Knock Prediction Improvement with Application to the Analysis of Knock Mitigation with Targeted Water Injection in SI Engines

Knock is a crucial limiting factor for improving the thermal efficiency of spark-ignition gasoline engines. However, effective suppression of knock in an engine and accurate prediction of knock occurrence are still present research challenges. This study proposed an improved Livengood-Wu (L-W) knock integral model, which employs the local thermodynamic parameters, rather than the global ones, in the three-dimensional computational domain and a multi-region knock precursor method based on the L-W integral transport equation. The improved knock model was applied to study the knock suppression effect of targeted water injection, which directs water sprays to knock-prone areas in the end gas region. It was found that although the accuracy of the improved model in predicting the crank angle of knock onset under normal conditions without targeted water injection is similar to the empirical formula-based L-W models, the improved L-W model can catch the spatial details of knock and the influence of the local thermodynamic changes, being effective in the cases with great inhomogeneity of temperature and components. The numerical results indicated that the knock suppression effect of targeted water injection was significant. The cooling effect of the water-to-fuel ratio (w/f) from 5% to 10% is more significant than that from 10% to 20%. More water injected does not bring more significant cooling benefits to the target region, and w/f of 10% shows a better knock suppression effect. In the studied engine, targeted water injection with w/f of 10% is predicted to advance the ignition timing by 3°crank angles. In addition, the knock suppression effects of targeted water injection and conventional direct water injection without spray injection preference are also compared, which reveals a better effect of the targeted injection with less water. However, the knock tendency suppression is mainly effective in the watering-specific region (e.g. the exhaust side) and has little impact on the non-watering regions where knock may occur. This indicates that the knock suppression effect is limited by the knock tendency in the region without water injection, and further optimization of the design and control strategy of the targeted water injection is needed.