A practical method for reliability evaluation of hybrid system under unit life submitted with different distribution

Life-type unit refers to the unit with knowable life distribution type, and hybrid system refers to the system composed of different life-type units. Our aim is to evaluate the hybrid system reliability, which is crucial for improving product competitiveness in heavy industries such as automobile tires and rocket engines, and reducing the maintenance costs. The Mean Confidence Limit Equivalent Folding method is employed in the study. Firstly, extracting a life-type unit and supposing a fitting function, the least squares method combines to verify and modify the supposed fitting function through iterative computations. Then, reliability point estimation and confidence lower limit of each life-type unit are calculated respectively, the formula of turning life-type data into success or failure-type data is derived, and the success or failure-type data is analyzed. Finally, compared with other methods such as the Monte Carlo method and the Levenberg-Marquardt method, the feasibility of the Mean Confidence Limit Equivalent Folding method is verified. The compared results show that a confidence lower limit error based on the method is about 1%, it can meet the parameter evaluation requirement of hybrid systems under constituent units obeying different distribution. The Mean Confidence Limit Equivalent Folding method considers the consistency of the confidence interval of reliability, and it can be suitable for all hybrid systems.