Comparative study of machine learning and deep learning in predicting the shear strength of marine sand and polymer layer interfaces interface under marine temperature effects

In marine engineering, polymer layers are anti-seepage barrier materials. The mechanical interaction between marine sand and polymer layer significantly affects overall structural stability. In this study, direct shear tests at different temperatures in the marine environment are simulated to evaluate the shear behavior of marine sand and polymer layer interface, and a database is developed. Based on the experimental data, the study employs the Back propagation Neural Network (BPNN), Genetic Algorithm and Particle Swarm Optimization BPNN, and convolutional neural network (CNN) models, which are trained and tested. The findings show that the CNN algorithm significantly outperforms other models in terms of prediction accuracy and efficiency. Sensitivity analysis shows that temperature, shear displacement, normal stress, and particle size have influence on interfacial shear strength, and the impact of normal stress is the greatest. In addition, an empirical formulation is proposed to provide tools for those without machine learning. Based on the research results, the deep learning CNN model developed in the study can accurately predict the shear strength of the interface between marine sand and the polymer layer, which provides an effective tool for the design and optimization of marine engineering.