The Mechanical Web

Smart Digital Supply Chains in Manufacturing Industry

Project description

Problem: Although it is recognised that the modelling and simulation significantly speeds-up and improves the design process, the integration of supplier models in the costumer’s chain is rarely done. The reason for this is because the exchange and integration of virtual analysis models between component supplier and system integrator is a complex and mostly manual task with many limitations in the functionality. Furthermore, detailed information of the supplier’s products, captured in virtual analysis models, should not be disclosed freely, and most engineering design problems require experiments as well as simulations to evaluate the design objective and constraint functions as a function of design variables. Today, the optimization of the integrators systems by model predictions becomes therefor nearly impossible or too costly. The information exchange between internal/external component suppliers and the integrating OEM/design department mostly relies on detailed but static documents to describe technical requirements and potential solutions. Initial proposals rarely allow the system designer to integrate supplier models directly into their evolving virtual engineering. Business Expectation: In small series production a late integration of supplier’s sub-models into the integrator’s virtual design environment may cause time delays if both models don’t harmonize as expected. An early adaption of supplier and system models would help to find a better technical design in shorter times. A smarter and more automated interaction between user search for potential solutions as well as detailed technical request and supplier response interaction are the right way to make such engineering and business processes more efficient. Suppliers of new technologies can provide interactive and functional models, which would demonstrate their capabilities in an early design stage. Technological Innovation: The main goal of the MechanicalWeb is to provide solutions for the efficient exchange of executable and smart surrogate models between component suppliers and the integrating system designer. Platforms may be web-based or run on open, community, or internally shared digital twins. The focus of the MechanicalWeb project is on AI and ML based methods to create, evaluate, and integrate supplier models into OEM’s complex system design environments. By introducing such smart surrogate models both – supplier and end-user – will benefit from faster integration times and better feedback on part behaviour in the target system already at early phases of design. Four industrial Use-Cases will highlight different technical aspects and business challenges: • ‘household equipment’ starts at the early phase of a product design, where OEM searches for suitable suppliers, runs cost estimations and checks technical compliance (Tridi, Telligence, Arzum - Turkey) • ‘online manufacturing’ intends to create an supplier platform, where users can find and request manufacturing solutions (AKKA, RapidMiner, Simutence, Oerlikon, KIT, Fraunhofer - Germany) • ‘shared digital twins for shaver design’ will be a kind of community environment, where partners share their models and data (Reden, Philips, RUG, InSumma, ThunderbyteAI, Unit040 - Netherlands) • ‘Virtual printer’ is defined as a company internal supply chain, where the design department needs to integrate parts from other departments for the final printer system definition (Canon Production Printing, Reden – Netherlands; AKKA, RapidMiner - Germany)


Project leader

Alt
Klaus Wolf
Fraunhofer SCAI, Germany
Project involvement The Mechanical Web, VMAP
Alt Alt Alt

Project publications