In this paper, the wave propagation mechanics theoretical model of hull grillage metastructures with local resonators is established, and the intelligent design method of hull grillage metastructures is proposed based on deep learning. The flexural vibration isolation characteristics of periodic hull grillage structures and the proposed hull grillage metastructures are theoretically studied using the spectral element method. Numerical calculation and experimental tests of the flexural wave vibration transmission of hull grillage metastructures are conducted to validate the proposed theoretical model. In addition, the data set is constructed by using the wave mechanics theoretical model of hull grillage metastructures, and the forward prediction neural network model of vibration transmission characteristics as well as the inverse design neural network model of hull grillage metastructures are established. Results show that compared with periodic hull grillage structures, the proposed hull grillage metastructures possess significant low-frequency flexural wave vibration isolation characteristics which almost completely covers the low-frequency range from 27 Hz to 169 Hz. Both forward prediction and inverse design network models have good performance, which verifies the correctness of the intelligent design method based on deep learning, and provides a new direction for the structural design on demand of low-frequency vibration reduction of ship and offshore structures.
Read full abstract