Fault detection in beam structure using adaptive immune based approach

Different structural and machine elements are used over the ages. These are subjected to various loads like static and dynamic load, temperature, corrosion etc. Due to the above-mentioned reasons, ageing of the structural elements occur. So, to enhance the designed lifetime of any structure continuous maintenance is required. One such method has been proposed in this research work and the proposed method can be employed as an online tool for the fault identification. Here dynamic analysis of structure has been conducted as the forward method to find out the modal natural frequencies related with the damage. Recently with the application of machine learning approaches and the soft computing, the damage can be detected easily. In this methodology, Clonal Section Algorithm (CSA) has been applied to find out the faults (crack locations and depth) in the structure initially. Later one such method has been developed in the concepts of adaptive immune based technique (Adaptive Clonal Section Algorithm-ACSA) which is the combination of an artificial immune (Clonal Selection Algorithm) and Regression Analysis (RA). The use of regression analysis makes the proposed method more adaptive and the residual error in the collection of vibration data is reduced. The mechanism and various steps involved in CSA, RA and ACSA are analyzed here in a precise manner. The key endeavor of this study is the development of ACSA and its implementation to condition monitoring of structure. To authenticate and check the accuracy of both the methods (CSA and ACSA), laboratory tests are carried out. The results obtained from each method are corroborated with other and found to be convergent.