Predicting Engine Oil Behavior Under Extreme Pressure

The Industrial Fluid Properties Simulation Challenge is an open competition designed to push improvements in the field of molecular modeling and ensure simulations are relevant to real world industrial applications. The organizing committee of the challenge want to bridge the gap often found between theory and experimental applications. The Industrial Fluid Properties Challenge asks researchers to theoretically predict a specific property for a fluid that is important in industry applications.

The 10th challenge, which wrapped up in November 2018, asked researchers to predict the shear viscosity of 2,2,4-trimethylhexane. Viscosity is a measure of how easily a liquid flows. Specifically, it looked at how the lubricant behaved as pressure rose from standard conditions to 10,000 times greater than normal conditions. The contestants entering the competition were given the challenge of predicting the viscosity at pressure levels of 0.1, 25, 50, 100, 150, 250, 400, 500, 600, 700, 800, 900, and 1000 MPa. The temperature at all levels was to remain at 20⁰C, or 293K.  In order to determine the winners, the committee ran an experiment, and the results of said experiment were then compared with the submitted simulations. Each entry was judged based on the results compared to the benchmark data determined from the experiment at the described state conditions and in comparison with the pressure viscosity coefficient. The simulation that most closely matches the results of the committee’s experiment is determined to be the winner.

Each entry had to include an analysis of any uncertainty in the results calculated. The results need to be statistically significant, with appropriate supporting evidence to support the results. The contest encourages multiple research groups to collaborate on submissions.

Seven teams entered the 10th challenge including groups from the University of Bath, Imperial College London, U.S. National Institute of Standards and Technology, and Shanghai Jiao Tong University. The winner of the 10th challenge was a team from Johns Hopkins, while the Russian team of Nikolav Kondratyuk and Vasily Pisarev came in second. Both Kondratyuk and Pisarev are affiliated with the MIPT Laboratory of Supercomputing Methods in Condensed Matter Physics. Kondratyuk feels that companies are in favor of using computer models because it speeds up the process of finding possible solutions. He also believes that sponsoring or running contests is a cost effective method for companies and industries to collect usable data to efficiently develop new lubricants.

Understanding how lubricants behave under extreme pressure is vital to various industries. Engineers require this knowledge to effectively prevent damage to expensive machinery due to the failure of lubricants. Understanding how lubricants behave is essential for companies and industries in order to prevent damage to machinery before it can occur, improving the efficiency and lifespan of expensive assets.

However, even supercomputers are limited when it comes to simulating the behavior of lubricants for periods longer than a microsecond. Consequently, the Russian researchers extrapolated the conclusions from the nonequilibrium method, which were then tested using an equilibrium method. The Russian model was accurate from 1 to 5,000 atmospheres, with only a 3% error, placing the Russian researchers in second place in the competition.

The Russian team used molecular dynamics methods in order to predict the viscosity dependence on pressure. The shear viscosity of the lubricant was calculated using Green-Kubo and Muller-Plathe methods. Determining the atomic interactions in the model was done using the COMPASS class II force field.  In order to determine a winner the organizers run an experiment to determine which simulations are the closest to reality.

Further information: Nikolay D. Kondratyuk et al. Calculation of viscosities of branched alkanes from 0.1 to 1000 MPa by molecular dynamics methods using COMPASS force field, Fluid Phase Equilibria (2019). DOI: 10.1016/j.fluid.2019.06.023