Estimation of a Craig-Bampton equivalent model from base excitation measurements

This paper investigates the potential of using data from base excitation tests to characterize the dynamic behavior of a structure, with a focus on its use in Dynamic Coupled Load Analysis (DCLA) for spacecraft launch phases. DCLAs are essential for verifying that spacecraft structures can survive launch loads and complete their missions. However, the accuracy of DCLAs relies heavily on the accuracy of the Craig-Bampton reduced form of the spacecraft's finite element model, which can be time-consuming to validate. In contrast, the proposed approach synthesizes a mathematical model of a generic structure using data from base excitation tests, offering a potential alternative to the CB model. The system identification problem is solved using a hybrid Particle Swarm Optimizer, with operators from Evolutionary Strategies and a local optimization algorithm to improve the solution and prevent premature convergence. The relationships between a continuous linear structure and a valid discrete model are studied, and a new strategy is implemented to define the search space.