PSI - Issue 39

Nabam Teyi et al. / Procedia Structural Integrity 39 (2022) 608–623 Author name / StructuralIntegrity Procedia 00 (2019) 000–000

615

8

safeguards against genetic loss. The crossover operator is the primary recombination operator that enables communication between candidate solutions. The number of iterations is a frequently used criterion for termination. The capacity of GAs to tackle non-linear, discontinuous, and poorly understood optimization problems is widely established. As a result, GAs solve inverse crack detection problem . A general flow of GA algorithm is shown in Fig. 8.

Fig. 8. Genetic algorithm.

He et al. (2001) proposed a numerical experimentation method for dealing with inverse problems based on GAs. According to preliminary findings, their approach may be utilised to address or pave the way for a wide variety of inverse flaw or defect detection and identification difficulties. The primary advantages of this methodology over conventional inverse problem search techniques are the avoidance of local optima and the absence of mathematical difficulties. GAs take longer than ten hours, even with tiny populations. Therefore, combining GAs with additional optimization techniques such as simulated annealing (SA) or gradient-based approaches should be considered. Saridakis et al. (2006) and Saridakis et al. (2007) calculated the location, depth, and relative angle of two cracks in a shaft. To solve the inverse crack identification problem, their analytical model was approximated using an ANN. By comparing the outputs to empirically measured responses, the GA attempted to find a solution. They used two of the five possible fuzzy logic objective functions. The proposed method’s accuracy was determined by the length of training for the neural network approximation. The stochastic nature of the evolutionary algorithm resulted in suboptimal results. The computing time was cut in half by using a neural network to approximate the analytical model and a GA whose objective function was based on a fuzzy logic representation to replace the exhaustive search of the solution space. They employed a mixed heuristic by subjecting the GA solution to a pattern search optimization approach. This meant fewer generations of GAs and less computational time. The framework may be applied to approximate various analytical models of multiple-crack shafts. Xiang et al. (2008) developed a method for locating and measuring shaft cracks. To detect cracks, the technology employed was wavelet-based elements and a GA. The GA was used to reduce the frequency differences between numerical simulation and experimental measurement. The proposed method could both numerically and experimentally detect a crack in a shaft. Vakil-Baghmisheh et al. (2008) developed a GA based defect identification approach. The cracked-beam structure was modelled using a model of a damaged cantilever beam based on analysis. Their technology employed evolutionary algorithms to monitor changes in the structure’s natural frequencies. The ideal location and depth of the fracture in the cantilever beam were determined by minimizing the cost function with respect to the difference between the measured and calculated natural frequencies. Agarwalla et al. (2015) used a GA-Fuzzy controller to detect damage in a steel cantilever beam subjected to natural frequency transversely. They discovered that the presence of cracks had a significant impact on the natural frequencies and mode shapes of the beam in question. Parhi et al. (2011) investigated