A novel ANN-RDT approach for damage detection of a composite panel employing contact and non-contact measuring data.

A non-destructive method is proposed in this study for damage detection of a composite panel using random decrement (RD) signature technique and artificial neural network (ANN) algorithm. For this purpose, a 10-layer rectangular composite panel is built and the experimental test is carried out employing contact and non-contact sensing equipment to capture the vibration characteristics of the system. The experimental scenarios are defined using both intact and damaged models in two pinned and fixed supporting condition. The free damping response of the structure is extracted by RD method and employed for training the ANN neurons. The proposed ANN-RDT method is compared with the multimode random decrement technique (MRDT) and analytical mode decomposition- random decrement technique (AMD-RDT) method in damage detection process considering two error indexes. It is shown that the present method performs better than the MRDT and AMD-RDT. According to the results, the proposed approach is shown to be effective for diagnosing the presence of damage in both specified supporting systems. The trained algorithm performed more successful in the scenarios with pinned conditions with a damage closer to the supports. As indicated, the efficiency of the method is highly sensitive to location of the damage and sensors along with the type of the support. The experimental non-contact data is also shown to be significantly beneficial to utilize a simpler linear equation and also to gain the system dynamic characteristics. [ABSTRACT FROM AUTHOR]