Optimal test plan for reliability demonstration of one shot systems

Optimal test plan for reliability demonstration of one shot systems

One-shot devices and one-shot systems are required to perform a function once only since their use is normally accompanied by an irreversible reaction or process, e.g. chemical reaction or physical destruction. One-shot devices and one-shot systems, electronic, mechanical or structural, by definition, cannot be fully tested prior to use because testing in the full mode of operation usually results in their total or partial destruction; or an irreversibility of a discrete part. The final system test is very difficult and expensive, therefore we should use component and sub-system tests to demonstrate reliability and number of test for integrated system should be decreased. But, one test is necessary. In this paper, we assume that components tested with zero failure and a hybrid model that integrates the Bayesian model with the variance propagation technique is implied. In this paper, we use GA to determine the best number of tests for each component to demonstrate system reliability. Target function is developed based on test cost and the best test plan for a desired reliability and confidence level is determined.

One-shot devices and one-shot systems are required to perform a function once only since their use is normally accompanied by an irreversible reaction or process, e.g. chemical reaction or physical destruction. One-shot devices and one-shot systems, electronic, mechanical or structural, by definition, cannot be fully tested prior to use because testing in the full mode of operation usually results in their total or partial destruction; or an irreversibility of a discrete part. The final system test is very difficult and expensive, therefore we should use component and sub-system tests to demonstrate reliability and number of test for integrated system should be decreased. But, one test is necessary. In this paper, we assume that components tested with zero failure and a hybrid model that integrates the Bayesian model with the variance propagation technique is implied. In this paper, we use GA to determine the best number of tests for each component to demonstrate system reliability. Target function is developed based on test cost and the best test plan for a desired reliability and confidence level is determined.