humanreliability of in mixed model assembly systems is very important. However, therelationship between multi-human reliability and single person reliability is not linear.Worker has many operation states, which increase the difficulty of the humanreliability analysis in mixed model assembly systems.The thesis researches the human’s multi-state in mixed model assembly systemsbased on the multi-state theory. A latent human error exists in the mixed modelassembly systems. And it can be propagated in the stages, finally causing thenonconforming product. To explore the latent human error propagation in assembly, areliability model of multi-state systems with propagated human error is put forward,while the assembly time is used as the metrics. The model is based on the universalgenerating function, taking the recovery factor into considered. Then an algorithm ofgeneralized reliability block diagram is used to resolve the proposed model. Finally,an illustrative example is provided. The case study shows that considering the humanerror propagation contributes to estimate the reliability of multi-stage assembly moreaccuracy and taking recovery factor into account makes the model meet the actualassembly situation.Besides, on one hand, the study has a vital theoretical significance on analyzinghuman error and reliability of mixed model assembly systems. On the other hand, theresearch has certain practical significance on assessing human reliability in mixedmodel assembly systems, improving the assembly workers’ efficiency and quality.