Predictive modeling of turbocharger performance and geometry using statistical regressions

Mohamed-amine Elhameur, Mahfoudh Cerdoun, Lyes Tarabet and Giovanni Ferrara

Energy, 2025, vol. 335, issue C

Abstract: This paper examines the potential of developing statistical correlations to predict turbochargers' performance and basic geometry, which enables rapid selection from vendors' catalogues for an adequate matching with internal combustion engines or eventual improvement of turbomachines design. In this aspect, simple and multi-linear regression analyses were achieved to generate predictive models. Dual validation framework (direct and inverse) were proposed and performed to explore the accuracy of the regression models. During the direct validation, aerodynamic performance of 103 turbochargers were forecasted and compared to the baseline data using statistical metrics. In the inverse validation, a suitable turbocharger was selected from a created database and matched numerically to a 1.5L, three-cylinder, diesel engine. Furthermore, the output performance and operating aero-thermodynamic parameters of the newly boosted engine were compared with the baseline ones. Multi-linear regression analysis's results show that the predictors explain 88.11 %, 96.53 %, 79.33 %, and 94.89 % of the variance of the compressor's mass flow, pressure ratio, rotational speed, and isentropic efficiency, respectively, all statistically significant at the 0.05 level. Moreover, the generated predictive models forecasted accurately centrifugal compressor performance with RMSE values up to 21 %, 15 %, and 5 % recorded for pressure ratio, efficiency, and mass flow, respectively. Besides, the predictive model of the compressor's mass flow rate demonstrated the most promising result, with an RMSE and R2 value of about 5 % and 30 %, correspondingly. On the other hand, selecting the appropriate turbocharger from the database, improved the engine's brake power by around 12.64 % at 1000 rpm, compared to the baseline engine. Finally, this study offers valuable correlations for rapid prediction of turbocharger performance or geometry, thereby supporting the turbocharger selection and design process allowing at the end to reduce engine-turbocharger matching effort.

Keywords: Turbocharger selection; ANOVA method; Database; Engine matching; Linear regression (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544225039805
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:335:y:2025:i:c:s0360544225039805

DOI: 10.1016/j.energy.2025.138338

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().