LESSONS LEARNED FROM DEVELOPING A DIGITAL PROTOTYPE WITHIN THE ARENA 2036 ENVIRONMENT AND IMPROVEMENTS WITH THE NEW VMAP STANDARD

In 2013, a new research campus called ARENA2036, which is the abbreviation for Active Research Environment for the Next Generation of Automobiles, has been established at the University of Stuttgart, Germany. The year 2036 accounts for the 150th anniversary of the patented version of Carl Benz’s first three-wheeled vehicle with a combustion engine which led to the first serial production of motorized vehicles as we know them today. The main focus of the research campus is to investigate the rising importance of Industry 4.0 applications and in a first phase, three leading technical projects worked on topics such as lightweight design and system/sensor integration (LeiFu), new and adaptive production lines in industrial environments (ForschFab), and enhancing simulation capabilities with the definition of a Digital Prototype (DigitPro) [Dittmann et. al. 2015]. The latter aimed to improve the daily life of simulation engineers by developing a digital prototype which defines a simulation data exchange platform based on a HDF5 data storage container and by improving mapping schemes to structural analysis meshes for selected process simulations such as braiding, resin infiltration, open-reed weaving (ORW), and draping of continuous fiber reinforced plastic (CFRP) materials. This should help to close the simulation process chain within the project and is now used as a basis for a subsequent 5-years governmental funded project called Digital Fingerprint [Böhler et. al. 2018]. To develop accurate and efficient mapping and data exchange interfaces, physical relevant parameters for the different processes have to be identified in a first step. In a second step, a proper data exchange format has to be defined, since the different partners use various software solutions which output results in different formats and therefore cannot be interpreted by other software tools. In this work, it will be shown why the HDF5 ® format seems to be a valid storage platform for a large amount of data which accumulates when following the full simulation process chain for all different kinds of materials, components, and processes. Furthermore, a first step has been made to define a common standard for simulation result data exchange and it will be discussed, why the ITEA-VMAP project [Wolf et. al 2018] seems to consist of the better consortium to target topics such as data semantics for resulting analysis data and their interpretation depending on the physical meaning in material modeling and what steps have already been taken to define a common data exchange format. In addition it will be discussed, how the resulting VMAP standard might influence the subsequent project Digital Fingerprint which aims for the inclusion of all relevant data being collected along a components life-time from first concepts and computer aided design via the engineering process and its virtual process simulation chain until its final usage in real applications and probable failure. Finally, an overview regarding the implementation status of the VMAP standard into the used mapping software envyo ® shall be given, showing the capabilities to cover the use cases defined within the VMAP project. These examples will include a use case coupling thermal and structural analysis and a use case showing the simulation process chain for continuous fiber reinforced plastics. Furthermore, enhancements made for mapping processes within the ARENA2036 project will be discussed. 1. Definition of the Digital Prototype within ARENA2036 DigitPro The sub-project Digital Prototype, or short DigitPro of the ARENA2036 research environment consisted of one OEM, three research partners, namely the Institute of Aircraft Design (IFB), University of Stuttgart, GER, the Deutsche Institute für Textil und Faserforschung (DITF), Denkendorf, GER, the German Aerospace Center (DLR), Stuttgart, GER and one smalland medium sized enterprise (SME) – the DYNAmore GmbH, also based in Stuttgart, GER. Given the natural interest of each of the above mentioned institutions, they were looking onto the aspects of numerical simulation of continuous fiber reinforce plastics (CFRP) with different targets and eventually on different length scales. The manufacturing processes chosen within this research project were the braiding process, the resin transfer molding (RTM) process, the open-reed weaving (ORW) process, and a preliminary or subsequent draping process of a preform of any of the above mentioned preprocessing steps. Besides investigating and improving the simulation processes with different simulation tools, one of the targets was also to allow the consideration of simulation results gained along the simulation process chain in the following analysis step. For evaluation purposes, a small component was chosen for each of the starting processes (braiding and ORW) and by the end of the five years project, two main targets should be fulfilled: Lessons Learned from Developing a Digital Prototype within the ARENA2036 Environment and Improvements with the new VMAP Standard the overall development time of a car floor should be reduced by 50 % and the weight saving factor should be around 10 %. The latter also explains why CFRP materials and components were chosen as demonstrator parts. Simulations were performed with different length scales and discretizations, taking into account the need for accuracy of the results in each of the simulation steps along the simulation process chain. As shown in fig. 1, different simulation disciplines are used along the process chain of a CFRP component – in this example a braided structure. To the upper right, material data is being generated for unknown materials, e.g. a biaxial or triaxial braid. Therefore, representative volume elements (RVEs) are used to calculate stressstrain responses into different loading directions and to get an estimation of the failure properties – maximum stresses and maximum strain. These simulations are performed on a microor mesoscopic level, depending on the resolution of the single fibers or fiber bundles. Process simulations such as braiding, draping or even the infiltration process are usually performed on a mesoor macroscopic scale. Thereby, smaller element sizes can be chosen to model a draping process compared to the standard element sizes being used in a structural analysis which usually uses element sizes of 3 – 5 mm. To increase accuracy, research institutions tend to use even beam or shell elements to model the fiber bundles in the braiding, weaving, or draping processes. Resin transfer molding (RTM) analysis can also be performed on different length scales. Depending on that, rather complex solution methods such as computational fluid dynamics (CFD) are used to calculate the locally varying Figure 1: Illustration of the various process simulations and length scales along the simulation process chain [Böhler et. al. 2018]. porosities and to consider them on a macroscopic RTM analysis [Dittmann et al. 2017]. On the other side, crash or structural analysis tend to use rather coarse meshes due to restrictions on the simulation time, e.g. a full car crash analysis should run over night – so no longer than 14 to 16 hours. In a following step, an optimization step should be possible to improve the component design, also considering production process specific needs. Finally, a link to a machine, in this case a weaving or braiding machine should be possible, allowing the translation of simulation input data into machine code such that a full computer aided manufacturing (CAM) process is possible. As shown in the center of fig. 1, all the data being generated shall be stored in a readable, easy accessible data storage container, allowing to track the development status of a component from a first material characterization until the translation into machine code for an optimized component manufacturing process. Various challenges regarding that data storage were faced: the Table 1: Overview of the various software tools being used throughout the project for the different processing steps as well as resulting data.