Sorted by business/research areas:
Coupling Environment
The »MpCCI CouplingEnvironment« has been developed at Fraunhofer SCAI in order to provide an application independent interface for the coupling of different simulation codes. MpCCI CouplingEnvironment has been accepted as a »de facto« neutral standard for simulation code coupling.
Download MpCCI_CouplingEnvironment [PDF, 1,1 MB]
Aerodynamic Loads in a Full-Vehicle NVH Analysis
Many vibration and acoustic effects in full vehicles are caused by fluctuating aerodynamic loads. Related tasks in automotive development are
• analysis of underbody paneling and attachment to body,
• analysis of engine hood flatter,
• interior acoustic analysis with external aero-acoustic loads.
In order to perform realistic noise vibration harshness (NVH) analysis, the exciting aerodynamic forces have to be taken into account.
Download MpCCI_FSIMapper_NVH [PDF, 659 KB]
Simulation of Electro-Magnetic Induced Vibrations
The rotating electro-magnetic field in electric motors produces periodic forces which excite vibrations of the structure and generate noise in the surroundings. On the one hand the electro-magnetic forces are computed in an EM-code. On the other hand the vibrations are computed in Computational Structural Dynamics (CSD) codes.
Download MpCCI_FSIMapper_EMAG_Vibrations [PDF, 918 KB]
Simulation of Flow-Induced Vibrations in Turbomachinery
Due to the interaction between rotating and stationary turbomachinery components, the transient fluid flow is characterized by periodic pressure fluctuations. Those oscillations excite vibrations of the blades.
Download MpCCI_FSIMapper_Flow_Vibrations [PDF, 618 KB]
Mapper – Independent Mapping Tool for Integrated Simulation Workflows
Mapper provides advanced and robust methods to map, compare and transfer simulation results and experimental data in integrated simulation workflows. Mapper supports a growing number of native file formats and can be used in a variety of engineering applications. Mapper is a
standard tool in the engineering departments of most German automotive OEMs and has been validated in the VDA / FAT working group »Formed Chassis Parts« for Forming to Crash workflows.
Optimization Using CFD Simulation Services
When developing technical devices andcomponents one is strongly interested in an efficient functionality of its design, meeting the requirements in a best manner. As numerical simulations have been well established in virtual design and computation times were further reduced, the application of sophisticated optimization algorithms gained lately in importance.
Download CFD_Optimization [PDF, 367 KB]
MpCCI Flowmaster Adaptor 1D-3D Simulations
The increasing demand for shorter product development lifecycles and fewer physical tests while maximising quality is challenging organizations to bring together their ‘best-in-class’ simulation tools for each discipline and to enable them to work together in pursuit of virtual prototypes for entire systems.
Download MpCCI_CouplingEnvironment_Flowmaster [PDF, 376 KB]
Simulation and Inversion of Wave Fields
The aim of the project WAVE is to develop a toolbox that provides methods for the simulation and inversion of wave fields on high performance architectures. WAVE is funded by the BMBF (German National Ministry of Education and Research). The project started in March 2016 and ends by July 2019.
Download WAVE [PDF, 347 KB]
Fast Hardware Independent Software Development
LAMA is a framework for developing hardware-independent, high performance code for heterogeneous computing systems. It facilitates the development of fast and scalable software that can be deployed on nearly every type of system, from embedded devices to highly parallel supercomputers, with a single code base. By using LAMA for their application, software developers benefit from higher productivity in the implementation phase and stay up to date with the latest hardware innovations, both leading to shorter time-to-market.
Download LAMA [PDF, 289 KB]
(Semi-)automatic code optimization and parallelization
Fraunhofer SCAI has a longterm working experience in massively parallel computing and the corresponding optimization. The HPC group at Fraunhofer SCAI is developing mechanisms to extract the maximal potential of a given hardware through
(semi-)automatic code transformation and parallelization techniques.
Download Compiler [PDF, 423 KB]
SAMG-Reservoir: a fast and robust solver for various kinds of reservoir simulations
SAMG-Reservoir handles linear systems with all relevant physical unknowns, and the coupling between them, in a robust and efficient manner via the SAMG-Coupled module. Only minimal additional information about the underlying physics are required. This is what reservoir simulators need: In fully and adaptive implicit simulations, the linear systems arising involve different, yet coupled types of unknowns with different physics behind them.
SAMG for groundwater simulation: mesh-agnostic yet fast
In order to produce reliable results in minimal computational time, groundwater simulations have to employ robust and efficient linear solvers. SAMG-Modflow allows for exploiting the most efficient solver algorithms via a simple plug-in. Hierarchical solver approaches like algebraic multigrid have proven to provide virtually optimal efficiency for numerical solution methods of diffusion problems such as pressure corrections without user interactions.
Accelerating MODFLOW with the enhanced SAMG linear solver software
SAMG is a software package for the highly efficient numerical solution of large sparse matrix problems. These problems occur in application areas such as fluid flow, structural mechanics, oil reservoir and ground water simulation, casting and molding, process and device simulation in solid state physics and circuit simulation. SAMG's solver modules incorporate modern hierarchical techniques (algebraic multigrid, AMG), making them especially suitable for solving large classes of discretised elliptic differential equations. Compared to classical solution methods, SAMG's solver modules have a major advantage in that they are numerically scalable. That is, the computational cost is only linearly proportional to the number of unknowns.
Fast auto-parallel linear solver for huge engineering applications
The original SAMG is a very efficient linear solver library based on algebraic multigrid (AMG), specifically developed for industrial applications. SAMG supports both serial and multi-core computations on single PC, workstations or compute nodes.
The new XSAMG library additionally exploits the parallelism offered by multiple nodes of a compute cluster by distributing linear systems across different nodes (based on MPI). Since this is done automatically, the calling simulation program itself does not have to be prepared for distributed computing.
Finding and delivering key information hidden in pharmaco-chemical patents
98% of human knowledge is in the form of digital records. Most of these records are unstructured – hence extracting information is a time and cost consuming task.
UIMA-HPC combines two key technologies that make it possible to get best time-to-solution results: Unstructured Information Management Architecture (UIMA) and Uniform Interface to Computing Resources (UNICORE). The first provides all UIMA-HPC workflows with software-component-based parallelization, the latter facilitating access to high performance computing (HPC) systems and the utilization of hardware-level parallelism. Combined for the first time as part of this project, these technologies synergize for maximum efficiency in solving information extraction tasks on huge unstructured document collections.
Knowledge discovery and semantic search in scientific literature
SCAIView is an advanced search engine and addresses questions of interest to general biomedical researchers. Most of the information is given as unstructured text (publications, text fields in databases).
Download SCAIVIEW [PDF, 712 KB]
ChemoCR® – tool for chemical compound reconstruction
SCAI Bioinformatics has a longtime working experience on the automated extraction of information from text in the biomedical literature. While established techniques exist for text mining, chemical structure extraction is a key feature and presents a significant technological challenge. In contrast to general purpose search engines such as GOOGLE the extraction of the so called connection table from images allows chemical relevant queries. Using GOOGLE image search a user could ask: »Find all documents which are showing aspirin in an image«. The resulting images are identified by searching for the term aspirin in the caption or surrounding text.
ProMiner® – recognition and Norm alization of Named Entities in Scientific Text
Scientific publications found in abstract databases, full text journals or patents are the main and most up-to-date information source, but the amount of text is overwhelming for most Life Science areas.
Recognition of Life Science terminology is a key prerequisite for performing automatic information retrieval and information extraction. Huge and complex terminologies with high numbers of synonymous expressions, ambiguous terminology and numerous generations of new names and classes present named entity recognition with a real challenge.
AutoBarSizer
The software AutoBarSizer generates optimized layouts for the cutting of stock items, namely steel profiles (metal beams) and other bars and rods, into shorter pieces. The generated layouts achieve an extremely high degree of material utilization (yield), i.e. the cutting waste is minimized.
PackAssistant – it’s child’s play!
The geobra Brandstätter Stiftung & Co. KG, manufacturer of the PLAYMOBIL figures, uses about 18 000 different injection molding tools to produce plastic parts for a large variety of »play worlds«. For the products introduced in 2015 this involved 1 000 new parts. The packaging of the parts has to be planned in order to estimate and optimize logistics costs.
PackAssistant – schnelle Behälterplanung für neue KTM-Flitzer
Zum Serienstart des Sportwagens X-Bow stand die KTM-Sportcar Produktions GmbH im Jahr 2008 vor der Herausforderung, in fünf Monaten eine neue Produktionsstätte mit einer Fläche von 5500 m2 zu planen und aufzubauen. Zur Vorbereitung der Fahrzeugproduktion und der Logistikprozesse gehörten auch der Entwurf und die Produktion von Verpackungen für die insgesamt 450 neuen Bauteile des Fahrzeugs.
AutoPanelSizer
AutoPanelSizer gives an answer to the question how rectangular pieces can be placed efficiently on rectangular sheets if the saw can only perform complete rectangular cuts. This question is particularly important in the woodworking industry, but also in the glass, metal, and plastics manufacturing industries.
AutoNester-L
AutoNester-L is a software package for automatic placing of patterns on leather hides. It is widely used in the leather, upholstery, and car manufacturing industry. It is able to nest any set of pieces within a very short time in an optimal way, minimizing wasted material, while taking into account various types of constraints. The effi ciency of the markers achieved by AutoNester-L is competitive to experienced human nesters.
AutoNester-T
AutoNester-T is a software package for automatic placing of patterns on fabrics. It is widely used in the garment and upholstery manufacturing industry. It is able to nest any set of pieces within a very short time in an optimal way, minimizing wasted material, while taking into account various types of constraints. The efficiency of the markers achieved by AutoNester-T is competitive to experienced human nesters.
CuboNester
The CuboNester software optimizes the arrangement of cuboid or nearly cuboid parts in three dimensions. The algorithm can combine a variety of different parts in one arrangement.
The software is available in two variants: CuboNester-C is designed for cutting blocks of material. It can be used for metal, marble, graphite, and foam, among others. CuboNester-P optimizes packing, for example of disassembled furniture into packages.
CuboNester-P
CuboNester-P arranges different cuboid parts in layers in an optimized way. The software is particularly suitable for packaging disassembled furniture into packages.
CoilNest – minimizes waste in pressing of car body parts
Within the CoilNest project, the Fraunhofer Institute SCAI has developed a software for cost and waste minimization in the pressing of car body parts.
In advance of the deep drawing process for car body production, between 50 and 100 blanks are pressed out of various coils. Optimized nesting of the blanks and pressing of the nested patterns allow considerable material savings compared to the conventional pressing of rectangular or parallelogram-shaped blanks. The software solves the resulting optimization problem to decide which blanks are pressed out of the same coil and how they are nested.
Automatic cut order planning in textile industry
CutPlanner is a software package for use in the textile manufacturing industry for automatic cut order planning. CutPlanner takes a customer's order for a clothing item and creates a cut plan for that item, including different sizes and different fabric types or colors, which minimizes production costs.
A cut plan is an assignment of sizes and fabric types to markers. For each of these markers, the required number of plies is computed to fulfil the order's specifics.
Längenoptimierung beim Zuschnitt von Stahlprofilen
Bei der Produktion warmgewalzter Stahlprofile werden zunächst Stahlblöcke unter hohem Energieaufwand auf 1250 °C erhitzt. Daraus werden lange Profile geformt und diese in einem zweistufigen Prozess (vgl. Abb. 3) in kundenspezifische, kurze Profilabschnitte unterteilt.
Die Kunden – z.B. Hersteller von Nutzfahrzeugen, Landmaschinen oder Gabelstaplern und Bauunternehmen – bestellen ihr individuelles Stahlprofil in unterschiedlichen Längen üblicherweise über mehrere Monate oder Jahre hinweg.
Dabei können die Profile bis zur Größenordnung der kurzfristig erwarteten Abnahmemenge auf Vorrat produziert und gelagert werden.
Coupling SOA-based applications with HPC resources
Fraunhofer SCAI is an established specialist in HPC and parallel applications. Through benchmarking activities in the last 10 years SCAI became an expert in measuring and optimizing parallel applications (commercial and Open-Source) on Linux and Windows HPC clusters. Besides simple runtime measurements and scale-out studies (with varied node, socket and core allocations), specific attention was paid to potential influences of the cluster’s interconnect type (Gigabit Ethernet vs. QDR Infiniband), different MPI versions (Intel MPI, MVAPICH, Open MPI, MS MPI) and MPI parameters or underlying file systems on the application’s overall performance. Impacts of different compilers (Intel, PGI etc.) and their respective flags have been analysed as well.
Interactive environment for parameter analysis and optimization
Technical processes and products typically depend on many parameters, e.g. geometrical, material and process parameters. The user would like to adjust these parameters to obtain the best design of the product or the optimal control of the production process.
Simulation, analysis and optimization of networks for gas, H2, CO2, and electricity
In the future, thousands of kilometers of new grids will be built in Germany to make greater use of electricity from renewable energies. These new “intelligent networks” (smart grids) increase complexity, costs, and vulnerability. Fraunhofer SCAI has developed a software tool to analyze and optimize transport networks for gas, electricity, and water using numerical simulations as early as the planning stage. This makes conversion and expansion more flexible for network operators, saves energy and expenses, and increases the security of the energy supply.
EnCoMOS – Energieoptimierung in komplexen Produktionsprozessen
Der Energie- und Ressourcenverbrauch macht bei vielen Unternehmen - etwa aus den Bereichen Chemie, Metall oder Pharma - über 60 Prozent der Produktionskosten aus. Effiziente Energienutzung ist daher für solche Unternehmen von besonders großer Bedeutung. Zur Optimierung der Energienutzung wurde von der Internationalen Organisation für Normung (ISO) die Norm DIN EN ISO 50001 Energiemanagementsysteme (EnMS) erarbeitet. Die praktische Umsetzung eines EnMS ist jedoch aufgrund der Vielzahl abhängiger Steuerparameter in komplexen Produktionsprozessen eine nennenswerte Herausforderung.
Simulation and optimization of data center energy flows
The reduction of energy consumption and the possible reuse of waste heat become increasingly important issues in the design and operation of large data centers. The increased variability in the energy consumption of HPC data centers is a great challenge for the current center infrastructure, which is mainly driven by the need to save electrical energy. Figure 2 shows the power consumption of the Leibniz Supercomputing Centre of the Bavarian Academy of Sciences and Humanitiesin Garching (LRZ). As can be seen, the overall power consumption varies with the power consumption of the super computer (SuperMUC). The combination of this load variability and the use of new cooling technologies in the data center (chillerless cooling) provides an opportunity to increase the energy efficiency of the data center.
Computer-aided molecular design
The Virtual Material Design (VMD) Group performs contract research and solves problems for the industry with state-of-the-art computer-aided molecular design (CAMD). Our aim is to support the industry in developing products and processes.
Multi-CPU accelerated numerical simulation
Recently Fraunhofer SCAI, in conjunction with the Institute for Numerical Simulation (INS) at the University of Bonn have been selected as one of the first CUDATM Research Centers in Germany based on the vision, quality and impact of their work. The research is focused on the development of massively paralleled multi-GPU based software packages for numerical simulation in the natural and engineering sciences. Companies benefit from the transfer of knowledge from basic research to practical applications.
SimCompare
SimCompare is a tool for detecting events – in this context deviations, such as anomalies or strange variations in deformations – arising in the results of similar simulation runs for crashworthiness analysis using finite element (FE) models. It provides a comparison of two FE simulation results based on arbitrary node / element data functions, e.g. displacements, plastic strains, or thicknesses, and time step. The parts with the largest differences are automatically highlighted based on several comparison measures and an appropriate threshold.
ModelCompare
The ModelCompare plug-in for finite element method (FEM) pre- and postprocessing tools provides a comparison of two similarly discretized FE models and portrays their differences in terms of geometry (mesh), material-ID and thickness.
ModelCompare identifies the differences between the two models based on the geometry described by the mesh and is independent of the Part ID (PID) or the NAME / TITLE of the parts. It uses specialized mapping techniques that result in extremely short run times.
FEMMINER – accelerated analysis for hundreds of simulations
The demand for improved product development in industry has led to a tremendous growth in the number of large-scale finite element simulations being performed. Dimension reduction techniques that extract the essential variations between designs and simulation results provide a new approach for analyzing product variants stored in huge simulation repositories. In this context Fraunhofer SCAI investigates innovative solutions for the automotive industry.
Fraunhofer Institute for Algorithms and Scientific Computing SCAI