-
Storm Surge Modeling in the AI ERA: Using LSTM-based Machine Learning for Enhancing Forecasting Accuracy
Authors:
Stefanos Giaremis,
Noujoud Nader,
Clint Dawson,
Hartmut Kaiser,
Carola Kaiser,
Efstratios Nikidis
Abstract:
Physics simulation results of natural processes usually do not fully capture the real world. This is caused for instance by limits in what physical processes are simulated and to what accuracy. In this work we propose and analyze the use of an LSTM-based deep learning network machine learning (ML) architecture for capturing and predicting the behavior of the systemic error for storm surge forecast…
▽ More
Physics simulation results of natural processes usually do not fully capture the real world. This is caused for instance by limits in what physical processes are simulated and to what accuracy. In this work we propose and analyze the use of an LSTM-based deep learning network machine learning (ML) architecture for capturing and predicting the behavior of the systemic error for storm surge forecast models with respect to real-world water height observations from gauge stations during hurricane events. The overall goal of this work is to predict the systemic error of the physics model and use it to improve the accuracy of the simulation results post factum. We trained our proposed ML model on a dataset of 61 historical storms in the coastal regions of the U.S. and we tested its performance in bias correcting modeled water level data predictions from hurricane Ian (2022). We show that our model can consistently improve the forecasting accuracy for hurricane Ian -- unknown to the ML model -- at all gauge station coordinates used for the initial data. Moreover, by examining the impact of using different subsets of the initial training dataset, containing a number of relatively similar or different hurricanes in terms of hurricane track, we found that we can obtain similar quality of bias correction by only using a subset of six hurricanes. This is an important result that implies the possibility to apply a pre-trained ML model to real-time hurricane forecasting results with the goal of bias correcting and improving the produced simulation accuracy. The presented work is an important first step in creating a bias correction system for real-time storm surge forecasting applicable to the full simulation area. It also presents a highly transferable and operationally applicable methodology for improving the accuracy in a wide range of physics simulation scenarios beyond storm surge forecasting.
△ Less
Submitted 7 March, 2024;
originally announced March 2024.
-
Ab initio investigation of the effects of B-do** on the adsorption of H2O, H2 and O2 molecules at diamond surfaces
Authors:
Stefanos Giaremis,
Maria Clelia Righi
Abstract:
Boron doped diamond is extensively studied for its use in tribological and electrochemical applications due to its remarkable physical and chemical properties. However, ambient conditions play a major role to its macroscopically observed behavior. In this study, the fundamental interactions between the low Miller index (001), (110) and (111) B-doped diamond surfaces with H2O, H2 and O2 molecules,…
▽ More
Boron doped diamond is extensively studied for its use in tribological and electrochemical applications due to its remarkable physical and chemical properties. However, ambient conditions play a major role to its macroscopically observed behavior. In this study, the fundamental interactions between the low Miller index (001), (110) and (111) B-doped diamond surfaces with H2O, H2 and O2 molecules, which are commonly present in ambient air and commonly involved in electrochemical reactions, are investi-gated by means of ab initio simulations. The results are presented in close comparison with previous studies on undoped diamond surfaces to reveal the impact of B on the adsorption properties. It is demonstrated that the B dopant is preferably incorporated on the topmost carbon layer and enhances the physisorption of H2O by forming a dative bond with O, while, in some cases, it can weaken the ad-sorption of O2, compared to the undoped diamond. Moreover, a noticeable displacement of the surface atoms attached to the fragment of the dissociated H2O and O2 molecules was observed, which can be associated to the first stage of wear at the atomistic level. These qualitative and quantitative results aim to provide useful insight towards the development of improved protective coatings and electrochemical devices.
△ Less
Submitted 27 April, 2023;
originally announced April 2023.
-
Density functional description of long-range electron Coulomb interactions in bulk SnS
Authors:
Stefanos Giaremis,
Joseph Kioseoglou,
Eleni Chatzikyriakou
Abstract:
A high-throughput benchmarking technique for testing the performance of different exchange-correlation functionals and pseudopotentials is proposed and applied to bulk SnS. It is shown that, contrary to the popular view that the local density approximation can best describe layered materials, a semilocal pseudopotential with a functional having a gradient dependence better described lattice vector…
▽ More
A high-throughput benchmarking technique for testing the performance of different exchange-correlation functionals and pseudopotentials is proposed and applied to bulk SnS. It is shown that, contrary to the popular view that the local density approximation can best describe layered materials, a semilocal pseudopotential with a functional having a gradient dependence better described lattice vectors and `tetragonicity' of the lattice. We classify the pseudopotentials based on this value and show that the participation ratio of maximally localized Wannier functions follows the theory which states that more distorted structures have higher anti-bonding hybridization as stabilizing factor. In order to classify pseudopotentials, the local and nonlocal potential contributions to the dynamical Born effective charges are taken for each pseudopotential. Finally, a strategy is proposed for learning exchange-correlation functionals based on the distinction between short and long range parts of the Kohn-Sham potential.
△ Less
Submitted 19 May, 2023; v1 submitted 20 February, 2023;
originally announced February 2023.
-
Decorated dislocations against phonon propagation for thermal management
Authors:
Stefanos Giaremis,
Joseph Kioseoglou,
Mykola Isaiev,
Imad Belabbas,
Philomela Komninou,
Konstantinos Termentzidis
Abstract:
The impact of decorated dislocations on the effective thermal conductivity of GaN is investigated by means of equilibrium molecular dynamics simulations via the Green-Kubo approach. The formation of "nanowires" by a few atoms of In in the core of dislocations in wurtzite GaN is found to affect the thermal properties of the material, as it leads to a significant decrease of the thermal conductivity…
▽ More
The impact of decorated dislocations on the effective thermal conductivity of GaN is investigated by means of equilibrium molecular dynamics simulations via the Green-Kubo approach. The formation of "nanowires" by a few atoms of In in the core of dislocations in wurtzite GaN is found to affect the thermal properties of the material, as it leads to a significant decrease of the thermal conductivity, along with an enhancement of its anisotropic character. The thermal conductivity of In-decorated dislocations is compared to the ones of pristine GaN, InN, and random and ordered InxGa1-xN alloy, to examine the impact of do**. Results are explained by the stress maps, the bonding properties and the phonon density of states of the aforementioned systems. The decorated dislocations engineering is a novel way to tune, among other transport properties, the effective thermal conductivity of materials at the nanoscale, which can lead to the manufacturing of interesting candidates for thermoelectric or anisotropic thermal dissipation devices.
△ Less
Submitted 17 December, 2019;
originally announced December 2019.
