Search

Scramjets have been identified as a potential long-term replacement for rocket and ramjet propulsion systems due to their enhanced performance at high Mach numbers. The introduction of solid fuels in these scramjet systems allows for shaping of the solid fuel cavity by additive manufacturing and introduces the possibility of enhancing combustion rates and stability. The present investigation aims to develop a coupled, high-order computational model to study the combustion of solid fuel scramjets. The primary objectives are to identify the effects of changing geometry on combustion and to better characterize the combustion process and flow patterns within a solid fuel scramjet engine. The high-Mach number of the air inflow over a scramjet cavity introduces a strong coupling between fluid dynamics, combustion, and regression time scales. Existing models often use simplified treatments of melt-layer conditions and combustion models that over-predict experimental rates, along with highly dissipative numerical schemes that inhibit the study of thermo-acoustic interactions between coherent pressure waves and the burning walls of the cavity. These limitations in current models suggest the need for a Navier-Stokes solver based on a high-order, discontinuous Galerkin method, incorporating melt layer equations and enhanced combustion manifolds. These manifolds should account for the effects of pressure and high oxidizer temperatures on flamelet dynamics. The focus is on modeling the flow field with accurate chemical heat release and residence time, to better study the effects of heat flux on the solid surface and the resulting coupling. An investigation of solid fuel scramjets was performed, and the numerical methodology with which the problem was tackled is described. A novel combustion mechanism was developed using a counterflow burner to study the combustion and regression of solid model fuel polymethyl methacrylate (PMMA). The diffusion flame between the fuel and oxidizer was studied numerically using a solid fuel decomposition and melt layer model to simulate convection and pyrolysis of the material. This model was validated using new experimental data as well as previously published works. The foam layer parameters are critical to the success of the validation. Results showed that the increased residence time of the gas in the bubbles facilitates the fuel breakdown. Fully coupled fuel injection and solid fuel surface monitoring was implemented based on this counterflow model and was a function of heat flux. Fuel regression was handled using adaptive control points for a B-Spline basis that updates based on surface movement. This methodology was used due to its resilience against the creation of surface discontinuities likely to result from large temperature gradients during combustion. Fourth-order computational simulations of ramjet combustion without regressing fuel walls using an in-house Discontinuous Galerkin approach were performed with a fully conjugate solution for the thermal wave in the solid. Results in ramjet geometries showed the turbulent combustion strongly affects the heat feedback to the walls and thus increases both the regression and fuel injection rates. Scramjet geometries were also simulated using the flamelet-progress variable approach in two different oxidizer conditions. All of these simulations showed strong agreement with experimental data and helped to uncover flame holding characteristics of the scramjet cavities and the strong coupling between the recirculation region and pyrolysis of fuel. The analysis has led to a better understanding of the effects of solid fuel scramjet geometries on mixing, enhanced modeling of acoustic instabilities in solid fuel air-breathing propulsion, and improved fuel chemistry modeling. It has been shown that cavity design significantly influences heat transfer to the solid fuel in both ramjet and scramjet conditions. The presence and thickness of the melt layer will guide designs that aim to reduce or enhance mechanical removal of fuel. Additionally, ramjet results indicate that longer cavities can couple with acoustics to induce self-excited conditions, leading to increased heat transfer to the solid. The importance of self-sustained instability and its coupling with melt layer fuel injection will contribute to improved acoustic stability. Developing pressure/temperature-dependent manifolds and melt layer models will advance our understanding of solid fuel supersonic combustion and its effects on phenomena such as blowout, fuel residence time, and solid fuel dual-mode transition.

The advent of detonation-based propulsion systems represents an opportunity for more sustainable combustion processes and hypersonic travel. In regular detonations, some yet to be resolved instabilities are attributed to the propagation and collision of triple points, formed at the intersection of a Mach stem, an incident shock and a transverse wave. Over time, the tracks observed by these points form a structure made of diamond-shaped cells. Ultimately, The ability to sustain these instabilities plays a key role in the propagation of detonations. The present work unveils the dynamics of gaseous detonations at a sub-cellular level. The experiments are conducted with hydrogen fuel which is of great potential for detonation engine applications. The hydrogen-oxygen mixtures are held at stoichiometry and the nitrogen dilution in oxygen is varied from 30% to 70%. This allows to observe the effects of activation energy through the dilution on the sub-cellular wave dynamics. Measurements of cell sizes and wave velocity are reported through shadowgraph imaging. A new methodology is developed for the simultaneous resolution of the velocity field and cellular structure. The statistical analysis is made possible due to the design of a fully automated detonation facility. The experiments are conducted in a thin channel to minimize gradients in the third direction and confine the detonation cells to a plane. The results in cell sizes are in good agreement with the literature and expand the conditions reported thus far. Local observations of the velocity within the cells are used to explain the regularity of the overall wave speed, found to increase at lower dilution. Lastly, high fidelity simulations are conducted to model the cellular structure in hydrogen-air oblique detonation waves. Similarly to the experiments, the velocity field is extracted along the detonation cells and reveals the effects of wave curvature on triple point dynamics.

Use of biofuels, especially in automotive applications, is a growing trend due to their potential to lower greenhouse gas emissions from combustion. Ketones are a class of biofuel candidates which are produced from cellulose. However, ketones received rather scarce attention from the combustion community compared to other classes such as, alcohols, esters, and ethers. There is little knowledge on their combustion performance and pollutant generation. Hence their combustion chemistry needs to be investigated in detail. Diisopropyl ketone (DIPK) is a promising biofuel candidate, which is produced using endophytic fungal conversion. A detailed understanding of the combustion kinetics of the oxidation of DIPK in advanced engines such as, the homogeneous charge compression ignition (HCCI) engine is warranted. This dissertation concentrates on the combustion kinetics of DIPK over a wide range of temperature and pressure with a focus on HCCI engine application. An existing DIPK kinetic mechanism has been reviewed and a single zone HCCI engine model has been modeled and validated against recent experimental data from Sandia National Lab. Therefore different HCCI modeling assumptions were tested and the DIPK reaction mechanism was modified with missing reactions and the required thermochemical data. As a result, the HCCI pressure trace, heat release rate and reactivity have been improved. In order to improve the ignition delay time simulation results, the low temperature oxidation of DIPK was studied as the fuel chemistry effects on the autoignition behavior becomes important in low temperature. Therefore DIPK low temperature oxidation experimental data was obtained from the synchrotron photoionization experiments conducted at the Advanced Light Source (ALS) so that the primary products as well as the dominant oxidation pathways are identified. Furthermore, the aldehydes oxidation, as a result of partial or incomplete combustion and as the primary stable intermediate products in oxidation and pyrolysis of biofuel were studied at low temperature in ALS. A high temperature reaction mechanism was created using the reaction class approach. The reaction mechanism for DIPK was improved using the experimental data along with quantum chemical calculation of activation energies and barriers as well as vibrational modes for the important reactions identified in ALS experiment. The rate constants for important reactions were calculated based on modified Arrhenius equation. DIPK oxidation and pyrolysis were studied at high temperature and pressure using UCF shock tube. The ignition delay times as well as the product (methane) time histories were investigated and used as validation targets for the new model.

Advanced combustion strategies provide significant efficiency and emissions benefits compared to conventional spark ignited (SI) combustion, but challenges related to combustion control and load limits have made these technologies difficult to implement in practical systems. Until now, low cost reduced order models necessary for large parametric and multi-cycle studies capable of accurately capturing the full range of combustion modes from homogeneous charge compression ignition (HCCI) and spark-assisted compression ignition (SACI) to SI have not been available. This important computational gap for advanced combustion engine research was the primary motivation for this doctoral work. The outcomes of this study include powerful new tools to evaluate advanced combustion strategies as well as novel methods to incorporate important advanced combustion characteristics into reduced order models. A reduced order thermodynamic model of advanced SACI combustion was first proposed. The model was used with available experimental data and previous high fidelity simulation results to develop a new empirical auto-ignition burn rate model that captures the effects of ignition timing, composition, temperature, pressure, engine speed, stratification and flame propagation. A complete engine model was then developed and incorporated into the commercial simulation software GT-Power. The model included chemical kinetics for low temperature heat release and auto-ignition detection and the empirical burn rate model for post-ignition heat release, as well as a new flame propagation model with improved physical groundings. The calibrated engine model showed good agreement with experimental trends of HCCI, SACI and SI combustion modes. The engine model was then used to assess practical strategies for accessing the advanced combustion regime and improving engine efficiency. The results showed HCCI and SACI provide a pathway for significant efficiency benefits compared to throttled SI, with efficiency improvements between 15-25% across a range of loads from 1-7 bar BMEP. Further efficiency gains appear possible beyond the experimentally observed SACI limit. As a further exercise, the load extension potential of boosted SACI combustion was conceptually investigated using a simple thermodynamic framework incorporating the empirical burn rate model and practical operating constraints. The results indicate boosted SACI can nearly double the maximum engine load compared to naturally aspirated operation.

Accurate models for the mixing of fuel and oxidizer at small, unresolved flow length scales are critical to the predictive skill of large eddy simulation (LES) of turbulent combustion. Subfilter scalar variance and subfilter scalar dissipation rate are important parameters in combustion modeling approaches based on a conserved scalar, but are prone to numerical and modeling errors due to the nature of practical LES computations. This work examines the errors incurred in these models using a novel method that couples LES scalar modeling with direct numerical simulation (DNS) of homogeneous isotropic turbulence and offers modeling and numerical techniques to address these errors. In the coupled DNS-LES method, DNS velocity fields are evolved simultaneously with LES scalar fields. The filtered DNS velocities are supplied to the LES scalar equations, instead of solving the LES momentum equations. This removes the effect of errors in the filtered scalar evolution from the scalar modeling analysis. Results obtained using the coupled DNS-LES approach, which permits detailed study of physics-related and numerical errors in scalar modeling, show that widely used algebraic dynamic models for subfilter scalar variance lack accuracy due to faulty equilibrium modeling assumptions and sensitivity to numerical error. Transport equation models for variance show superior performance, provided that the scalar dissipation rate model coefficient is set appropriately. For this purpose, a new dynamic approach for nonequilibrium modeling of subfilter scalar dissipation rate is developed and validated through a priori tests in an inhomogeneous jet flow and using the coupled DNS-LES method for assessment of numerical error effects. Explicit filtering is assessed as means to control numerical error in LES scalar modeling and the scalar equations are reformulated to account for the explicit filtering technique. Numerical convergence of the mean subfilter scalar variance prediction with increasing grid resolution is demonstrated.

Accurate models for the mixing of fuel and oxidizer at small, unresolved flow length scales are critical to the predictive skill of large eddy simulation (LES) of turbulent combustion. Subfilter scalar variance and subfilter scalar dissipation rate are important parameters in combustion modeling approaches based on a conserved scalar, but are prone to numerical and modeling errors due to the nature of practical LES computations. This work examines the errors incurred in these models using a novel method that couples LES scalar modeling with direct numerical simulation (DNS) of homogeneous isotropic turbulence and offers modeling and numerical techniques to address these errors. In the coupled DNS-LES method, DNS velocity fields are evolved simultaneously with LES scalar fields. The filtered DNS velocities are supplied to the LES scalar equations, instead of solving the LES momentum equations. This removes the effect of errors in the filtered scalar evolution from the scalar modeling analysis. Results obtained using the coupled DNS-LES approach, which permits detailed study of physics-related and numerical errors in scalar modeling, show that widely used algebraic dynamic models for subfilter scalar variance lack accuracy due to faulty equilibrium modeling assumptions and sensitivity to numerical error. Transport equation models for variance show superior performance, provided that the scalar dissipation rate model coefficient is set appropriately. For this purpose, a new dynamic approach for nonequilibrium modeling of subfilter scalar dissipation rate is developed and validated through a priori tests in an inhomogeneous jet flow and using the coupled DNS-LES method for assessment of numerical error effects. Explicit filtering is assessed as means to control numerical error in LES scalar modeling and the scalar equations are reformulated to account for the explicit filtering technique. Numerical convergence of the mean subfilter scalar variance prediction with increasing grid resolution is demonstrated.

A diesel engine was tested in a laboratory setting over the range that it is expected to operate in a hybrid vehicle. An efficient experiment design was created to minimize both the amount of required data and error introduced into the final results. Through combustion modeling, collected data for the engine’s intake air and fuel mass flow as well as volumetric exhaust content data was used to determine levels of engine-out mass flow of NOx over the engine’s operating domain. Several fuel consumption and NOx emission parameters were calculated and regression models were developed to produce baseline engine maps. Based on the baseline maps, targeted engine operation points were selected to examine how the vehicle’s hybrid control strategy might be tuned towards engine operation that provides lowered NOx emission at the cost of fuel economy. Results show that quite significant levels of NOx reduction can be had at a small cost in increased fuel use. However, even at reduced engine-out levels, NOx emission is still relatively considerable in terms of meeting standards set for by the United States Environmental Protection Agency. The use and effectiveness of selective catalyst reduction by injection of urea into the exhaust stream to treat engine-out NOx is also explored in this thesis.

In the first part of this work, instability is correctly predicted for a Rijke tube with a new two-term acoustic forcing term derived from a one-dimensional flame dynamics model. The new two-term acoustic forcing term, which is comprised of the summation of chemical heat release rate and heat transfer due to convection, correctly predicts instability where older models of acoustic forcing based solely on chemical heat release rate incorrectly predicted stability. This stability analysis correctly predicts the inlet conditions of the instability in addition to the frequency of instability. In the second part of this work, networks of dynamic well-stirred reactors are used to model qualitative behavior observed in turbulent combustion. First a model of dynamic well-stirred reactor is derived, and then several reactors are coupled together by recirculation. The dynamics of the various models are computed and assessed. The models exhibit interesting behavior that has been viewed experimentally including hysteresis and peaking in the dynamic response.

The value of combustion models cannot be overstated. As a major component of engine models they are widely used as research tools in the continuing effort to improve existing combustion systems and to develop completely new concepts. Models also find use in the classroom where they allow students to manipulate control parameters and study the effect on performance, efficiency and pollutant formation without ever visiting an engine test cell. Models of varying complexity are available for premixed combustion but there is a lack of published work on simplified modeling of the more complex case of stratified-charge combustion. In this investigation, such a model is developed and validated against experimental data for stratified combustion in a constant volume chamber. This first iteration is intended as an oversimplified foundation to which additional complexity can be added in subsequent iterations after weaknesses are identified. Data are obtained for a range of test cases from a purpose-built test stand. Each of the experimental cases is simulated and the results compared. Simulation results for pressure and flame radius match well with the experimental data. The timing of key events is also predicted with relative accuracy. Significant errors are observed only in the period where the model is predicted to break down. The overall positive result shows that it should be possible to capture the essential characteristics of stratified-charge combustion in a simple model, given adequate development time.