Damage Detection of the Rod in the Crossflow Using Surrogate-Based Modelling

An effective and accurate methodology is developed to create an inverse surrogate model for the mass reduction analysis of the rod in the rod bundle inserted in the crossflow. The performance of two surrogate modelling approaches has been evaluated. These models are the Response Surface Method and Legendre polynomial approximations. The relationship between dominant frequencies, support stiffness and rod mass derived from Computational Fluid Dynamics simulations is used as input data for approximations. The selection of sample points is implemented with a new type of orthogonal design. The results have shown that the proposed methodology can reliably replace the finite volume model and drastically reduce computational time.