This Website is not fully compatible with Internet Explorer.
For a more complete and secure browsing experience please consider using Microsoft Edge, Firefox, or Chrome

Engineering Optimisation with Bayesian Optimisation and CFD

This presentation was made at the NAFEMS European Conference on Simulation-Based Optimisation held on the 15th of October in London.

Optimisation has become a key ingredient in many engineering disciplines and has experienced rapid growth in recent years due to innovations in optimisation algorithms and techniques, coupled with developments in computer hardware and software capabilities. The growing popularity of optimisation in engineering applications is driven by ever-increasing competition pressure, where optimised products and processes can offer improved performance and cost-effectiveness which would not be possible using traditional design approaches. However, there are still many hurdles to be overcome before optimisation is used routinely for engineering applications.

The NAFEMS European Conference on Simulation-Based Optimisation brings together practitioners and academics from all relevant disciplines to share their knowledge and experience, and discuss problems and challenges, in order to facilitate further improvements in optimisation techniques.

Resource Abstract

Computational Fluid Dynamics (CFD) is an important component of engineering simulation, covering the solution of the fundamental equations of fluid dynamics, together with supplementary models for turbulence and a range of other physical processes such as heat transfer, combustion, chemical reaction and multiphase flow. It is well known to be computationally very costly, which until recently has restricted its design use. Improvements in computing power now make it possible to perform the multiple runs necessary to optimise real engineering systems, but automated optimisation algorithms still have to be carefully designed to minimise the number of individual simulations performed. Bayesian Optimisation is a technique from Machine Learning in which the emphasis is on the process of learning an objective function which is an approximation to the cost function for the design; once an initial objective function has been learnt the process of finding the optimum solution(s) is relatively straightforward, and the objective function may be incrementally updated and refined as search progresses. We have developed a Bayesian Optimisation toolkit in Python, using the open source CFD code OpenFOAM, and demonstrated its utility on a number of engineering and industrial test cases. We will present an overview of the Bayesian Optimisation process and its application for a number of cases including heat exchangers, draft tubes and particle separators.

Document Details

AuthorDaniels. S
Date 15th October 2019
OrganisationUniversity of Exeter


Back to Search Results