Design engineers seek to use engineering simulations to produce designs with higher quality at lower cost and in less time than those who use physical prototypes. However, competent analysts with experience in the simulation domain are needed to perform trustworthy engineering simulations. The idea of democratization of engineering simulation (DoES) was introduced at NWC 2015 to make it easier for design engineers to use the potential of engineering simulation, and an ASSESS theme positioning paper on DoES was released by ASSESS Initiative in 2018. The idea of DoES is comprehensive, but it doesn't seem to include a systematic way for design engineers who aren't familiar with sophisticated design methodologies to combine them with engineering simulations. This study suggests a designer-oriented AI-aided design optimization (AADO) process as a way of democratization of design optimization (DoDO) beyond DoES. There are three sequential stages in the process. In the initial stage for autonomous metamodeling, labeled data received from conducting engineering simulations is used to autonomously and automatically generate metamodels (surrogates; reduced order models; machine learning models) for responses of interest. We developed an AI for picking the best metamodel adaptable to the labeled data received and a rule-based sequential sampling strategy to build credible metamodels in the smallest number of engineering simulations. In the second stage for optimization, an appropriate optimization method that takes into account design problem formulation is autonomously selected, and optimization is then automatically carried out. In the third stage for data storytelling, a designer-oriented explainable design outcome report is automatically generated utilizing data collected during and as a consequence of design optimization. The data storytelling stage consists of five steps for producing the three components of data storytelling: data, narrative, and visuals. The AADO process autonomously and automatically provides the designer-oriented explainable design result report in Excel once parametric engineering simulation models have been supplied. Case studies of industrial design problems will be presented to show the efficacy of the suggested AADO process.
Reference | NWC23-0174-extendedabstract |
---|---|
Authors | Choi. D-H Choi. B-L Lee. Y Ryu. D Moon. H |
Language | English |
Audiences | Designer Analyst |
Type | Extended Abstract |
Date | 17th May 2023 |
Organisation | Pidotech Inc. |
Region | Global |
Stay up to date with our technology updates, events, special offers, news, publications and training
If you want to find out more about NAFEMS and how membership can benefit your organisation, please click below.
Joining NAFEMS© NAFEMS Ltd 2025
Developed By Duo Web Design