Development of a predictive 0/1D model for lubricating oil induced pre-ignitions at an optical gas/dual-fuel engine

Abstract

Undesired self ignition of premixed gas-air charges in dual-fuel engines caused by burning lube oil droplets continue to endanger both performance and structural integrity of those engines. Since experimental investigations in engines are demanding and associated with enormous costs a novel experimental test facility featuring dual-fuel operation was used to provide reference data forb suitable 0/1D pre ignition model development and validation. The test bench is equipped with an optically accessible combustion chamber, whose adaptable setup allows mounting plug-ins for specific lube oil addition. Moreover, it features adjustable operation at engine relevant conditions (compression pressure/temperature) as well as a wide range of speed, tune able flow (turbulence), variable gas/air charge composition and high flexibility by pneumatic driven valve train. Already established optical acquisition of ignition and flame propagation is applicable to detect inflammation and flame kernel growth of pre ignition phenomena. Comprehensive instrumentation determines in-cylinder pressure, process gas temperature and other boundary conditions needed for 0/1D model validation and development. Phenomenological modelling needs a deep understanding of the underlying processes. Pre ignition denotes a complex interaction between the cycle-to-cycle variations of flow field, turbulence, and levels of in homogeneities, and further between reaction kinetic processes and the random existence of potential ignition kernels. The optical investigations and fundamental thermodynamic measurements performed at the experimental test facility determined pre ignition influencing parameters. Simulation of the pre ignition occurrence including detailed reaction kinetic mechanism is validated against test bench investigations. Initial results of the pre ignition model show good agreement with the optical test bench measurements. The fast 0D/1D model is capable to predict pre ignition events with boundary conditions from the test bench using lube oil droplet evaporation combined with detailed reaction kinetics.