Book Chapter

Uncertainty Analysis and Model Reduction Based Global Optimisation of Distributed Large-scale Systems

Tao, Min, Theodoropoulos, Constantinos

2020

Overview

Uncertainty arises in many large-scale distributed industrial systems, needing efficient computational tools. Uncertainty propagation techniques have been developed and applied including power series expansions (PSE) and polynomial chaos expansions (PCE). However, such fast low-order approximate models generate errors and, in general, require prior knowledge about uncertainty distribution. In this work, the recursive projection method (RPM) was adopted to accelerate the computation of steady state solutions of complex large-scale dynamic systems. These accelerated models including uncertainty were subsequently utilised in an efficient Bayesian global optimisation framework. The performance of the proposed robust optimisation framework was demonstrated through an illustrative example: a tubular reactor where an exothermic reaction takes place.

• Subject: Bayesian global optimisation / Distributed parameter systems / Recursive projection method / Uncertainty analysis
• ISBN: 9780128233771, 012823377X