Prediction of oil dilution formation rate due to post injections in diesel engines by using Gaussian process

Post-injection strategies are commonly used in diesel engines for various purposes, such as cleaning diesel particulate filters from soot, heat-up and heat maintenance strategies. Usage of post injections in diesel engines with upcoming legislation is expected to be increased because of increasing regeneration frequencies, de-sulfation of catalysts and challenges in cold emission. Post injections with retarded timing and high quantity under certain charge conditions can cause oil dilution by creating liquid fuel impingement on lubricated cylinder walls and the fuel absorption into the oil film. Fuel contamination to engine oil can cause deterioration in oil properties. Engine components can only bear a certain level of degradation in oil properties. Thus, exceeding a certain level of oil dilution should be avoided not to harm engine durability. The engine oil must be replaced within a specific interval due to various mechanisms affecting engine oil lubrication performance such as ageing, soot content etc. Depending on the fuel rate going into the engine oil, oil dilution due to post injections can limit engine oil change intervals. Engine oil change interval is essential for the cost of ownership and environment. Parameters such as charge conditions at post-injection timing and post-injection configuration are effective over the formation rate of oil dilution due to post injections. Therefore, the formation rate of oil dilution due to post injections must be minimized and modelled not to exceed a specific limit in different cycles. With this study, a series of engine testing is conducted to investigate the correlation between empirical spray parameters and fuel in oil formation rate due to post injections. During the studies, various charge conditions and post-injection configurations were tested. In addition to fuel addition to oil, fuel evaporation from oil was also investigated. A correlation between penetration depth and the formation rate of oil dilution was observed. However, a correlation between Sauter means diameter and formation rate of oil dilution was only kept when penetration depth was filtered within a specific range in test results. Evaporation of fuel in oil has shown an asymptotic behaviour based on test duration and initial heavy fraction ratio under constant oil temperature conditions. A statistical modelling approach is applied to predict the formation rate of oil dilution as a function of normalized spray penetration depth, Sauter means diameter and engine speed. Moreover, such a model was validated over various transient cycles.