Use of tribological and AI models on vehicle emission tests to predict fuel savings through lower oil viscosity

2021-36-0038

02/04/2022

Features
Event
SAE BRASIL 2021 Web Forum
Authors Abstract
Content
On urban and emission homologation cycles, engines operate predominantly at low speeds and part loads where engine friction losses represent around 10% of the consumed fuel energy but would account for 25% of the fuel consumption once combustion efficiency is taken into account. Under such mild conditions, engine and engine oil temperatures are also lower than ideal. The influence of oil viscosity on friction losses are significant. By reducing lubricant viscosity, engine friction, fuel consumption and emissions are reduced. Tribological and machine learning models were investigated to predict the effect of oil viscosity on fuel consumption during the FTP75 emission cycle with the use of detailed actual emission test measurements. Oil viscosity was calculated with the measured oil temperature. As the same vehicle transient is followed in the cold and hot phases, the models were evaluated by comparing their prediction of fuel consumption in the hot phase versus the measured value. The models were able to predict the fuel consumption in the hot phase within 1% tolerance for more recent vehicles equipped with GDI and more detailed test data. The proposed methodologies have the advantage of being able to have their reliability tested before application. One can test the model ability to predict the fuel consumption hot phase before applying them to predict fuel saving with lower viscosity oils than the reference one used in the actual emission test. The models were then used to predict fuel saving in the cold phase by change of the tested 5W30 oil for a 0W20 one.
Meta TagsDetails
DOI
https://doi.org/10.4271/2021-36-0038
Pages
12
Citation
Tomanik, E., Tomanik, V., and Morais, P., "Use of tribological and AI models on vehicle emission tests to predict fuel savings through lower oil viscosity," SAE Technical Paper 2021-36-0038, 2022, https://doi.org/10.4271/2021-36-0038.
Additional Details
Publisher
Published
Feb 4, 2022
Product Code
2021-36-0038
Content Type
Technical Paper
Language
English