Use of Digital Twin to Analyse the Effect of Graphene as a Lubricant Additive for Diesel Engines
Two engine oil additives with graphene were tested in Diesel engines. The first one was a graphene oxide (GO) based, commercially available with the supplier recommendation of using 3% V/V concentration in engine oil. The second one was a graphene nanoplatelets (GNP) based, under development, more concentrated allowing to add a much smaller amount of additive volume. The use of GO additive resulted in a reduction of brake specific fuel consumption by 0.2%-0.7%. The use of 0.1% V/V of GNP brought 0.4% of fuel savings on an ESC emission cycle. Further increasing the GNP concentration to 0.2% did not further reduce fuel consumption. Tests with up 0.2% V/V of GNP did not impact the particulate emissions. Because the expected engine performance benefits from using improved oils were small, almost the same as the measurement uncertainties, the applicability of machine learning using engine on-board diagnostics (OBD) readings to analyse the impact of lubricant additives was tested. The use of Random Forest, machine learning digital twins, was able to reproduce the actual values with excellent accuracy.
Tematyka artykułu: Paliwa i smary, trybologia silników
Autor: Eduardo Tomanik
Współautor(zy): Grzegorz Koszałka, Thiago Marinho Maria, Jacek Hunicz and Wania Christinelli