The damage level assessment of equipment function based on Bayesian networks and transfer learning

Abstract

The damage level assessment of equipment function is an important part of equipment battle damage assessment. In practice, it is often difficult to obtain accurate damage level assessment results due to a lack of damage test data and insufficient modeling. Aiming at this problem, a functional damage assessment method based on Bayesian networks and transfer learning is proposed in the case of small sample test data. First, a Bayesian network model considering the correlation of component damage is constructed, which can more accurately reflect the damage results of equipment when incomplete damage information is obtained. Then, an improved TrAdaboost transfer learning method is proposed for the Bayesian network model, which overcomes the disadvantage that the traditional TrAdaboost method is unable to transfer the results with randomization. Finally, the method proposed in this paper is applied to the Asia network and a certain type of radar vehicle functional damage level assessment process, and the results prove the effectiveness and superiority of the proposed method.