Work Group Knowledge-Based Systems Engineering

In research, we contribute to the dissemination and application of knowledge-based systems and artificial intelligence methods for the development of customer-centred, adaptive solutions. For developers, we offer methods and tools to model solution spaces for multi-variant product and product-service systems with time-varying and fuzzy requirements and constraints, to support decisions during solution space exploration and to automate routine tasks during the design process.

We pursue two approaches: In the solution-centric approach, solution spaces are spanned by the explicit mapping of formalised knowledge, for which we further develop product configurators and design assistants as well as their implementation methods, among other things. In the problem-centred approach, the design process is formalised by algorithms so that solution spaces are mapped implicitly without modelling specific product variants. We implement tools, e.g. based on computational design synthesis, generative design and numerical optimisation, in order to build up knowledge about development objects and design them robustly.

In turn, our research flows into teaching in order to train students in the modelling of constructive solution spaces as well as in the development of adaptable systems.
Our aim is to use practical content to teach problem-solving skills that enable students to independently develop methods and tools and apply them to problems. They themselves benefit from the fact that we also use knowledge-based systems in teaching for individualised, skills-oriented training.

Projects

  • Design of hybrid solid components (SFB 1153) - The starting point of the project is the question of how tailored forming components must be designed so that the potential of hybrid solid components can be exploited. Based on this, guidelines are provided for the design of hybrid solid components, e.g. for the material combination of steel and aluminium.
  • User interface for working with generative models - The research deals with the modelling of an application of the "Generative Design Approach" using knowledge-based development methods. This creates a development environment that generates different variants from a large solution space during the development process.
  • Development environment for systems for hybrid value creation (Smart Hybrid) -In this project, development methods and tools for PSS are created and transferred into a generally valid approach with an associated integration and transfer model.
  • Computer environment for designing with design catalogues - The use of computer-aided design catalogues as a knowledge base for KBES is being investigated
  • Engineering tools for company-typological variant and complexity management
  • Multi-agent systems as decision support in design: In this project, we are virtualising the design review and automating the assessment of individual part designs with regard to their manufacturability and functional reliability through the use of virtual agents. These represent autonomous, decentralised images of experts from different domains who can jointly find an optimal component design through perception, communication and negotiation.
  • Computational Design Synthesis for the development of bone-anchored implants: The research aims to automate the design of implants, e.g. of the hip. We use algorithmic methods to build a model of the design process itself, which is used to generate a customised implant for each application. CT scans form the basis for this and the implants are then additively manufactured. Surface structuring using selectively applied grid structures supports bone ingrowth.
  • Learning support through digital teaching programmes in design engineering: In these projects, we are researching the use of knowledge-based systems for training design engineers and integrating intelligent tutoring systems. These automatically provide our students with individualised feedback for uploaded design tasks and thus supplement our classroom teaching.
  • Tailored Seat: Project in cooperation with our partner Forvia.

   

Expertise

  • CAD: Autodesk Inventor, SolidWorks, Grasshopper
  • FEM: Autodesk Simulation Mechanical, Ansys, Abaqus
  • Experimental investigations
  • PSS product service systems
  • Knowledge-based engineering

Equipment

  • CAE Pool
  • Test benches
  • Powerwall

 

Contact

Dr.-Ing. Paul Christoph Gembarski
Management
Address
An der Universität 1
30823 Garbsen
Building
Room
303
Dr.-Ing. Paul Christoph Gembarski
Management
Address
An der Universität 1
30823 Garbsen
Building
Room
303