A robust optimization for damage detection using multiobjective genetic algorithm, neural network and fuzzy decision making

An inverse problem of damage identification and localization in a structure is modelled as a robust optimization problem. In the robust optimization problem, the optimum value and small variations around this optimum value are considered. The structural health monitoring damage detection problem is solved using a multiobjective genetic algorithm. So, the robust optimum value is obtained by solving a multiobjective problem where a functional and a variance function of this functional are used. This variance function is obtained by a Design of Experiment with regression and also through a relation between functional variance and damage parameters found by artificial neural network. As a multiobjective genetic algorithm obtains multiple solutions, a fuzzy decision making technique finds the better tradeoff solution for the problem. Boundary element method is utilized to obtain the distribution of stress to elastostatic problem. Numerical results clearly show that the proposed strategy and the use an optimized fuzzy decision making results in accurate damage identification and represents a powerful tool for structural health monitoring. Based on the analysis and numerical results, suggestions to potential researchers have also been provided for future scopes.