Approximation capabilities of multilayer feedforward networks
Neural Networks
Hi-index | 0.00 |
One of major difficulties with numerical tidal models is accurate inversion of open boundary conditions. A data-driven model based on artificial neural network is developed to retrieve open boundary values. All training data are calculated by numerical tidal model, so the tidal physics are not disturbed. The basic idea is to find out the relationship between open boundary values and the values of interior tidal stations. Case testes are carried out with a real ocean bay named Liaodong Bay, part of the Bohai Sea, China. Four major tidal constituents, M2 , S2 , O1 and K1 , are considered in coupled inversion method. Case studies show that the coupled inversion for open boundary conditions can make a more satisfactory inversion for a practical problem.