Reduced dimensional model for inexact quasi-Newton acceleration and stabilization of a computational multi-physics framework

The project focuses on enhancing a computational multi-physics framework for analyzing or designing engineering systems where more than one physical phenomena are present and interact with each other – for example, fluid-structure interaction or conjugate heat transfer. In cases where the solution algorithm can be unstable or slow to converge, the partner organization has found it beneficial to use an inexact quasi-Newton method to stabilize the problem and speed up convergence. However, this method can become prohibitively expensive in some cases. The project will be to develop and provide an approximate reduced dimensional model for the quasi-Newton algorithm which reduces the cost without sacrificing too much accuracy.

Faculty Supervisor:

Rajeev Jaiman

Student:

Partner:

ANSYS Canada Ltd.

Discipline:

Engineering

Sector:

Information and cultural industries; Professional, scientific and technical services

University:

The University of British Columbia

Program: