| In recent years, the practice of concurrent engineering requires reliability technology to be updated in the following three respects: (1) Connecting reliability to product failure mechanism; (2) Connecting reliability to product design; (3) Connecting reliability to product quality control. Investigating from the basic points of reliability theory, we found that these three problems are related to the binary state assumption, which serves as one of the basic assumptions in the current reliability technology. To resolve the above three problems, we have to extend the binary state assumption to the general continuous state assumption and redirect reliability technology from failure time analysis to failure process analysis. In this dissertation, we investigate reliability analysis from the point of continuous state analysis. Continuous state reliability analysis is essentially the failure process analysis. Because most of the failure processes can be seen as the degradation processes, we extend degradation process analysis to more general sense to be the basis of continuous state reliability analysis. To connect product reliability to the root cause of failures, we investigate the approaches to establish reliability models based on failure mechanism and product parameters. Dynamic reliability models based on strength-stress interference failure mechanism under the assumptions that both strength and loading force are random processes are established. To connect reliability analysis to quality techniques, we investigate the relationship between product reliability and robust design technique, which is one of the most effective methods in improving product quality. Our basic conclusion is: In order to update the current reliability technology to satisfy the requirement of engineering practice, we have to put reliability technology into a more extensive theoretical framework, use both discrete and continuous state mathematical tools to be the analytical basis of reliability modeling, and apply both deterministic and stochastic methods to product reliability analysis and design. This should be the new direction of the research on reliability technology in the near future. |
