An Integrated Framework for Bayesian Uncertainty Quantification and Probabilistic Multi-Criteria Decision Making in Aero-Engine Preliminary Design

Abstract

The following paper presents a novel framework that enables making early design decisions based on probabilistic information obtained from fast, deterministic, low-fidelity tools, calibrated against high-fidelity data that is supported by experts’ knowledge. The proposed framework integrates a Probabilistic Multi-Criteria Decision Making technique with Bayesian Uncertainty Quantification concepts supported by the Kennedy and O’Hagan Framework. It allows continuous improvement of low-fidelity design tools as high-fidelity data is gathered and therefore facilitates investigation into the impacts the accumulation of high-fidelity data has on preliminary design process risk. The paper discusses theoretical concepts behind the framework and demonstrates its relevance by application in an illustrative combustor preliminary design case study.