A problem solving environment for multidisciplinary coupled simulations in computational grids
Introduction
Coupling of different simulation codes, each specialized for a specific physical regime, is becoming more and more important for numerical simulations, both in industry and in research. The reason is that in many real-world applications the interaction of different physical phenomena must be taken into consideration in order to achieve high quality predictions. Often the performance requirements for such complex simulations can be met only by exploiting distributed computing resources, for example, provided by a computational grid [1]. Managing these distributed resources compounds the complexity of handling the different elements of the coupled simulation. This increases the time for performing simulations and forms an upper bound for the complexity of a simulation being manageable.
To improve the building and managing of multidisciplinary coupled simulations—especially in computational grids—we have integrated existing software tools to a problem solving environment that allows users to easily perform these kind of simulations using a graphical user interface.
The software tools, that have been integrated, are as follows:
- •
The software integration and workflow management system TENT [2] as the combining framework for integrating the tools mentioned below and as the graphical user interface for setting up, starting, and monitoring the coupled simulation.
- •
The numerical code coupling library MpCCI [3], [4]. MpCCI is a library based on the Message Passing Interface (MPI) for data exchange and interpolation between different simulation codes.
- •
The grid-enabled MPI implementation MPICH-G2 [5] as the underlaying MPI library for the MpCCI library. MPICH-G2 allows one to start the MpCCI-coupled codes in computational grids built with the Globus Toolkit [6].
The new and major work during the integration of these different tools was the adaption of the MpCCI library to support MPICH-G2 by the generation of proper Globus Resource Specification Language (RSL) files for MPICH-G2 from the coupling control parts of the TENT system. The RSL file contains the resources (hostnames and number of processors), the path to the executables, and all other necessary information to start the codes in the grid.
To actually solve problems with this environment, a number of numeric simulation codes have been integrated into the system. For example, the CFD codes TRACE [7], FLOWer [8], and TAU [9] and the commercial structure mechanics codes ANSYS [10] and NASTRAN [11]. Also, many additional tools such as filters (pre- and post-processing tools) and visualization tools (e.g., AVS [12] and Tecplot [13]) have been integrated to prepare and visualize data files.
Compared to many other grid aware problem solving environments (PSEs) (e.g., such as the systems presented in [14]), the system presented in this paper was build by integrating already existing legacy software tools (TENT, MpCCI, MPICH-G2). The main goal was the development of an industrial strength system for multidisciplinary coupled simulations. Other aspects like free availability of the software were negligible; in particular the coupling library MpCCI is a commercial product but it is used in a variety of industrial applications and already integrated in many commercial tools.
Section snippets
Software tools
This section gives a short overview of the software tools used and technologies that have been integrated into the problem solving environment for coupled simulations.
Integration
This section describes how the different technologies have been combined into a system that allows one to setup and control the coupled simulation with a graphical user interface. First, we summarize the tasks that are necessary to setup and run a coupled simulation with TENT, MpCCI, and MPICH-G2. After that, some of these tasks are explained in more detail.
Fig. 5 shows how the tools are integrated in TENT. Each of the numerical application codes is provided with a (TENT) wrapper to allow users
Applications
This section describes some applications of the presented system to various coupling problems in the field of airplane and space craft design:
- •
Fluid-thermal-coupling: Simulation of an air cooled blade. The thermal load of a turbine blade becomes an increasingly important aspect in the design phase of a new turbine. The applied loads are an important factor in determining the lifetime of a blade. To take the heat loads into account already during the design process a coupled simulation of the gas
Conclusions and future work
We have described a problem solving environment for multidisciplinary coupled simulations in computational grids, that was built by integrating already existing software tools. The integration system TENT, as well as the coupling library MpCCI, are tools which are already being used for scientific and commercial application scenarios. The new integration with the grid-enabled MPI implementation MPICH-G2 allows one to start coupled simulations in computational grids easily because of its
Software availability
The integration system TENT is available on request and free of charge for non-commercial use, see the Website http://www.dlr.de/tent for contact information and download of binary packages. The coupling library MpCCI is a commercial product, see http://www.pallas.com/e/products/mpcci/ for details.
Andreas Schreiber received a diploma in industrial mathematics from Technical University Clausthal in 1996. He is currently a research scientist in the Simulation and Software Technology division of the German Aerospace Center (DLR) in Cologne, Germany, where he leads the Distributed Systems and Component Software branch. From 1997 to 1998 he worked as a software developer at the German Federal Armed Forces (Bundeswehr). In 2001, he has been a visiting scientist at the Argonne National
References (28)
- et al.
MEGAFLOW: Parallel complete aircraft CFD
Parallel Comput.
(2001) - et al.
MPICH-G2: a grid-enabled implementation of the message passing interface
J. Parallel Distrib. Comput. (JPDC)
(2003) The integrated simulation environment TENT
Concurr. Comput.: Practice Exp.
(2002)- MpCCI Website....
- et al.
A communication library to couple simulation codes on distributed systems for multi-physics computations
- MPICH-G2 Website....
- Globus Project Website....
- et al.
The generation of artificial counter rotating vortices and the application for fan-shaped film-cooling holes
The DLR-TAU Code—an Overview
Cited by (14)
Steering and visualization: Enabling technologies for computational science
2010, Future Generation Computer SystemsCitation Excerpt :Thus computational steering has itself grown up alongside Grid computing, and a technique with once-limited applicability now has much wider appeal. Modern computing infrastructures have therefore undoubtedly improved the potential of computational steering, but at the cost of increased complexity in the software that must be integrated and the hardware configurations it runs on (see e.g. [26]). This paper explains the principles and aims of steering and puts forward a simple reference model that allows current, distributed steering approaches to be classified easily.
Fluid structure interaction numerical simulation of wiresweep in electronics packaging
2016, Telkomnika (Telecommunication Computing Electronics and Control)Model of multidisciplinary simulation integration in helicopter rotor blade design process
2014, International Journal of Computer Integrated ManufacturingFSI analysis of wire sweep in encapsulation process of plastic ball grid array packaging
2013, Isi Bilimi Ve Teknigi Dergisi/ Journal of Thermal Science and TechnologyFSI simulation of wire sweep PBGA encapsulation process considering rheology effect
2012, IEEE Transactions on Components, Packaging and Manufacturing TechnologyFluid/Structure interaction investigation in PBGA packaging
2012, IEEE Transactions on Components, Packaging and Manufacturing Technology
Andreas Schreiber received a diploma in industrial mathematics from Technical University Clausthal in 1996. He is currently a research scientist in the Simulation and Software Technology division of the German Aerospace Center (DLR) in Cologne, Germany, where he leads the Distributed Systems and Component Software branch. From 1997 to 1998 he worked as a software developer at the German Federal Armed Forces (Bundeswehr). In 2001, he has been a visiting scientist at the Argonne National Laboratory working in the Globus Project group. His research interests include grid computing, high-performance computing, software engineering, and Web services.
Thijs Metsch is a student of information technology at the University of Cooperate Education (Berufsakademie) located in Mannheim, Germany, and he works in the Simulation and Software Technology division of the German Aerospace Center (DLR) in Cologne, Germany. His working fields include grid computing, user interface development, and automatic code generation.
Hans-Peter Kersken received his diploma in physics and his PhD from the University of Bonn, Germany, in 1989 and 1992, respectively. Currently, he is a scientific staff member in the Simulation and Software Technology division of the German Aerospace Center (DLR) in Cologne, Germany. Previously, from 1997 to 1998 he was with the Computing Center of the University of Stuttgart, Germany, and from 1992 to 1997 with the Scientific Computing Group at the Computing Center of the Alfred-Wegener Institute for Polar and Marine Research at Bremerhaven, Germany, where he worked in the field of parallel numerical mathematics and parallelization of ocean models. His research interests are in the fields of coupled computations, numerical mathematics, parallel computing, and computational fluid dynamics.