Search Results

A Machine Learning-Genetic Algorithm (ML-GA) approach was developed to virtually discover optimum designs using training data generated from multi-dimensional simulations. Machine learning (ML) presents a pathway to transform complex physical processes that occur in a combustion engine into compact informational processes. In the present work, a total of over 2000 sector-mesh computational fluid dynamics (CFD) simulations of a heavy-duty engine were performed. These were run concurrently on a supercomputer to reduce overall turnaround time. The engine being optimized was run on a low-octane (RON70) gasoline fuel under partially premixed compression ignition (PPCI) mode. A total of nine input parameters were varied, and the CFD simulation cases were generated by randomly sampling points from this nine-dimensional input space. These input parameters included fuel injection strategy, injector design, and various in-cylinder flow and thermodynamic conditions at intake valve closure (IVC).

The 4th Workshop of the Engine Combustion Network (ECN) was held September 5-6, 2015 in Kyoto, Japan. This manuscript presents a summary of the progress in experiments and modeling among ECN contributors leading to a better understanding of soot formation under the ECN “Spray A” configuration and some parametric variants. Relevant published and unpublished work from prior ECN workshops is reviewed. Experiments measuring soot particle size and morphology, soot volume fraction (fv), and transient soot mass have been conducted at various international institutions providing target data for improvements to computational models. Multiple modeling contributions using both the Reynolds Averaged Navier-Stokes (RANS) Equations approach and the Large-Eddy Simulation (LES) approach have been submitted. Among these, various chemical mechanisms, soot models, and turbulence-chemistry interaction (TCI) methodologies have been considered.

Numerical simulations of soot formation were performed for n-dodecane spray using the transported probability density function (TPDF) method. Liquid n-dodecane was injected with 1500 bar fuel pressure into a constant-volume vessel with an ambient temperature, oxygen volume fraction and density of 900 K, 15% and 22.8 kg/m3, respectively. The interaction by exchange with the mean (IEM) model was employed to close the micro-mixing term. The unsteady Reynolds-averaged Navier-Stokes (RANS) equations coupled with the realizable k-ε turbulence model were used to provide turbulence information to the TPDF solver. A 53-species reduced n-dodecane chemical mechanism was employed to evaluate the reaction rates. Soot formation was modelled with an acetylene-based two-equation model which accounts for simultaneous soot particle inception, surface growth, coagulation and oxidation by O2 and OH.