Forces Identification Of Skeletal Structures Based On Tikhonov Regularization And Genetic Algorithm

Forces identification is very important to not only the structural health monitor, but also to the daily maintenance and assessment of structures. In this thesis, the forces identification of skeletal structures from the measured response by using both Tikhonov regularization and Genetic Algorithm (GA) is studied.Up to now, Tikhonov regularization approach has been one of the most effective ways to solve ill-posed problems. In this approach, the appropriate choice of a regularization parameter is the most important. Considering the regularization solution should be close to the real value, the regularization parameter must be smaller; and from the view of numerical stabilization, the parameter must be larger. Usually, iterative methods are used for obtaining the parameter. However, the iterative solutions are often sensitive to the initial values of the parameter, and different initial values may lead to quite different solutions. Presently, a method based on the Genetic Algorithm (GA) is proposed, which can achieve an optimal solution of the parameter in the whole range and therefore provide an efficient way for the parameter calculation. Here the generalized cross validation criterion, L-curve method and the Engl's criterion are used as the objective functions of GA.The forces of skeletal structures are identified using the method mentioned above. The relation between loads and strains of structural elements is deduced. Meanwhile, the effects of the measurement number, their positions and error distribution on the identification results are discussed. Accuracy of the method is demonstrated through numerical results.Furthermore, the dynamic loads of a structure are identified by using dynamic stiffness matrix through Tikhonov regularization and Genetic Algorithm (GA). Numerical results also show that good results can be achieved.