High-Speed Neural Network Tsunami Predictor with Multiple Adjustable Parameters

Overview

Due to global warming, glaciers have been melting at an accelerated rate, causing coastlines to encroach further inland. As a result, the danger that tsunamis pose to wildlife and the built environment has worsened. Thus, to understand the impacts of these tsunamis, there is a need for a fast and accurate tsunami simulation tool. To this end, we have created a neural network that can simulate tsunami waves for two arbitrary parameters: ocean basin steepness and ocean basin depth. Our neural network is up to 117 times faster than a traditional solver with a mean squared error of only 8.678E-06. While the neural network currently accepts two parameters, we plan to add enough parameters for the neural network to understand even the most complex ocean basin topographies. Our neural network approach will provide a high-speed tool for researchers to study tsunamis. Ultimately, our tool will allow for planners in coastal communities to appropriately design and place levees and property boundaries.

Competition history

• AJAS 2024 Category not listed

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

• All Earth and Environmental Sciences projects
• AJAS Earth and Environmental Sciences projects
• Every AJAS 2024 project