An improved Taguchi method and its application in finite element model updating of bridges

In the model updating process, the objective function is usually set as the weighted sum of the difference between analytical and experimental dynamic characteristics. But it is difficult to select the weighting factors since the relative importance of each parameter to updated results is not obvious but specific for different problem. To overcome this problem, multi-objective genetic algorithm (GA) is introduced into model updating by Gyeong-Ho Kim since there is no need for selecting weighting values in multi-objective optimization technique. To complex structures, however, it is difficult to update the structural models by GA because of the relative low efficiency. While Taguchi updating method, deemed as an efficient and robust method, is a good choice to update the models of large structures. But Taguchi method is only applied to solve the single objective optimization problem of model updating. Therefore, this paper proposed improved Taguchi updating method to deal with the problem of model updating using multi-objective optimization technique. Then the proposed method is applied to update the model of a steel suspension bridge with measured modal parameters. The updated results show that the proposed method is promising to structural model updating. © (2011) Trans Tech Publications, Switzerland.