Prediction of wave overtopping discharge on coastal protection structure using SPH-based and neural networks method
Dang, B.-L.; Dang, Q.V.; Abdel Wahab, M.; Nguyen-Xuan, H. (2022). Prediction of wave overtopping discharge on coastal protection structure using SPH-based and neural networks method, in: Abdel Wahab, M. (Ed.) Proceedings of the 4th international conference on numerical modelling in engineering. Lecture Notes in Civil Engineering, 217: pp. 71-79. https://dx.doi.org/10.1007/978-981-16-8185-1_6 In: Abdel Wahab, M. (Ed.) (2022). Proceedings of the 4th international conference on numerical modelling in engineering. Lecture Notes in Civil Engineering, 217. Springer: Singapore. ISBN 978-981-16-8184-4; e-ISBN 978-981-16-8185-1. IX, 100 pp. https://dx.doi.org/10.1007/978-981-16-8185-1, more In: Lecture Notes in Civil Engineering., more |
| Available in | Authors | | Document type: Conference paper
|
| Keyword | | | Author keywords | coastal structure; wave overtopping; SPH model; neural networks |
| Authors | | Top | - Dang, B.-L., more
- Dang, Q.V.
- Abdel Wahab, M., more
- Nguyen-Xuan, H.
| | |
| Abstract | This paper adopts DualSPHysics, the powerful SPH models, to investigate a large-scale 2-D numerical simulation of wave-structure interactions. As a case study, a non-conventional seawall structure built at Vietnam's coastline is considered. The hydraulic performance of such a structure is assessed using the value of wave overtopping over structure. It is one of the most important considerations when evaluating the efficiency of proposed designs. Due to the geometrical differences, traditional methods such as empirical equations are inconvenient for analyzing such novel structure design with complicated shapes. As a supplement to the experimental study, numerical modeling and machine learning approaches are being studied for assessing such problems. The reliability and effectiveness of two approaches have been proven in several studies in literature. In this work, a large-scale computational model of wave-structure interaction under regular wave conditions is carried out. The simulation results demonstrate good agreement when compared to neural network-based prediction approaches, and analytical solution as well. |
|
