SCCS Colloquium - Feb 21, 2019
|Date:||Feb 21, 2019|
|Time:||15:00 - 15:30|
Arbab Akhtar: Monte Carlo Methods for Shallow Water Equations
This is a Master's Thesis introduction talk advised by Anne Reinarz.
The partial differential equations for shallow water equations are solved with ADER-DG method in ExaHyPE where uncertain measurements are quantified with Monte Carlo Method.
Keywords: Shallow Water Equations, ADER-DG, Monte Carlo Method
Vyshakh Unnikrishnan: Implementation of a deep learning based model for rainfall-runoff modelling
This is a Master's Thesis introduction talk advised by Ivana Jovanovic.
Rainfall-runoff modelling is vitally important from a hydrology perspective, i.e. accurate runoff prediction of the big watercourses leads to better preparation when facing a flooding scenario. In this thesis, we aim to develop a data driven model, more precisely a long short term memory (LSTM) recurrent neural network architecture (RNN) that can accurately predict the daily runoff values of a particular region. Rainfall data together with other meteorological data for the past period would be used as the input to the LSTM-RNN model. We hypothesize that the deep learning model trained with the vast amount of data will be able to adequately model the hydrological processes and storage effects in the region, and capture the dynamics of a full annual cycle. Moreover, we believe that the properly trained data driven model might eventually be used to enhance, or even substitute, the complex and computationally expensive physically based hydrological models used for the purpose of runoff prediction.
Keywords: Deep Leraning, LSTM Neural Networks, Hydrological Model