Session: CIE-14/15-01 CAPPD: Joint Topics

Paper Number: 148327

148327 - The Simulation Informed Computer-Aided Design Framework: Exploring How Simulation Drives Iteration and Decision Making in Cad 

Introduction  

Simulation-informed design has been recognized as an essential approach in product development [1], [2], enhancing design performance and improving decision-making support [3]. Moreover, the synergy between simulation-informed design and robust automated optimization has proven to assist designers in making informed decisions [5], [6], [7]. 

Recent advances have integrated simulation capabilities like Finite Element Analysis into CAD modeling environments, enabling rapid simulation-informed iteration. Despite the advances in CAD analytics [4], [8], there is a gap in the understanding of how simulation drives design iteration. 

RQ: How do engineers use simulation results to inform their iterative design decisions?

In this paper, we propose the Simulation Informed CAD (SICAD) framework for analyzing how designers use structural simulation while iterating on a design, then use it to analyze data collected during a 2-hour design challenge to demonstrate potential applications.

The SICAD Framework

The Simulation Informed CAD (SICAD) framework was developed to analyze the usage of real-time simulations in CAD environments, such as Onshape Simulation [9], during part modifications. 

Its goal is to serve as a novel roadmap for researchers exploring the effect of CAD environments on design decision making. The data input consists of the Backend Analytics Data, a log of all actions users take in CAD, and details about the Model Geometry as it changes. Each modification to part geometry and visualization of simulation results constitutes one iteration, with metrics such as number of actions or part complexity measured for each successive cycle. 

Action Type Classification

The first part of the SICAD Framework was adapted from Gopsil et al. [2016], similar to Deng et al. [2022]. Actions collected from the back-end audit trail are grouped into six categories aimed at understanding how a user iterated on their design with simulation; Simulation Definition, Viewing Simulation, Creating/Editing, Reversing/Deleting, Viewing, and Other.

Iteration Trajectory Analysis

The second part of the framework attempts to quantify how a part's geometry changes and how a user's actions change in successive iterations between viewing simulation results. We measure aspects of the part geometry such as “Number of Edges” to see how part complexity changes over time, as well as the “Number of Actions” the user performed during each iteration cycle.

Demonstration of SICAD Framework

To demonstrate the potential application of the SICAD Framework, it was used to analyze data collected during a 2-hour hackathon with university students. 7 groups were tasked with modifying a starter model to achieve a target stress-strain curve on a force test stand. More details on the design challenge can be found in the Onshape Learning Center [11].

Discussion and Preliminary Results

Action Type Classification and Iteration Trajectory Analysis were performed on the collected data. In both cases, the CAD Documents were ordered by the relative performance of the design for the specified target in the design challenge from best to worst.

Not enough data have been collected for meaningful conclusions into how engineers use simulation results to drive design decisions, but the preliminary analysis demonstrates potentially valuable insights. For example, the user who created the best performing design was the only one with more than 1% of their actions in the “Simulation Definition” category, and the number of actions per iteration cycle tended to decrease overall for all users.

Conclusion and Future Work

We have proposed the Simulation Informed CAD (SICAD) Framework as a roadmap for researchers seeking to understand how designers use analysis to inform their iterative decision making process. Potential applications of the Action Type Classification and Iteration Trajectory Analysis are demonstrated through the analysis of design data. As more data are collected through experimentation and real-world design activities, the SICAD Framework can help researchers improve tools and training to help engineers more effectively use simulation tools in their design processes.

Presenting Author: Matthew Mueller PTC

Presenting Author Biography: Matthew Mueller is the manager of Education Innovation at PTC where he is the product manager for Onshape’s education features and leads academic research collaborations. He completed his PhD in Mechanical Engineering at Tufts University where his research focused on engineering education. His current research focuses on how CAD can be used to understand and improve design.

Authors:

Matthew Mueller PTC
Aliaa Maar Pennsylvania State University
Debrina Roy Pennsylvania State University
Alison Olechowski University of Toronto
Jessica Menold Pennsylvania State University