ISBN 9783843934046

978-3-8439-3404-6, Reihe Ingenieurwissenschaften

Emmeram Meindl
Numerical and Experimental Investigation of Knock in Turbocharged Direct Injection Spark Ignition Engines

195 Seiten, Dissertation Technische Universität Bergakademie Freiberg (Sachsen) (2017), Softcover, A5

Zusammenfassung / Abstract

Due to the ongoing trend of engine downsizing and increasingly tighter emission standards knock is still a crucial aspect of engine and combustion process development. The scope of this thesis was to evaluate the impact of varied engine operating parameters on the global and local knock behavior of turbocharged direct injection spark ignition engines and to verify, develop and validate proper methods, procedures and numerical models for the analysis and prediction of knock. In addition, this thesis deals with the occurrence of cyclic knock variability and possible contributions to this phenomenon. Both experimental and numerical investigations were performed. The impact of the air-fuel equivalence ratio λ, intake-manifold pressure, intake-manifold temperature and engine speed was investigated by a comprehensive testing program on the engine test bench. The mean knock trends and the individual cycle behavior were evaluated. The influence of the thermochemical state of the end-gas on auto-ignition, the prerequisite of knock, was analyzed with the help of a zero-dimensional chemistry solver. A zero-dimensional knock modeling approach incorporating complex chemical kinetics was implemented and validated against experimental data. This tool was applied in order to assess the impact of NO on auto-ignition under engine operating conditions. Furthermore, this modeling approach was used to investigate how and to what extent cyclic variations of the trapped air mass, residual gas mass fraction and composition contribute to cyclic knock variations. The influence of the varied engine parameters on the preferred locations of auto-ignition was evaluated by 3D CFD simulations. The generalized knock integral method (gKIM) was chosen for the 3D CFD knock modeling. The gKIM was modified and its capability to predict the mean crank angle of knock onset was demonstrated. This modeling approach was also used to assess the impact of fuel spray fluctuations on the local mixture and temperature distribution and ultimately on the cyclic variations of knock.