A hybrid surrogate vibration optimization method for the propulsion shaft system in an unmanned underwater vehicle

A hybrid surrogate-assisted vibration optimization method is proposed for the propulsion shaft system of the underwater vehicle. The previous studies mainly analyzed the influences of selected parameters on the vibrations. Thus, the proposed method performs an automatic multiparameter optimization of the propulsion shaft system. The intelligent optimization algorithms and surrogate model methods are combined. The agents generated by the intelligent optimization algorithms are used to establish the surrogate model, which could replace the real model in the iterations of the intelligent optimization algorithm to reduce the calculating consumption. The real model is used to find the local optimal at the end steps. The proposed method determines the optimal parameter configuration of the propulsion shaft system to reduce vibration. Compared with the full real model method, the proposed approach reduces the computational time by 18.3%, while the deviation of the final objective function value is only 0.6%. These results confirm the efficiency and accuracy of the proposed method.