Data-driven design & fabrication methods
ME-428
Week 1: Introduction and Defining a Design Problem (20/02/23)
For the first week of the course we have an introductory lecture, and also discuss how to define the design problem. Slides are given below- video of the lecture will be posted after the event.
Project work for this week:
- Install Matlab + optimization toolbox - familiarize yourself with this toolbox
- Identify possible topics for design optimization and form groups of 3-4 for the project
Week 2: Encoding a design space (01/03/23)
We will discuss a number of methods for encoding a design space including parameterization and also learning based methods.
Week 3: Design of Experiment (DoE)
We will recap VAEs and then focus on DoE.
Week 4: Genetic Algorithmns
Introduction to genetic algorithmns.
Week 5: Bayesian Optimization (Flipped Classroom)
Recap of heuristics, discussion of Design of Experiments. Next we start looking at learning based approaches for surrogate modelling.
Week 6: Heuristics & Design of Experiment
Week 7: Simulation & Industry Perspective on DDD
No slides in advance this week
Week 8: Method Selection
We will finish off Surrogate Models by looking at Bayesian Optimization, both the theory and the practical implementation
Week 8: First Presentations
See schedule above
Week 9: Simulation
Introduction to surrogate modelling and Bayesian Optimization.
Week 10: Implementation of BO and Simulation
Week 11: Future Outlooks and Review
Week 12: Fabrication & Future Outlook
Discussion of a variety of additative and subtractive manufacturing processes, and the pipeline from design to part.
Week 13: Challenges/Opportunities for Data Driven Design
Finish discussion on 3D printing and then conclude with a review on the challenges and limitations of data-driven design.
Week 14: Final Presentations
Content:
13:15 – 14:00 | Introduction
|
14:15 – 15:00 | Workflow Example: Design Optimization of Mechanical Systems
|