CFD Analysis of Diesel Pilot Injection for Dual-Fuel Diesel–Hydrogen Engines

Formulation and importance of conservative transport in non-premixed flamelet models

Representative Interactive Flamelet (RIF) and Detailed Chemistry based combustion models are two commonly used combustion models for non-premixed diesel engine simulations. RIF performs transient chemistry calculations on a one-dimensional grid based on the mixture fraction coordinate. Hence, the chemistry calculations are essentially decoupled from the computational fluid dynamics (CFD) grid. The detailed chemistry model, on the other hand, solves transient chemistry in the 3D CFD domain. An efficient parallelization strategy is used for the computation of the multiple flamelets RIF model. The multiple flamelets RIF and detailed chemistry combustion models are applied for modeling a constant volume spray combustion case and a diesel engine case, with a view to compare the differences between the two models. Results for ignition delay, flame lift-off length, cylinder pressure, and emissions are compared with experimental data. The effect of number of flamelets is evaluated. Finally, the effect of spray cooling is investigated based on the results from the RIF model and the detailed chemistry based combustion model.

An extended flamelet model for multiple injections in DI Diesel engines

Subject of this work is the recently introduced extended Representative Interactive Flamelet (RIF) model for multiple injections. First, the two-dimensional laminar flamelet equations, which can describe the transfer of heat and mass between two-interacting mixture fields, are presented. This is followed by a description of the various mixture fraction and mixture fraction variance equations that are required for the RIF model extension accounting for multiple injection events. Finally, the modeling strategy for multiple injection events is described: Different phases of combustion and interaction between the mixture fields resulting from different injections are identified. Based on this, the extension of the RIF model to describe any number of injections is explained. Simulation results using the extended RIF model are compared against experimental data for a Common-Rail DI Diesel engine that was operated with three injection pulses. Simulated pressure curves, heat release rates, and pollutant emissions are found to be in good agreement with corresponding experimental data. For the pilot injection and the main or post injection, respectively, different ignition phenomena are pointed out and the influence of the scalar dissipation rate on these ignition phenomena is detailly investigated. 2009 SAE International.

Diesel engine simulation results using two different combustion models are presented in this study, namely the representative interactive flamelet (RIF) model and the direct integration of computational fluid dynamics and CHEMKIN. Both models have been implemented into an improved version of the KIVA code. The KIVA/RIF model uses a single flamelet approach and also considers the effects of vaporization on turbulence-chemistry interactions. The KIVA/CHEMKIN model uses a direct integration approach that solves for the chemical reactions in each computational cell. The above two models are applied to simulate combustion and emissions in diesel engines with comparable results. Detailed comparisons of predicted heat release data and in-cylinder flows also indicate that both models predict very similar combustion characteristics. This is likely due to the fact that after ignition, combustion rates are mixing controlled rather than chemistry controlled under the diesel conditions studied.

Diesel engine simulation results using two different combustion models are presented in this study, namely the Representative Interactive Flamelet (RIF) model and the direct integration of CFD and CHEMKIN. Both models have been implemented into an improved version of the KIVA code. The KIVA/RIF model uses a single flamelet approach and also considers the effects of vaporization on turbulence-chemistry interactions. The KIVA/CHEMKIN model uses a direct integration approach that solves for the chemical reactions in each computational cell. The above two models are applied to simulate combustion and emissions in diesel engines with comparable results. Detailed comparisons of predicted heat release data and in-cylinder flows also indicate that both models predict very similar combustion characteristics. This is likely due to the fact that after ignition, combustion rates are mixing controlled rather than chemistry controlled under the diesel conditions studied.

Representative interactive flamelet model and flamelet/progress variable model for supersonic combustion flows

The present study is focused on the development of the RIF (Representative Interactive Flamelet) model which can overcome the shortcomings of conventional approach based on the steady flamelet library. Due to the ability for interactively describing the transient behaviors of local flame structures with CFD solver, the RIF model can effectively account for the detailed mechanisms of NOx formation including thermal NO path, prompt and nitrous NOx formation, and reburning process by hydrocarbon radical without any ad-hoc procedure. The flamelet time of RIFs within a stationary turbulent flame may be thought to be Lagrangian flight time. In context with the RIF approach, this study adopts the Eulerian Particle Flamelet Model (EPFM) with mutiple flamelets which can realistically account for the spatial inhomogeneity of scalar dissipation rate. In order to systematically evaluate the capability of Eulerian particle flamelet model to predict the precise flame structure and NO formation in the multi-dimensional elliptic flames, two methanol bluffbody flames with two different injection velocities are chosen as the validation cases. Numerical results suggest that the present EPFM model has the predicative capability to realistically capture the essential features of flame structure and NOx formation in the bluff-body stabilized flames.

<div class="section abstract"><div class="htmlview paragraph">A wide variety of spray models and their associated sub-models exist to assist with numerical spray development studies in the many applicable areas viz., turbines, internal combustion engines etc. The accuracy of a simulation when compared to the experiments varies, as these models chosen are varied. Also, the computational grid plays a crucial role in model correctness; a grid-converged CFD study is more valuable and assists in proper validation at later stages. Of primary relevance to this paper are the combustion models for a grid-converged Lagrangian spray modeling scenario. CONVERGE CFD code is used for simulation of split injection diesel (n-heptane) sprays and a structured methodology, using RNG k-ε turbulence model, is followed to obtain a grid-converged solution for the key Computational Fluid Dynamics (CFD) parameters viz., grid size, injected parcels and spray break-up time constant. Four combustion models namely the SAGE model, the Representative Interactive Flamelet model, the 3-Zone Extended Coherent Flamelet model, and SHELL+Characteristic Time Combustion model will be examined using the grid-converged CFD settings and validation against experimental flame luminosity data will be conducted with soot (empirical Hiroyasu model) as a key validating parameter. The primary focus of this work is to select a suitable combustion model for split injection combusting sprays, which would not only exercise moderate runtimes but also keep intact the interaction physics and flame characteristics. This is achieved through analysis of flame characteristics like ignition delay, flame lift-off, flame penetration length, global structure of the flame and consequent validations with the experiments.</div></div>

An equivalent dissipation rate model for capturing history effects in non-premixed flames

Multiple injections are an important aspect in modern direct-injection diesel engine development. The representative interactive flamelet (RIF) model, which was successfully used previously for simulations of diesel engine combustion, was recently extended to model multiple injections. In this paper this new RIF model is applied to model ignition and combustion with a pilot and a main injection with various dwell times, start of injection timings, and loads. Special emphasis is placed on the ignition of the main injection. It is shown that, for the investigated cases, the main injection does not auto-ignite but it is ignited by a strained premixed flame that propagates from the pilot injection to the mixture field of the main injection. The structure of that flame and the influence of the scalar dissipation rate on the propagation speed are investigated in detail. In addition to pressure curves, modelling results for NOx and soot emissions are compared with experimental data, showing good agreement.

Scalar dissipation rate based multi-zone model for early-injected and conventional diesel engine combustion

In this work, a progress variable approach is used to model diesel spray ignition with detailed chemistry. The flow field and the detailed chemistry are coupled using the flamelet assumption. A flamelet progress variable is transported by the computational fluid dynamics (CFD) code. The progress variable source term is obtained from an unsteady flamelet library that is evaluated in each grid cell. The progress variable chosen is based on sensible enthalpy. By using an unsteady flamelet library for the progress variable, the impact of local effects, for example variations in the turbulence field, effects of wall heat transfer etc. on the autoignition chemistry can be considered on a cell level. The coupling between the unsteady flamelet library and the transport equation for total enthalpy follows the ideas of the representative interactive flamelet (RIF) approach. The method can be compared to having an interactive flamelet in each computational cell in the CFD grid. The results obtained using the proposed model are compared to results obtained using the RIF model. Differences are exhibited during the autoignition process. After ignition, the results obtained using the proposed model and RIF are virtually identical. The model was used to study lift-off lengths in sprays as function of nozzle diameter and injection pressure. A good agreement between model predictions and experimental trends was found.

Numerical simulation and laser-based imaging of mixture formation, ignition, and soot formation in a diesel spray

<div class="htmlview paragraph">Three different approaches for modeling diesel engine combustion are compared against cylinder pressure, NOx emissions, high-speed soot luminosity imaging, and 2-color thermometry data from a heavy-duty DI diesel engine. A characteristic time combustion (KIVA-CTC) model, a representative interactive flamelet (KIVA-RIF) model, and direct integration using detailed chemistry (KIVA-CHEMKIN) were integrated into the same version of the KIVA-3v computer code. In this way, the computer code provides a common platform for comparing various combustion models. Five different engine operating strategies that are representative of several different combustion regimes were explored in the experiments and model simulations. Two of the strategies produce high-temperature combustion with different ignition delays, while the other three use dilution to achieve low-temperature combustion (LTC), with early, late, or multiple injections.</div> <div class="htmlview paragraph">Comparison of simulated cylinder pressure and heat-release rates with the experimental results shows that all the models predict the cylinder pressure and heat release rate reasonably well. The KIVA-CTC model under-predicts ignition delay and over-predicts rates of combustion for the late-injection conditions. All the models predict NOx emissions trends very well but the absolute magnitudes are different from the experimental measurements. The KIVA-CHEMKIN model better predicted the emissions for LTC conditions. For the KIVA-RIF model, the NOx emissions predictions were only slightly sensitive to the number of flamelets used. With a single flamelet assumed in the entire computational domain, the emissions are under-predicted but the predictions are little better with the use of multiple flamelets. Detailed comparisons of the spatial and temporal evolution of modeled in-cylinder soot with the experimental data are also presented using the direct integration and the KIVA-RIF combustion models, and good qualitative agreement is found with both the models.</div>