19009 DEFAINE (Design Exploration Framework based on AI for froNt-loaded Engineering)
| Reference number | |
| Coordinator | Chalmers Tekniska Högskola AB - Industri och materialvetenskap |
| Funding from Vinnova | SEK 8 576 634 |
| Project duration | September 2020 - February 2024 |
| Status | Ongoing |
| Venture | Eureka cluster co-funding |
Purpose and goal
The project has, with the help of AI technology such as machine learning has developed methods and tools in several disciplines to be able to effectively explore and optimize a larger number of construction alternatives, a design space. Being able to prepare (Front-loading) project for this type of activities aims to: · Reduce recurring costs when designing aircraft by 10% and · Reduce design update lead time by 50%. Methods and tools have been demonstrated in various use cases. The project has been presented in various events and publications.
Expected results and effects
The targeted KPIs were met, demonstrated in the project´s use cases: - Number of design objectives traded (target:20 achieved:24) - Design space dimensionality (target 50% increase, achieved = 400%) - Lead time for design update (target 50% decrease, achieved: 91%) - Number of design variants evaluated(target: 10x, achieved:40x-steady state, 15x -transient model) - Lead-time for design update (target: 8-40hrs, achieved: 0.5-1hrs ) - Simulation workflow set up (target: 1 week, achieved: 5 hours) - Conference publications (target: 8, achieved: 12) - Events (achieved: 7)
Planned approach and implementation
The work has been carried out with Swedish partners and partners from the Netherlands through workshops (physical and virtual). The project suffered from a slow start due to covid and the inability to have a proper physical kick-off of the project. After the first real physical workshop, work could really begin and virtual workshop was possible in an effective way. Another contingency was due to the fact that key partners from two countries (Germany and Belgium) were not funded within their own national funding applications, leading to a significant and time-consuming re-scoping.