Search Results

Computational fluid dynamics of gas-fueled large-bore spark ignition engines with pre-chamber ignition can speed up the design process of these engines provided that 1) the reliability of the results is not affected by poor meshing and 2) the time cost of the meshing process does not negatively compensate for the advantages of running a computer simulation. In this work a flame propagation model that runs with arbitrary hybrid meshes was developed and coupled with the KIVA4-MHI CFD solver, in order to address these aims. The solver follows the G-Equation level-set method for turbulent flame propagation by Tan and Reitz, and employs improved numerics to handle meshes featuring different cell types such as hexahedra, tetrahedra, square pyramids and triangular prisms. Detailed reaction kinetics from the SpeedCHEM solver are used to compute the non-equilibrium composition evolution downstream and upstream of the flame surface, where chemical equilibrium is instead assumed.

Three time integration schemes and four finite volume interpolation schemes for the convection term in momentum equation were tested under turbulent planar gas jet and Sandia non-reacting vaporizing Spray-H cases. The three time integration schemes are the first-order Euler implicit scheme, the second-order backward scheme, and the second-order Crank-Nicolson scheme. The four spatial interpolation schemes are cubic central, linear central, upwind, and vanLeer schemes. Velocity magnitude contour, centerline and radial mean velocity and Reynolds stress profiles for the planar turbulent gas jet case, and fuel vapor contour and liquid and vapor penetrations for the Diesel spray case predicted by the different numerical schemes were compared. The sensitivity of the numerical schemes to mesh resolution was also investigated. The non-viscosity based dynamic structure subgrid model was used. The numerical tool used in this study was OpenFOAM.

The paper reports the application of a look-up table approach within a LES combustion modelling framework for the prediction of knock limit in a highly downsized turbocharged DISI engine. During experimental investigations at the engine test bed, high cycle-to-cycle variability was detected even for relatively stable peak power / full load operations of the engine, where knock onset severely limited the overall engine performance. In order to overcome the excessive computational cost of a direct chemical solution within a LES framework, the use of look-up tables for auto-ignition modelling perfectly fits with the strict mesh requirements of a LES simulation, with an acceptable approximation of the actual chemical kinetics. The model here presented is a totally stand-alone tool for autoignition analysis integrated with look-up table reading from detailed chemical kinetic schemes for gasoline.

Fundamental simulations in a quiescent cell under adiabatic conditions were made to understand the effect of temperature, equivalence ratio and the components of the recirculated exhaust gas, viz., CO2 and H2O, on the combustion of n-Heptane. Simulations were made in single phase in which evaporated n-Heptane was uniformly distributed in the domain. Computations were made for two different temperatures and four different EGR levels. CO2 or H2O or N2was used as EGR. It was found that the initiation of the main combustion process was primarily determined by two competing factors, i.e., the amount of initial OH concentration in the domain and the specific heat of the mixture. Further, initial OH concentration can be controlled by the manipulating the ambient temperature in the domain, and the specific heat capacity of the mixture via the mixture composition. In addition to these, the pre combustion and the subsequent post combustion can also be controlled via the equivalence ratio.

A novel algebraic similarity model for subgrid scalar dissipation rate has been developed as part of the Large Eddy Simulation (LES) package KIVA3V-LES for diesel engine study. The model is proposed from an a priori study using Direct Numerical Simulation (DNS) of forced isotropic turbulence. In the a posteriori test, fully resolved turbulent passive scalar field measurements are used to validate the model in actual engine flows. For reason of the length limit by SAE and the specific interest in engine applications, only a prior test and a posteriori test in engine flows are included in this paper. A posteriori tests in isotropic cube flow, turbulent round jet and flame cases will be presented in separate papers. An engine LES simulation of multi consecutive cycles was performed in this study.

Diesel engine combustion is simulated using Large Eddy Simulation (LES) with a multi-mode combustion (MMC) model. The MMC model is based on the combination of chemical kinetics, chemical equilibrium, and quasi-steady flamelet calculations in different local combustion regimes. The local combustion regime is identified by two combustion indices based on the local temperature and the extent of mixture homogeneity. The LES turbulence model uses the dynamic structure model (DSM) for sub-grid stresses. A new spray model in the LES context is used, and the Reynolds-averaged Navier-Stokes (RANS) based wall model is retained with the LES derived scales. These models are incorporated in the KIVA3V-ERC-Release 2 code for engine combustion simulations. A wide range of diesel engine operating conditions were chosen to validate the combustion model.

Detailed knowledge of hydrocarbon fuel combustion chemistry has grown tremendously in recent years. However, the gap between detailed chemistry and computational fluid dynamics (CFD) remains, because of the high cost of solving detailed chemistry in a large number of computational cells. This paper presents the results of applying a suite of techniques aimed at closing this gap. The techniques include use of a surrogate blend optimizer and a guided mechanism reduction methodology, as well as advanced methods for efficiently and accurately coupling the pre-reduced kinetic models with the multidimensional transport equations. The advanced methods include dynamic adaptive chemistry (DAC) and dynamic cell clustering (DCC) algorithms.

Neural networks are useful tools for optimization studies since they are very fast, so that while capturing the accuracy of multi-dimensional CFD calculations or experimental data, they can be run numerous times as required by many optimization techniques. This paper describes how a set of neural networks trained on a multi-dimensional CFD code to predict pressure, temperature, heat flux, torque and emissions, have been used by a genetic algorithm in combination with a hill-climbing type algorithm to optimize operating parameters of a diesel engine over the entire speed-torque map of the engine. The optimized parameters are mass of fuel injected per cycle, shape of the injection profile for dual split injection, start of injection, EGR level and boost pressure. These have been optimized for minimum emissions. Another set of neural networks have been trained to predict the optimized parameters, based on the speed-torque point of the engine.

Neural networks have been used for engine computations in the recent past. One reason for using neural networks is to capture the accuracy of multi-dimensional CFD calculations or experimental data while saving computational time, so that system simulations can be performed within a reasonable time frame. This paper describes three methods to improve upon neural network predictions. Improvement is demonstrated for in-cylinder pressure predictions in particular. The first method incorporates a physical combustion model within the transfer function of the neural network, so that the network predictions incorporate physical relationships as well as mathematical models to fit the data. The second method shows how partitioning the data into different regimes based on different physical processes, and training different networks for different regimes, improves the accuracy of predictions.

A modeling study has been conducted in order to characterize the heat transfer in an automotive diesel exhaust system. The exhaust system model, focusing on 2 exhaust pipes, has been created using a transient 1-D engine flow network simulation program. Model results are in excellent agreement with experimental data gathered before commencement of the modeling study. Predicted pipe exit stream temperatures are generally within one percent of experimental values. Sensitivity analysis of the model was the major focus of this study. Four separate variables were chosen for the sensitivity analysis. These being the external convective heat transfer coefficient, external emissivity, mass flow rate of exhaust gases, and amplitude of incoming pressure fluctuations. These variables were independently studied to determine their contribution to changes in exhaust gas stream temperature and system heat flux. There are two primary benefits obtained from conducting this analysis.

The present study investigated HCCI combustion in a heavy-duty diesel engine both experimentally and numerically. The engine was equipped with a hollow-cone pressure-swirl injector using gasoline direct injection. Characteristics of HCCI combustion were obtained by very early injection with a heated intake charge. Experimental results showed an increase in NOx emission and a decrease in UHC as the injection timing was retarded. It was also found that optimization can be achieved by controlling the intake temperature together with the start-of-injection timing. The experiments were modeled by using an engine CFD code with detailed chemistry. The CHEMKIN code was implemented into KIVA-3V such that the chemistry and flow solutions were coupled. The model predicted ignition timing, cylinder pressure, and heat release rates reasonably well. The NOx emissions were found to increase as the injection timing was retarded, in agreement with experimental results.

Artificial neural networks (ANN) have been recognized as universal approximators for nonlinear continuous functions and actively applied in engine research in recent years [1, 2, 3, 4, 5, 6, 7 and 8]. This paper describes the methodology and results of using the ANN to model a turbocharged DI diesel engine. The engine was simulated using the CFD code (KIVA-ERC) over a wide range of operating conditions, and numerical simulation results were used to train the ANN. An efficient data collection methodology using the Design of Experiments (DOE) techniques was developed to select the most characteristic engine operating conditions and hence the most informative data to train the ANN. This approach minimizes the time and cost of collecting training data from either computational or experimental resources. The trained ANN was then used to predict engine parameters such as cylinder pressure, cylinder temperature, NOx and soot emissions, and cylinder heat transfer.

DPF regenerations involve a trade-off between fuel economy and DPF durability. High temperature regenerations of DPFs have fewer fuel penalties but simultaneously tend to give higher substrate temperatures, which can reduce thermal reliability. In order to weaken the trade-off, the integrated system-level model [1,2,3,4] is used to conduct optimization studies and explore novel regeneration strategies for DPF aftertreatment devices. The integrated model developed in the Engine Research Center (ERC) includes sub-models for engines, emissions, aftertreatment devices and controllers. Based on the engine and regeneration fuel economy, multiple and single cycle regeneration tests are performed and analyzed. The optimal soot loadings to initiate and terminate regenerations are discussed. A pulsed regeneration strategy, which is characterized by injecting multiple pulses of fuel (upstream of a DOC) during regenerations, is investigated.

In this paper, large eddy simulation (LES) coupled with two uncertainty quantification (UQ) methods, namely latin-hypercube sampling (LHS) and polynomial chaos expansion (PCE), have been used to quantify the effects of model parameters and spray boundary conditions on diesel and gasoline spray simulations. Evaporating, non-reacting spray data was used to compare penetration, mixture fraction and spray probability contour. Two different sets of four uncertain variables were used for diesel and gasoline sprays, respectively. UQ results showed good agreement between experiments and predictions. UQ statistics indicated that discharge coefficient has stronger impact on gasoline than diesel sprays, and spray cone angle is important for vapor penetration of both types of sprays. Additionally, examination of the gasoline spray characteristics showed that plume-to-plume interaction and nozzle dribble are important phenomena that need to be considered in high-fidelity gasoline spray simulations.

A swept-volume method of calculating the volume swept by the flame during each time step is developed and used to improve the calculation of fuel reaction rates. The improved reaction rates have been applied to the ignition model and coupled with the level set G-equation combustion model. In the ignition model, a single initial kernel is formed after which the kernel is convected by the gas flow and its growth rate is determined by the flame speed and thermal expansion due to the energy transfer from the electrical circuit. The predicted ignition kernel size was compared with the available experimental data and good agreements were achieved. Once the ignition kernel reaches a size when the fully turbulent flame is developed, the G-equation model is switched on to track the mean turbulent flame front propagation.

Using non-viscosity dynamic structure Large Eddy Simulations (LES) turbulence model, spray=induced turbulence is investigated on a number of different Computational Fluid Dynamics (CFD) grids of varying mesh sizes (from 0.5 to 2 mm mesh). Turbulent flow is induced inside a quiescent chamber by liquid fuel spray and then left to decay after end of injection by virtue of its molecular viscosity and turbulent dissipation. Coherent structures (CS) of this turbulent flow are constructed and visualized using λ2 definition. Using CS, analysis is performed on the turbulent flow around the liquid spray jet. These CS from LES are then compared against the results from RANS calculations as well. The visualization of CS helps to explain the mechanism of fuel-air mixing obtained from LES results and its difference with RANS calculations.

The paper reports the application of Proper Orthogonal Decomposition (POD) to LES calculations for the analysis of combustion and knock tendency in a highly downsized turbocharged GDI engine that is currently under production. In order to qualitatively match the cyclic variability of the combustion process, Large-Eddy Simulation (LES) of the closed-valve portion of the cycle is used with cycle-dependent initial conditions from a previous multi-cycle analysis [1, 2, 3]. Detailed chemical modelling of fuel's auto-ignition quality is considered through an ad-hoc implemented look-up table approach, as a trade-off between the need for a reasonable representation of the chemistry and that of limiting the computational cost of the LES simulations. Experimental tests were conducted operating the engine at knock-limited spark advance (KLSA) and the proposed knock model was previously validated for such engine setup [3].

In this work non-reacting spray simulations are performed using two Computational Fluid Dynamic (CFD) codes, KIVA and OpenFoam. The metric used is the liquid tip penetration which is compared with experimental data from the Engine Combustion Network at Sandia National Laboratories. Some important spray sub-models, available in KIVA, are implemented in OpenFOAM so that the two codes have more common models. In addition, model coefficients and computations cells used in the simulations are the same in both codes. The differences in spray source terms formulations and other spray sub models between the codes and their effect on liquid penetration are discussed.

Large Eddy Simulation (LES) with a flamelet time scale combustion model is used to simulate diesel combustion. The flamelet time scale model uses a steady-state flamelet library for n-heptane indexed by mean mixture fraction, mixture fraction variance, and mean scalar dissipation rate. In the combustion model, reactions proceed towards the flamelet library solution at a time scale associated with the slowest reaction. This combination of a flamelet solution and a chemical time scale helps to account for unsteady mixing effects. The turbulent sub-grid stresses are simulated using a one-equation, non-viscosity LES model called the dynamic structure model. The model uses a tensor coefficient determined by the dynamic procedure and the subgrid kinetic energy. The model has been expanded to include scalar mixing and scalar dissipation. A new model for the conditional scalar dissipation has been developed to better predict local extinction.

In this work the modeling aspects of fuel vaporization are studied. To start with, the effects of vaporization model on engine simulations are studied. This is done by using two different fuel surrogates. Next a set of non-reacting spray simulations were performed under different ambient and operating conditions and for two different fuels. This was done for spray model validation and to look at the effect of vaporization model on liquid penetration length. Following an observed discrepancy in one of the spray cases, effect of ambient temperature on liquid length, two sensitivity analyses were performed. These analyses take into account the effects of each spray-sub model on vaporization and effects of spray breakup constants on liquid penetration. Using the results from the sensitivity analyses and linearized stability theory an empirical correction factor was developed to correct the spray behavior at low ambient temperatures.