Doctor Abel Coll
Abel Coll is a Doctor in Civil Engineering from the Technical University of Catalonia (BarcelonaTech). He is the Head of the Pre and Post-processing department, also known as GiD Group, of the International Center for Numerical Methods in Engineering (CIMNE) at Barcelona, Spain. CIMNE was created in 1987 and specializes in the development and applications of numerical methods and software to find solutions to a variety of problems in engineering and applied sciences.
He begun its research at CIMNE on 2004 focusing on Computational Geometry and Mesh Generation for numerical simulations, and has collaborated in a wide list of European and National research projects related to different scientific and engineering fields (such as Computational Fluid Dynamics, Computational Electromagnetics, Computational Mechanics, Structural analysis or multi-physics problems). His research interest and expertise include preprocessing operations, including computer aided design (CAD) tools, mesh generation and software integration), as well as post-processing operations and visualization of results attached to meshes and coming from numerical simulations. Other research lines of his interest are cloud computing, high performance computing and Big Data for simulation software. Most of this work is part of the development of the GiD suite, a software developed and marketed by the GiD group at CIMNE, for pre and postprocessing data for engineering simulation software (www.gidhome.com).
This talk will present an overview of VELaSSCo project (www.velassco.eu), which is one of the pioneer European projects proposing to merge the Big Data technology with the engineering simulation tools in order to allow the post-processing of huge simulation data (data analytics) and its visualization in an efficient and a more understandable way for the decision making process, what is currently a bottleneck in the simulation pipeline.
Talk: Using Big Data for extreme large scale simulations analysis and visualization.
Numerical simulations are often included in some Decision Support Systems (DSS) when accurate predictions of the physical effects of some events in the domain of study are necessary for the assessment of the best solution. As the phenomena to be studied grows in complexity, or the domain space where the phenomena occurs increases, the related numerical simulation generates greater amounts of data to be analyzed and processed, usually in the range of billions of records.
For these huge simulations it is common to use High Performance Computing (HPC) infrastructures, which typically work on distributed machines with the goal of speed-up the computation times. This generates big amounts of distributed data that need, first, to be post-processed (new data must be generated from them, as iso-surfaces, stream-lines, color maps, etc…) and, second, to be visualized in a reasonably interactive time, almost real-time.
A tool for post-processing and visualize the results of huge numerical simulations in an efficient way is crucial to provide a rapid understanding of the effect of some event in the environment of study for the DSS, and allow a fast integration of this information in the whole decision making process. Furthermore, it is common to use techniques as Artificial Neural Networks (ANN), Model Reduction or optimization techniques based on genetic or Monte-Carlo methods, which also need to run a big number of simulations before the data from the simulation can be reliable. This fact increases even more the data from simulations that need to be processed before feeding the DSS.
VELaSSCo project, Visual Analysis for Extremely Large-Scale Scientific Computing (www.velassco.eu) is a FP7 research project led by CIMNE, funded by the European Community’s Seventh Framework Programme (FP7/2007-2013) under grant agreement n° 619439, which has developed a post-processing and visualization platform for huge distributed simulation data using Big Data technologies and architectures. This talk will give an overview of VELaSSCo project and its potential for integrating huge numerical simulations in the DSS.