Modeling Non-Premixed Combustion Using Tabulated Kinetics and Different Fame Structure Assumptions

Abstract

Nowadays, detailed kinetics is necessary for a proper estimation of both flame structure and pollutant formation in compression ignition engines. However, large mechanisms and the need to include turbulence/chemistry interaction introduce significant computational overheads. For this reason, tabulated kinetics is employed as a possible solution to reduce the CPU time even if table discretization is generally limited by memory occupation. In this work the authors applied tabulated homogeneous reactors (HR) in combination with different turbulent-chemistry interaction approaches to model non-premixed turbulent combustion. The proposed methodologies represent good compromises between accuracy, required memory and computational time. The experimental validation was carried out by considering both constant-volume vessel and Diesel engine experiments. First, the ECN Spray A configuration was simulated at different operating conditions and results from different flame structures are compared with experimental data of ignition delay, flame lift-off, heat release rates, radicals and soot distributions. Afterwards, engine simulations were carried out and computed data are validated by cylinder pressure and heat release rate profiles.

Experimental Validation of Combustion Models for Diesel Engines Based on Tabulated Kinetics in a Wide Range of Operating Conditions

“…2 and the user is referred to [12] for further details. To ensure a consistency between TWM and the corresponding approach based on direct integration, avoiding anticipated ignition due to progress variable diffusion from rich to lean side after the cool flame, reaction rates are set to zero in regions where dual-stage ignition does not happen, corresponding approximately to an equivalence ratio value φ = 3.…”

“…A preliminary assessment and validation of the TFPV model was performed in a similar way as done for the other approaches in [12]. In particular, the well-known constantvolume spray-A experiment from ECN [28,29,30] Figs.…”

“…TWM, TPPDF and TRIF are extensively described and validated under constant-volume combustion conditions in [12]. Fig.…”

“…After every time-step, the progress variable C is evaluated and the equivalent reactor chemical composition is computed by means of the virtual species approach [14,12]. Progress variable is defined as in [13], with C being equal to the heat released by combustion, computed as the difference between the current and the initial value of the reactor formation enthalpy, also known as h 298 :…”

“…The most widely used approaches consider a progress variable to characterize the advancement of the fuel oxidation reactions, by solving a transport equation with a source term which depends on local thermodynamic conditions, the progress variable itself and parameters characterizing the assumed flame structure, like the mixture fraction variance or the scalar dissipation rate [7,8,9,10,11]. In a previous work [12], the authors have implemented different models for Diesel combustion into the Lib-ICE code based on tabulated kinetics, demonstrating its consistency with the approaches based on direct-integration following extensive validation against the spray-A experimental data-sets from the Engine Combustion Network. Purpose of this work is the extension of the previously proposed approach for tabulated kinetics to include an additional combustion model as well as the possibility to predict NO x emissions.…”