09:00 – 09:10

Opening & Welcome

09:15 – 10:45

Workshop: Benchmarking for model verification and validation

Abstract:

Virtual labs, virtual twins, and virtual design are innovations with significant potential to transform industrial practices. The tools used in these virtual representations are often based on partial differential equations (PDEs) and require numerical solutions through complex software implementations. Additionally, there are challenges related to parameter identification and comparison with real experimental data. Therefore, to fully realize this potential, it is essential to rigorously verify the tools and software implementations and validate the methods employed.

The primary goal of this workshop is to identify the challenges faced by the community in this area, assess the current state of the art, and review existing benchmarks across different domains. During the workshop, we will collaboratively identify key questions that a verification and validation benchmark platform should address. Additionally, we will discuss the necessary functionalities for effective user interaction and work towards identifying a community that would benefit from and contribute to such a platform.

Join us for this engaging session as we explore these critical topics and work together to advance the field of simulation model verification and validation.

Publications:

Diercks, P , Gläser, D , Lünsdorf, O , Selzer, M , Flemisch, B & Unger, J. (2023) ‘Evaluation of tools for describing, reproducing and reusing scientific workflows’, ing.grid. 1(1) doi: 10.48694/inggrid.3726

10:45 – 11:00

Break

11:00 – 11:30

Showcase „Betty’s Research Engine“ (NFDI4ING Service)

Abstract:

coming soon

Publication:

Seibert, Vasiliy (2024): Betty’s Re Search Engine Talk. (Screencast) https://doi.org/10.5446/66914

11:30 – 12:15

Software solutions for construction engineering and architecture (CC45)

Abstract:

The presentation will focus two solutions: the NFDI4ING Data Ingest Service and the FAIR data development kit RDF store. The first allows publishing and archiving research data from built environment research. It supports typical open data formats from this field such as IFC (building information models) or e57 (point clouds) and generates 3D previews from them for fast inspection. The second is intended to serve as a modular, expandable software framework for research data management interlinked with NFDI. It can be customised to create domain-specific tools that respond to the heterogeneity of data and methods of modern research in, but not limited to, built environment.

12:15 – 12:45

Participant Contribution: “The I in FAIR: Solutions for HeterOgeneous Workflows in Model-based Engineering applications – SHOWME.how”

Abstract:

Model-based development, design, decision support, and diagnostics in engineering applications often require complex workflows that involve heterogeneous computational models, programming languages, operating systems, and computing infrastructures. The diversity of workflow components presents significant challenges in interoperability, limiting the usability and subsequent reusability of the generated digital assets, including software and data.

To address these challenges, we present an approach to orchestrate various computational workflow components, enhancing their combined value through synergy. Additionally, it facilitates the seamless integration of state-of-the-art research methodologies into application-driven engineering tasks. We call this method SHOWME.how. It is enabled by technologies like package managers (Conda, Mamba), containerization (Apptainer, Docker), data exchange protocols (HTTP, Filesystem), and workflow managers (Nextflow).

Using SHOWME.how, we develop blueprints to perform high-throughput tasks for applications in heat transfer and free-surface flow, that use different software for computational models (OpenFOAM, FEniCS) and packages in various languages (Python, Julia, R). Our approach enables the reuse of existing research software in its native environment, eliminating the need to build wrappers, develop language-bridging interfaces, or rely on suboptimal implementations. It also streamlines the prototyping of complex computational workflows for engineering studies,  facilitating their creation and reuse with minimal effort.

The demonstrated blueprints can be readily adapted to develop highly interoperable and reusable digital assets and software frameworks, ultimately increasing their longevity and value.

Keywords: Computational Workflows, Model-based Engineering, High-Throughput Computing, Uncertainty Quantification

12:45 – 13:00

Wrap up