|
Title:
|
A WEB COMPUTATIONAL PLATFORM FOR LITTORAL HYDRODYNAMICS MODELLING |
|
Author(s):
|
Romain Chailan, Fabien Rétif |
|
ISBN:
|
978-989-8533-45-6 |
|
Editors:
|
Hans Weghorn |
|
Year:
|
2015 |
|
Edition:
|
Single |
|
Keywords:
|
High Performance Computing, Web services, Hydrodynamics Modelling. |
|
Type:
|
Short Paper |
|
First Page:
|
171 |
|
Last Page:
|
175 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Weather and climate forecast are sciences of great interest considering their major impact on our society, e.g., flash flood or storm forecasting. Such a forecast relies on complex physical processes, which are in practice solved by resource demanding numerical models. For instance, a regional sea-state conditions forecast requires using various models. This implies technical issues, like handling large amount of IO data and spending thousands of hours dedicated to their computation, their extraction and their analysis. Coastal engineers and scientists are generally not IT engineers. They need tools to avoid big data manipulations and the computational burden out-coming from such modelling processes. We propose a solution aiming at easing the simulation of numerical models used in the study of littoral hydrodynamics processes. The proposed solution is based on a Web Oriented Architecture (WOA) and can be viewed as a Scientific Workflow Management System (SWfMS). The strength of the solution is to bring to end-users a friendly platform able to seamlessly chain hydrodynamics model, to transform and to explore multidimensional IO data sets with an ad-hoc visualisation tool. This approach also offers a user-transparent way to run numerical models on a High Performance Computing (HPC) cluster. The solution is demonstrated by chaining two numerical models in the goal of analysing coastal hazard in extreme conditions. |
|
|
|
|
|
|
