Tampere University of Technology

TUTCRIS Research Portal

Customized dimensional analysis conceptual modelling framework for design optimization—a case study on the cross-flow micro turbine model

Research output: Contribution to journalArticleScientificpeer-review

Details

Original languageEnglish
Pages (from-to)1168-1184
JournalEngineering Optimization
Volume51
Issue number7
Early online date2018
DOIs
Publication statusPublished - 2019
Publication typeA1 Journal article-refereed

Abstract

Dimensional Analysis Conceptual Modelling (DACM) is a framework used for conceptual modelling and simulation in system and product designs. The framework is based on cause–effect analysis between variables and functions in a problem. This article presents an approach that mobilizes concepts from the DACM framework to assist solve high-dimensional expensive optimization problems with lower computational costs. The latter fundamentally utilizes theories and concepts from well-practised dimensional analysis, functional modelling and bond graphing. Statistical design-of-experiments theory is also utilized in the framework to measure impact levels of variables towards the objective. Simplifying as well as decomposing followed by optimization of expensive problems are the focuses of the article. To illustrate the approach, a case study on the performance optimization of a cross-flow micro hydro turbine is presented. The customized DACM framework assisted optimization approach converges faster and returns better results than the one without. A single-step simplification approach is employed in the case study and it returns a better average optimization result with about only one fifth of the function evaluations compared to optimization using the original model.

Keywords

  • Causality analysis, conceptual modelling, DACM assisted optimization, high-dimensional optimization, micro cross-flow turbine

Publication forum classification

Field of science, Statistics Finland