Warship reliability evaluation based on dynamic bayesian networks and numerical simulation

Abstract The systematic implementation of reliability engineering is a necessary measure to improve warship reliability. In warship reliability engineering, reliability evaluation is a key step because it reveals whether warships feature an acceptable level of reliability. Current reliability evaluation methods, such as the analytic method, multilevel synthesis method and numerical simulation method, are all based on static logic. However, warships are typical dynamic systems, and their reliability cannot be evaluated using the aforementioned methods. On the basis of the characteristics of warships, including their small sample system and dynamic and multistage missions, this study proposed a new reliability evaluation method based on dynamic Bayesian networks and numerical simulation. The proposed method is the inheritance and development of current methods for reliability evaluation based on numerical simulation. Hence, it overcomes the limitations of the analytic method and multilevel synthesis method, as well as provides an effective means for the reliability evaluation of warships. The proposed method is suitable for the overall reliability evaluation not only of warships but also of other complex dynamic systems.