Optimal design of composite structures

A composite material is a macro-level combination of two or more material whose properties can be tuned based on the macro-scale distribution of the material. D.I. Self-Composite Alloys Inc., are working on developing a new generation of materials. Their preliminary findings have shown that it is possible to create metals with improved mechanical properties by just tuning the manufacturing process. They are interested in a design optimization tool for composites. Since traditional design optimization would an iterative time-intensive process, the project will aim to “teach” machine learning algorithm optimal solutions for different designs. Specifically, design optimization will be performed for a sample application over a constraint space, and an artificial neural will be trained on this data. This train network can then predict the optimal design without the need to perform a full optimization.

Faculty Supervisor:

Krishna Vijayaraghavan

Student:

Wesley Romey

Partner:

DI Self-Composite Alloys Inc.

Discipline:

Engineering - mechanical

Sector:

Manufacturing

University:

Simon Fraser University