Large eddy simulation of a spark-ignition engine using tabulated chemistry approach
Published on 28. May 2017
Book
Paperback/Softback
160 pages
978-3-8439-3131-1 (ISBN)
Description
Due to the limited fuel reserve and climate changing, high thermal efficiency and low pollution emission are the essential requirements for future Internal Combustion (IC) engines, which makes the engine design a challenging task. In this regard, a fundamental understanding of in-cylinder processes such as the Cycle-to-Cycle Variations (CCV) is necessary. The scale-resolving modeling approach, Large Eddy Simulation (LES), is employed in this work since it enables engine designers to gain deeper insight into the CCV.
The CFD code Kiva was used and extended with Smagorinsky and WALE model to perform LES. In the context of LES, the tabulated chemistry approach Flamelet Generated Manifold (FGM) and the Artificially Thickened Flame (ATF) model were implemented into the Kiva code. The model implementation was successfully verified. These turbulence and combustion models were validated by simulating a bluff body stabilized turbulent premixed combustion system under non-reactive and reactive conditions.
Then, the LES of the non-reactive flow in the motored SI engine was conducted for 50 cycles. The in-cylinder pressure, mean velocities, and their standard deviations were qualitatively and quantitatively compared with experimental measurements. A good agreement was obtained. Moreover, the state and the evolution of the flow anisotropy tensor were studied by means of Lumley triangle. A deeper insight into the tumble breakdown, which was responsible for the increase of the turbulence level and CCV, was demonstrated.
Before performing the LES of the reactive flow in the fired SI engine for 60 cycles, uncertain parameters in the combustion model were studied and an adjustment was proposed. The LES with adjusted parameters provided a very good accuracy by predicting the in-cylinder pressure, the pressure CCV, the flame propagation, the CCV of flame propagation, and the flame structure in comparison with experiments. Furthermore, the numerical results were analyzed to gain a deeper understanding of combustion fundamentals in this SI engine. The following conclusions were made. First, the pressure CCV were caused by the CCV of flame propagation. Second, the flame propagation was dominated by the in-cylinder velocity field. Third, the vortex location which arose near the spark plug was very crucial for the flame kernel development. Fourth, whether an engine cycle was strong or weak could be identified in the earlier stage of the flame kernel development.
The CFD code Kiva was used and extended with Smagorinsky and WALE model to perform LES. In the context of LES, the tabulated chemistry approach Flamelet Generated Manifold (FGM) and the Artificially Thickened Flame (ATF) model were implemented into the Kiva code. The model implementation was successfully verified. These turbulence and combustion models were validated by simulating a bluff body stabilized turbulent premixed combustion system under non-reactive and reactive conditions.
Then, the LES of the non-reactive flow in the motored SI engine was conducted for 50 cycles. The in-cylinder pressure, mean velocities, and their standard deviations were qualitatively and quantitatively compared with experimental measurements. A good agreement was obtained. Moreover, the state and the evolution of the flow anisotropy tensor were studied by means of Lumley triangle. A deeper insight into the tumble breakdown, which was responsible for the increase of the turbulence level and CCV, was demonstrated.
Before performing the LES of the reactive flow in the fired SI engine for 60 cycles, uncertain parameters in the combustion model were studied and an adjustment was proposed. The LES with adjusted parameters provided a very good accuracy by predicting the in-cylinder pressure, the pressure CCV, the flame propagation, the CCV of flame propagation, and the flame structure in comparison with experiments. Furthermore, the numerical results were analyzed to gain a deeper understanding of combustion fundamentals in this SI engine. The following conclusions were made. First, the pressure CCV were caused by the CCV of flame propagation. Second, the flame propagation was dominated by the in-cylinder velocity field. Third, the vortex location which arose near the spark plug was very crucial for the flame kernel development. Fourth, whether an engine cycle was strong or weak could be identified in the earlier stage of the flame kernel development.