A Complex Multi-level System Reliability Distribution Estimation Method Based on Bayesian Melding Method

As the complexity of system structure increases, system reliability estimation is challenging for lack of time and cost. How to enhance the system reliability distribution for complex multi-level structures is of great value. Besides, when there exists reliability distribution information of subsystems or components, how to fully use the limited information of different system levels for the system reliability distribution estimation has great significance. Therefore, considering that Bayesian Networks can construct complex system structures intuitively as well as evaluate the system reliability effectively, and the Bayesian Melding Method can integrate the multi-level information adequately, a complex multi-level system reliability distribution estimation method is proposed. The complex system structure is first decomposed to many blocks by BN. The single block system reliability distribution is obtained based on the Bayesian Melding Method with information integrated and Bayesian Networks with reliability estimated. The final system reliability distribution is estimated by merging the distributions obtained in different blocks so as to enhance the system reliability distribution with more information from different levels. To demonstrate and verify the effectiveness of the proposed method, one numerical example is employed.