Study On The Decision Fusion Of Structure Distributed Damage Detection

It is not easy to detect the damages of complicate structures by single Damage Detection Method (DDM), and it is necessary to combine several DDM together. Therefore, distributed damage detection algorism, the key of which is how to deal with various detection results for the same situation, is adopted in the paper.Firstly the structure damage detection and the decision fusion technologies are reviewed, including the present status, development, perspectives and the corresponding methods of which are introduced. Then, distributed damage detection frame is created and both damage detection resolution and decision fusion algorithm are discussed. When the Bayesian decision fusion theory is used in the same-resolution damage detection case, it is found that the fusion result is much better than any single DDM and the usage of information obtained from Health Monitoring System becomes much higher.Meanwhile, multi-resolution damage detection fusion is considered in the paper. Firstly three damage detection organs made by BP neural network for different resolution detection are studied as sub-DDM. Then support function S, object function T and the related fusion algorithms are put forward based on the D-S evidence theory. Therefore, the multi-resolution damage detection fusion is completed, through which the final damage detection result is gained, and the damage case which is undeterminable by single method can be identified using multi-resolution damage detection fusion. Finally the D-S evidence theory is applied to the same-resolution damage detection fusion to compare with the Bayesian algorithm and it is obtained that both fusion methods are helpful for the detection probability, but the former shows the superiority in the system performance.