The Pennsylvania State University The Graduate School Department of Mechanical and Nuclear Engineering

Transcription

1 The Pennsylvania State University The Graduate School Department of Mechanical and Nuclear Engineering AUTOIGNITION OF HYDROGEN AND SYNGAS WITH AIR IN A TURBULENT FLOW REACTOR A Thesis in Mechanical Engineering by Daniel J. Elies c 2012 Daniel J. Elies Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science December 2012

2 The thesis of Daniel J. Elies was reviewed and approved by the following: Robert J. Santoro George L. Guillet Professor of Mechanical Engineering Thesis Advisor Thomas A. Litzinger Professor of Mechanical Engineering Karen A. Thole Professor of Mechanical Engineering Head of the Department of Mechanical and Nuclear Engineering Signatures are on file in the Graduate School.

3 Abstract A good deal of attention has been given recently to combustion of syngas in gas turbines used for power generation. Syngas is a mixture of hydrogen and carbon monoxide produced from coal gasification, a process where coal is partially oxidized producing a gaseous product with high concentrations of hydrogen and carbon monoxide. Although coal gasification is not a new technology, recent interest has been spurred by concerns about climate change due in large part to increased levels of carbon dioxide in the atmosphere. Coal-fired power plants in 2011 produced approximately 46% of the electricity used in the United States, but contributed 79% of the energy related carbon dioxide emissions. The reason that coal contributes such a large portion of the carbon dioxide emissions is that the carbon to hydrogen ratio of coal is high as compared to other hydrocarbon fuels. As a result, more of the energy released from burning coal comes from the oxidation of carbon rather than hydrogen, which increases the amount of carbon dioxide emitted per kilowatt hour. To reduce these high carbon dioxide emissions, the use of syngas as fuel is part of the Clean Coal effort that will use carbon sequestration to remove carbon dioxide from the combustion products and store it underground. One of the challenges of using syngas as a fuel is the variable composition in terms of the amount of hydrogen and carbon monoxide present in the fuel. This variable composition results from the wide variety of coal that can be used to produce syngas. These compositional variations alter the combustion characteristics of syngas. Additionally, present gas turbine technology for power generation utilizes lean-premixed conditions to reduce oxides of nitrogen formation, that is fuel and hot air from the compressor are premixed prior to combustion. If premixed syngas and air was to ignite in the gas turbine premixer, severe damage would occur. Consequently, one of the combustion characteristics of particular importance is the autoignition time. Autoignition is a measure of the time for a mixture iii

4 of fuel and oxidizer at some elevated temperature to spontaneously ignite. Thus, the present study specifically addresses measurements of the autoignition time for hydrogen and hydrogen/carbon monoxide mixtures under conditions relevant to gas turbines used for power generation. Experiments were conducted using a turbulent flow reactor for the purpose of examining autoignition times for hydrogen and hydrogen/carbon monoxide mixtures with air. Experiments with only hydrogen as fuel were conducted at experimental conditions including ignition delay times of 130 and 210 ms, equivalence ratios of and 0.750, and pressures of 10 and 15 atm. Temperatures could be varied between 800 and 900 K using a combination of an electric heater and a hydrogen and oxygen fueled preburner. Experiments were also conducted with a mixture of hydrogen and carbon monoxide, to simulate Syngas, at a pressure of 15 atm, an ignition delay time of 130 ms, and at equivalence ratios and The temperatures that resulted in autoigntion of hydrogen at the above conditions ranged between 840 to 890 K. The temperatures that resulted in autoignition for the syngas experiments were 8 to 23 K lower than those observed for hydrogen at the same pressure and equivalence ratio conditions. Previous experiments conducted at similar conditions using hydrogen and hydrogen/carbon monoxide mixtures in turbulent flow reactors reported much shorter autoignition times. Furthermore, these earlier studies showed significant disparities, as much as 1 or 2 orders of magnitude, between predictions of autoignition using homogeneous chemical kinetic models and measurements of autoignition time. The results of the current study show agreement with the homogeneous chemical kinetics model within at most a factor of five. iv

5 Table of Contents List of Figures List of Tables Acknowledgments vii x xi Chapter 1 Introduction Motivation Literature Review Autoignition and Ignition Delay Time Shock Tubes Flow Reactors Other Flow Reactor Facilities Experiment Chapter 2 Experimental Apparatus and Procedure Components Flow Reactor Test Section Main Injector Radial Injector Axial Injector Heated Air System Preburner and Fuel Heating System Valves Nozzle v

6 2.1.7 Critical Orifices Clamshell Heater Experimental Variables Experimental Procedure Setup Determining Test Temperature Chapter 3 Data Analysis Introduction Mixture Temperature Energy Balance Case 1: No-Ignition with Long Fuel Duration Case 2: Ignition or No-Ignition with Short Duration Fuel Case 3: Radial Injector Discussion Autoignition Delay Time Summary Chapter 4 Results and Discussion Introduction Interpreting Results Results Axial Injector Trends Syngas Radial Injector Discussion Chapter 5 Conclusion 83 Appendix A Appendix B Bibliography 94 vi

7 List of Figures 1.1 Diagram of a IGCC system [1] Comparison of flow reactor data and shock tube data to chemical kinetics model [7]. Flow reactor data is adjusted to 20 atm Test apparatus schematic Test section model showing modular sections, main injector, and nozzle Pictures of radial injector Radial injector model with cross section view. Air flows from left to right Model of axial injector Cross section model of preburner with labels. Igniter chamber not shown Preburner and fuel mixing chamber Characteristic fuel pressure trace taken downstream of orifice Cross section view of nozzle showing water passages Schematic of the preburner indicating locations of temperatures discussed Characteristic flow reactor pressure trace Plot of reactor pressure trace during fuel injection for an experiment with an ignition event. The ignition event was characterized by the onset of the pressure oscillations. The valve was closed when the system reads a temperature in the reactor above a threshold. The threshold temperature was set at 1300 F (978 K) Interpreting data while conducting experiments Temperature trace of a run with a long duration where there was ignition Temperature traces for a run with a 3.9 s duration where an ignition event was not observed vii

8 3.3 Correlations that relate T2 at a specific time to the steady state mixture temperature as measured by Tcl Comparison of raw data with the normalized data. The lines plotted are different thermocouples in the reactor recorded at 50 khz Series of experiments taken with the axial injector at an equivalence ratio of 0.375, a pressure of 15 atm, and the in diameter nozzle ( 130 ms). The experiments that resulted in no-ignition are plotted with open symbols and the experiments that resulted in ignition are plotted with closed symbols. The no-ignition points are plotted as a function of the residence time corresponding to the exit of the flow reactor ( 130 ms). The ignition points are plotted at the residence time calculated from the location of ignition as determined by the thermocouples CHEMKIN simulation of hydrogen with air in homogeneous batch reactor using the O Conaire model at an equivalence ratio of and a pressure of 15 atm with an initial water mole fraction of 0.05 [20] Axial injector experiments conducted with the in diameter nozzle, at an equivalence ratio of 0.375, and at a pressure of 15 atm Axial injector experiments conducted with the in diameter nozzle, at an equivalence ratio of 0.750, and at a pressure of 15 atm Axial injector experiments conducted with the in diameter nozzle, at an equivalence ratio of 0.375, and at a pressure of 10 atm Axial injector experiments conducted with the in diameter nozzle, at an equivalence ratio of 0.750, and at a pressure of 10 atm Axial injector experiments conducted with the in diameter nozzle, at an equivalence ratio of 0.375, and at a pressure of 15 atm Axial injector experiments conducted with the in diameter nozzle, at an equivalence ratio of 0.750, and at a pressure of 15 atm Axial injector experiments conducted with the in diameter nozzle, at an equivalence ratio of 0.375, and at a pressure of 10 atm Axial injector experiments conducted with the in diameter nozzle, at an equivalence ratio of 0.750, and at a pressure of 10 atm Results from Table 4.1 plotted. Results are from the application of criteria to the experiments at the eight nominal conditions including all combinations of two equivalence ratios, and 0.750, two pressures, 10 and 15 atm, and two nozzle diameters, and in, which give residence times of approximately 130 and 210 ms viii

9 4.12 Axial injector experiments conducted with the in diameter nozzle, at an overall equivalence ratio of 0.5, and at a pressure of 15 atm. The fuel was approximately 75% hydrogen and 25% carbon monoxide by mole Axial injector experiments conducted with the in diameter nozzle, at an overall equivalence ratio of 1, and at a pressure of 15 atm. The fuel was approximately 75% hydrogen and 25% by mole Note that the equivalence ratio listed here is based only on the hydrogen content only, refer to Table 4.2 for the overall equivalence ratio Radial injector and axial injector data taken at an equivalence ratio of 0.75 and a pressure of 15 atm with the in diameter nozzle. The no ignition points are open symbols and the ignition points are closed symbols. The no ignition points are plotted against the maximum residence time of the tube. The black arrows represent the expected behavior. The dashed line shows the progression of the experiment when conducted with the radial injector Comparison of current data with homogeneous reactor simulations using O Conaire [20] and Burke [26] mechanisms Calculated activation energy for hydrogen air system using the axial injector Comparison of current experimental data to other flow reactor data and two homogeneous reactor simulations using the Burke et al. and O Conaire et al. reaction mechanisms [20, 26] ix

10 List of Tables 2.1 List of thermocouples and pressure transducers on high speed data acquisition system. A and B refer to the two high speed A to D boards of the high speed DAQ system Test section thermocouple axial locations from injection point for the axial injector. Pressure measurements were made at the same locations as T2, T5, and T Table of available nozzles with corresponding residence times for air at 800 K List of available orifices with calibration Preburner flow rate adjustments for different nozzles and nominal pressures. The in diameter nozzle at 15 atm was the baseline ratio. The other three cases have 10% and 20% mass flow rate increases of hydrogen and oxygen Equilibrium calculations showing constant pressure adiabatic flame temperatures for a pressure of 15 atm Summary of valve timing. Values not given for long duration are same as short duration. A full test duration corresponds to 50 s. A duration of 60.1 s corresponds to the full duration Temperature drop on Mixed Fuel measured at 44 seconds to steady state value Results from application of criteria to the experiments at the eight nominal conditions including all combinations of two equivalence ratios, and 0.750, two pressures, 10 and 15 atm, and two nozzle diameters, and in, which give residence times of approximately 130 and 210 ms Results from aluminum bottle CO experiment series B.1 Curve-fit coefficients for thermodynamic properties [19] B.2 Constant c P values for Argon and Helium at 1 atm and 300 K [28]. 93 x

11 Acknowledgments I would like to thank my advisor Dr. Robert Santoro for the opportunity to pursue my graduate education under his guidance. I would especially like to thank him for his encouragement and support in writing and reviewing this thesis. Dr. Sibtosh Pal helped immensely in designing the experiment and evaluating problems. Dr. Roger Woodward was invaluable in providing data acquisition software and experimental support. I would also like to thank Mr. Larry Schaff for setting up and running the experiment. I greatly appreciate Dr. Thomas Litzinger s time and effort in reviewing this thesis. I am truly honored to work with all of these individuals. Finally, I would like to thank my parents for their love and encouragement. xi

12 Chapter 1 Introduction 1.1 Motivation The world s demand for energy and the ability to meet this demand with progressively cleaner and more efficient technologies has been an engineering and scientific problem since the realization that there are significant environmental and economic implications that accompany energy production. Further complicating the prob- Figure 1.1. Diagram of a IGCC system [1].

13 2 lem is the desire of every country to be energy independent. The emissions and pollutants produced from various energy sources, mainly hydrocarbon fuels, can have serious environmental impacts. These pollutants include oxides of nitrogen (NO x ), oxides of sulfur (SO x ), volatile organic compounds (VOC), and particulate matter (PM). In addition, the two major combustion products of hydrocarbon fuels, carbon dioxide (CO 2 ) and water (H 2 O), are both green house gases (GHG) and are likely to have a detrimental effect currently or in the future on the climate through increasing the global temperature. Therefore, looking at possible ways to decrease GHG emissions continues to be a priority. Efficiency increases are one clear way to decrease CO 2 emissions because less fuel is used. Other technologies such as carbon capture and storage or sequestration (CCS) have been identified as a means to reduce total CO 2 output to the atmosphere. Coal-fired power plants in 2011 produced approximately 46% of the electricity used in the United States, but contributed 79% of the energy related carbon dioxide emissions [2]. The reason that coal contributes such a large portion of the carbon dioxide emissions is that the hydrogen to carbon ratio of coal is very low. As a result, much of the energy released from burning coal comes from the oxidation of carbon rather than hydrogen, i.e. the amount of CO 2 produced per unit energy is high compared to other fuels such as natural gas. Another major contributor of CO 2 emissions is from transportation where the 2010 AER reports 28 percent of delivered energy consumption is from the transportation sector [3]. Reduction of CO 2 emissions in this category would mainly involve switching to a different energy source. Coal gasification is a technology that is not new but has gained interest due to the fact that it could reduce CO 2 emissions produced from coal combustion and also provide the potential for making alternative transportation fuels. Coal gasification provides the opportunity to implement CCS more effectively on either the precombustion or postcombustion streams. Gasification is a process where coal is partially burned producing a product stream called Syngas with high hydrogen (H 2 ) and carbon monoxide (CO) concentrations. In the case of electric power generation, the syngas can be used in a gas turbine to produce electricity. The diagram above from the NETL website, Figure 1.1, gives an illustration of a

14 3 Integrated Gasification Combined cycle plant (IGCC). Secondary processes such as the water gas shift reaction basically can convert carbon monoxide (CO) to H 2 and CO 2. If the process is used to shift CO to CO 2 then a concentrated CO 2 stream can be separated and sequestered, leaving only H 2. Alternatively, the Fisher Tropsch process can be used to create liquid fuels from syngas. These are two examples that have been proposed for creating alternative fuels for transportation. These alternate fuels pose new problems due to the variety of compositions and therefore combustion characteristics. Autoignition delay time in particular is an important design criterion for lean premixed gas turbines. Land based gas turbines use lean premixed combustion to reduce NO x emissions [4]. The potential of catastrophic autoignition is possible when fuel is mixed with the air or oxygen that has been heated and compressed by the turbine s compressor. Conducting experiments and developing models to predict autoignition delay times for these applications is required if gasification is to be utilized. The apparatus most similar to a gas turbine premixing section is a turbulent flow reactor. The majority of the data on hydrogen and syngas type fuels, that is fuels containing a mixture of CO and H 2, have been collected using shock tubes at temperatures that are generally higher than typical gas turbine premixing sections. There is very limited turbulent flow reactor data available. In fact there have been only three studies of hydrogen or high hydrogen content fuels in turbulent flow reactors [4, 5, 6]. The data from these experiments has been compared to chemical kinetics model predictions and an alarming discrepancy exists. The model s chemical kinetics mechanism is largely based on shocktube data. Moreover, the model validation is also at a higher temperature range than is applicable to gas turbines. The works of Petersen et. al. [7] have recently attempted to confirm the discrepancy of the flow reactor data by conducting shock tube experiments at these lower temperatures. Also, Beerer and McDonell [5] conducted a separate flow reactor experiment for comparison.

15 4 1.2 Literature Review The primary focus of this review will be on presenting the relevant papers that led to the specific set of experiments that were conducted. First, a general overview of autoignition phenomena will be presented. Second, a brief description of the shock tubes and flow reactors will be presented. Last, the importance of the present work will presented with reference to previous work Autoignition and Ignition Delay Time The general definition of autoignition is ignition in the absence of an external ignition source. Ignition is defined as the event that is described either as an explosion or an event of very fast chemistry that releases energy in the form of heat and radiation. The classical concept of explosion limits therefore becomes applicable. An explosion limit essentially means that there are limits of initial conditions (i.e. pressure, temperature, concentration) that will result in an explosion after some induction period and all states in between these limits will not result in an explosion. The elapsed time between the initial state of the reactants and the explosion is defined as the ignition delay time or autoignition delay time. The processes that occur during the autoignition delay time involve oxidation reactions where fuel and oxidizer break down or dissociate into intermediate species and eventually enter a chain branching regime where an explosion will occur. This induction period could also involve processes such as mixing, pyrolysis, evaporation, and thermal chain branching (see Chapters 2, 3, and 7 of Glassman [8]). The chemical kinetic portion of the autoignition delay time is dependent on the rate of the reactions which are dependent on the local conditions. The primary conditions that affect the rate of a reaction are the species concentration (i.e. pressure and equivalence ratio), temperature, and external effects such as radiation and the presence of a catalyst or inhibitor [8]. Because the rates of the reactions are functions of pressure and temperature the thermodynamic process that occurs during the induction time can have a significant effect on the autoignition delay time. This is also important when making assumptions (e.g. constant pressure and enthalpy) in chemical kinetics models for comparison. There are several devices that have been developed to measure autoignition

16 5 delay times including: flow reactors, shock tubes, rapid compression machines, and constant volume bombs. There are a variety of devices for measuring autoignition delay because they were developed to replicate a certain practical device or a specific thermodynamic process. For example, a turbulent flow reactor is most like the premixing section of a lean premixed gas turbine. The thermodynamic process for this application is essentially constant pressure/constant enthalpy. Shock tubes typically are most like a constant volume/constant energy process Shock Tubes A shock tube contains two sections separated by a diaphragm or burst disc. The first section is called the driver section and is pressurized. The second section is the driven section and contains the fuel oxidizer mixture. When the diaphragm is ruptured a shock wave forms and consequently propagates through the driven section where is eventually reflects off the end of the tube leaving a stagnant heated mixture. The time between when the waves reflects off the end wall and an explosion event occurs is the autoignition time. Typically pressure is used to identify ignition but many methods for detecting ignition are used (see [9] for a list induction time definitions). Shock tubes can produce very high temperatures nearly instantaneously. These high temperature produce very short autoignition delay times (on the order of micro seconds). The benefit of the short delay time and near instantaneous increase in temperature is that the experiments are highly ideal. There are little fluid mechanic effects and no mixing component in the ignition delay time Flow Reactors Flow reactors are essentially continuous flow devices where fuel and oxidizer mix rapidly in a tube and then flow along the tube, eventually igniting if the conditions needed for autoignition are present. Thus, flow reactors have a finite mixing time and a finite chemical time that contribute to the time it takes for autoignition to occur. τ total = τ mixing + τ chemical

17 6 Since autoignition is a chemically controlled phenomenon, it is important that mixing time be short as physically possible. Thus a critical characteristic of all flow reactors is mixing. The benefit of a turbulent flow reactor is that mixing is fast if the Reynolds number is large as is typical in a turbulent flow reactor. The questions then for any turbulent flow reactor are does mixing have a significant impact on the measured autoignition time and does the measured autoignition time represent accurately the autoignition times that will be present in practical devices such as gas turbines. The mixing time is not only dependent on the geometry of the apparatus, but also the operating conditions and flow rates of the flow reactor. In other words, autoignition in turbulent flow reactors may be apparatus and measurement dependent. The design of a flow reactor should therefore provide rapid mixing and uniform flow conditions. Consequently in turbulent flow reactors, autoignition time has typically been measured by calculating the average velocity in the flow reactor tube and then dividing the length by that average velocity. Two methods of conducting experiments have been used and will be described below Other Flow Reactor Facilities The first reactor used to study syngas was designed and built in the early 1980 s by Peschke and Spadaccini [4]. The first methodology was used by Peschke and Spadaccini and involved increasing the oxidizer (air) temperature while continuously flowing air and fuel into the flow reactor until ignition occurred, which was measured by observing a temperature or pressure rise using a thermocouple or pressure transducer respectively [4]. The ignition delay time that was then reported was just the residence time of the reactor, that is the length of the flow reactor divided by the bulk velocity. The reactor used water cooled walls to eliminate ignition in the boundary layer. The reported temperature was calculated based on the measured fuel and air temperatures. The injector used in these experiments was a multiple element converging diverging injector. This injector had 19 elements, each with fuel injection near the throat [4]. The second flow reactor used by Beerer and McDonell [5] used a second methodology. Instead of ramping the temperature while continuously flowing fuel, the air temperature and flow reactor conditions were established and then fuel was in-

18 7 jected. The experiment can thus result in no ignition or an ignition event. A series of experiments are then done at different air temperatures to obtain the temperature, measured by a thermocouple, that corresponds to ignition at the residence time of the reactor. Using the same apparatus the ignition delay time was measured with a second method that measured the elapsed time between when the fuel valve opened and when ignition occurred as measured with a PMT looking axially down the flow reactor. Beerer and McDonell state that there is as much as a 50% difference between the two measurements of ignition delay. The injector used in these experiments was also a converging diverging injector similar to Peschke and Spadaccini [4], however, it was composed of only a single element design. The multiple element converging diverging injector of Peschke and Spadaccini [4] is the basis of one the injectors used in the current experiments. The difference is that there are 7 elements instead 19. The injector for current experiments is described later as the radial injector because it injects fuel radially at the throat. The design of the radial injector was based on the one reported by Peschke and Spadaccini [10]. The injector had multiple injector elements that incorporated converging-diverging venturis in which fuel was injected. This injector has been proven to have excellent mixing characteristics. In fact, this has been shown by Spadaccini and Santoro (current radial injector) for a multiple element converging-diverging injector and by McDonell for a single element convergingdiverging injector [6, 10, 11]. The same basic design of injector was used in all these studies. The measurements made of autoignition for high hydrogen content fuels conducted using this injector design have had significant discrepancies from homogeneous one dimensional model predictions [5, 7]. Additionally, Petersen conducted shock tube experiments at lower temperatures attempting to narrow the temperature gap between flow reactor and shock tube experimental data [7]. Figure 1.2 is a plot showing flow reactor data from Peschke and Spadaccini [10] (called UTRC data (1985) on the figure) and from Beerer and McDonell [5], shock tube data from Petersen [7], and various models. When initially published the shock tube data seemed to confirm the flow reactor data because the shock tube data shows an increasing deviation from the model as temperature decreases. Late, papers have been published stating that the disparities in observations and kinetic pre-

19 8 Figure 1.2. Comparison of flow reactor data and shock tube data to chemical kinetics model [7]. Flow reactor data is adjusted to 20 atm. dictions were a result of the ideal modeling assumptions applied and their inability to represent experimental conditions appropriately [12]. Essentially, when conducting shock tube experiments with long residence times and non-dilute mixtures the ideal behavior of a shock tube is lost. The discrepancy between the turbulent flow reactor data and the chemical kinetics model prediction, however, still exists. The conditions used in the present study, which were chosen to simulate the temperatures and pressures found in lean premixed gas turbines, were at relatively low temperatures, T < 1000 K, and high pressures, 10 < p < 30 atm [13]. This regime had proven difficult to obtain experimentally. Flow reactors either have not achieved high enough temperatures or lack the length for long residence times associated with lower temperatures. Thus, there has been a lack of overlap of conditions for direct comparison. Furthermore, it has been noted by Dryer and Chaos [12, 14, 15, 13] that this specific regime of the explosion limit curve, mild ignition, is defined by the region above the extended second limit where ignition

20 9 proceeds by thermal chain ignition. Dryer [13] describes possible non-ideal behavior can be due to the presence of contaminants in the reactants or on experimental surfaces, compressible fluid dynamic effects, inhomogeneous mixing, and catalysis from particles on surface materials. 1.3 Experiment The purpose of this work was to investigate autoignition behavior of hydrogen and syngas in a turbulent flow reactor. The results of previous flow reactor studies on syngas fuels have not compared well to predictions using established chemical kinetics models. The results of Beerer and McDonell and Peschke and Spadaccini show that autoignition of syngas or hydrogen has a small temperature dependence at the conditions studied compared to the predictions [5, 4]. Aside from deviation from chemical kinetics models, experimentalists noted that when studying syngas fuels in flow reactors that for the same conditions quite different autoignition times are measured [5, 6]. Since autoignition in flow reactors is not only coupled to the chemical kinetics but also the thermal and fluid characteristics of the apparatus there are many potential sources of error in performing or interpreting autoignition experiments. The major problems that are commonly discussed with flow reactors are recirculation zones, mixing, wall effects, boundary layer effects, and non uniform flow properties. The main objectives for the present work were: Conduct experiments at higher temperatures to address lack of overlap in experimental data between shock tubes and flow reactors in the below 1000 K temperature regime. Investigate the potential fluid mechanic effects of converging diverging injectors. Consider other possible problems such as catalytic and boundary layer effects.

21 Chapter 2 Experimental Apparatus and Procedure 2.1 Components The autoignition experiments were performed in a flow reactor consisting of an instrumented modular tube, injector, and nozzle. In addition, supplementary components included a preburner, electric clamshell heater and an electric air heater. Compressed air was provided from two large blow down tanks. The tanks were filled to a pressure of 700 psia. Hydrogen, oxygen, argon, nitrogen, and carbon monoxide were supplied from standard gas bottles of commercial purity, at least 99% pure. Ultra high purity, 99.9% pure, carbon monoxide was also used. All gases were metered using calibrated critical orifices. Data acquisition was done with LABVIEW, while valve control was done with a MODICON sequencing system. The following is a description of the components of the flow reactor. The process diagram is shown diagrammatically in Figure 2.1. In order to distinguish between the fuels used in the preburner and the flow reactor, the following nomenclature is adopted. The preburner hydrogen and the preburner oxygen used in the preburner will be referred to as the propellants. The fuel used for the experiment, hydrogen, will simply be referred to as fuel.

22 11 Air main O2 air addition Preburner H2 Preburner O2 Preburner Ar Fuel H2 Fuel CO water P,T P,T P,T P,T P,T P,T P,T P P,T P,T Air Heater P,T T Air Addition P,T P,T P,T Flow Reactor vent Nitrogen Pressurized Tank P,T Drain Figure 2.1. Test apparatus schematic. P T Filter Pneumatic Valve Dome Loaded Regulator Metering Orifice Check Valve Solenoid Valve Pressure Sensor Thermocouple Exhaust

23 12 The flow reactor was designed to allow control of the primary variables important in autoignition delay: pressure, temperature, equivalence ratio, and residence time. The temperature was varied until ignition occurred at the selected residence time. Residence time was controlled by a choked nozzle at the end of the flow reactor tube, pressure was controlled by the total mass flow rate, air temperature was controlled by the electric air heater, the fuel temperature was controlled by the ratio of preburner propellants and diluent to fuel; and the equivalence ratio was determined by the ratio of fuel to oxidizer. A more detailed relationship between the variables and the residence time was derived assuming isentropic compressible flow and a discussion of the detailed relationship can be found in Section Flow Reactor Test Section The flow reactor test section, Figure 2.2, consisted of six modular flanged sections that were externally heated by a clamshell heater. Smooth wall grade 310 stainless steel with a in (45.19 mm) inner diameter was used for the entire test section. Small holes in the wall, 0.03 in (0.76 mm) diameter, allowed access for temperature and pressure measurements. A tube was welded to the reactor around these holes to allow access to the reactor outside the clamshell heater (Section 2.1.8). At the end of the tube, a three way tee fitting was attached. Along the axis of the tube thermocouples were inserted to a distance of approximately 0.1 in (2.54 mm) into the tube and held in place by a graphite ferrules. A pressure transducer was connected to the other end of the three way tee fitting. The thermocouples were sheathed, ungrounded, K type, with a in (0.508 mm) outer sheath diameter, and were manufactured by Omega (part number hkmqxl-020u-12). The thermocouples in the test section were used to determine the location of autoignition and were numbered from one to thirteen. The main injector was preceded by thermocouple T1 and was followed by thermocouples T2 through T13 at specified axial distances from the fuel injection location (see Table 2.2). The last 26 in (660 mm) of the reactor have the highest density of thermocouples (T6-T13) and allowed for higher accuracy in determining the ignition location. Pressure transducers were located at the same positions as thermocouples 1, 2, 5, and 13. The pressure transducers were Setra gauge pressure transducers

24 13 with zero to five volt output and typically had a range of 0 to 3000 psig (20684 kpa). Pressure transducers with a range of 0 to 1500 psig were used upstream of the carbon monoxide metering orifice used in syngas studies and for the reactor pressure at location P13. Ranges of 0 to 1000 psig were used for reactor pressures located at P1 and P2 and also for the measurement of the air pressure after the critical orifices. The pressure upstream of the air critical orifices was taken using a 0 to 500 psig range pressure transducer. The accuracy for these pressure transducers was ±0.11% of full scale. The majority of the thermocouples in the test section and two of the pressure transducers, 2 and 5, were recorded by the high speed data acquisition system at a data rate of 5 khz. Table 2.1 lists which channel of the A to D converter was used to acquire the measurements of the thermocouples and pressure transducers. A separate data acquisition system recorded temperature at 10 Hz and pressure at 200 Hz for all thermocouple and pressure transducers in the system. The distance of the thermocouples from the fuel injection location of the main injector are tabulated in Table 2.2. The distances in Table 2.2 are for the axial injector with the addition of a thermocouple that was added to provide a centerline temperature measurement. This centerline thermocouple was referred to as Tcl and is located between the T4 and T5 thermocouples. The same type and diameter thermocouple was used for the centerline measurement however, a tube was used to stiffen the thermocouple to eliminate movement in the high flow environment leaving only approximately 0.1 in (2.54 mm) of the thermocouple fully exposed near the centerline of the flow reactor. Thus, the response time of the thermocouple was maintained. A second injector, referred to as the radial injector, increases the distances in Table 2.2 by 2 in (50.8 mm). These distances were for a cold tube. The actual distances were slightly different due to thermal expansion of the tube that was approximately 1 in (25.4 mm) when the reactor was heated to test conditions. The analysis conducted used the cold distances Main Injector The injector is a critical component in flow reactor autoignition experiments. There are two factors that make up the measured autoignition time: mixing time and

25 14 Table 2.1. List of thermocouples and pressure transducers on high speed data acquisition system. A and B refer to the two high speed A to D boards of the high speed DAQ system. Date of Change A0 A1 A2 A3 A4 A5 A6 A7 B0 B1 B2 B3 B4 B5 B6 B7 Original Setup PB Outlet T2 T2 T3 Mixed T5 P2 P5 valve Fuel T3 signal 11/21/2011 T1 11/22/2011 T5 12/16/2011 Tcl 1/11/2012 PB Outlet T2 = CO T2 T6 T7 T8 T9 T10 T11 T12 T13 Main Injector Nozzle Figure 2.2. Test section model showing modular sections, main injector, and nozzle. chemical time. The chemical time is of primary interest in the present work. It is beneficial to minimize the mixing time because it can be apparatus dependent and can affect the autoignition results if mixing is not achieved rapidly. Temperature and equivalence ratio uniformity are both important since both affect the autoignition time measured. Essentially if there is poor mixing, there is a range of temperatures and equivalence ratios that exist and the condition of autoignition are unknown. The autoignition times from the present experiments are compared with homogeneous chemical kinetic model predictions of chemical time for autoignition to occur and therefore, assume perfect mixing. In the present studies, two distinctly different injectors were fabricated and are described below. The first is a converging diverging injector with three radial injection sites in each of the seven converging sections. The second is a distributed axial injector with seven injection tubes.

26 15 Table 2.2. Test section thermocouple axial locations from injection point for the axial injector. Pressure measurements were made at the same locations as T2, T5, and T13. Thermocouple Location (in) (m) T T T Tcl T T T T T T T T T Radial Injector The radial injector design is based on a fast mixing converging-diverging nozzle used by Spadaccini in flow reactor experiments [4, 10, 16]. This injector, shown in Figure 2.3, is made in two pieces. The converging section (right) has three grooves in each nozzle section that make up the fuel injection location when mated to the diverging section (left). The back of the diverging section that mates with the converging piece has a manifold to distribute fuel to each of the fuel injection locations as seen in Figure 2.4. The other side of the diverging section that looks spiked was designed to prevent recirculation zones in the flow path by creating a smooth expansion from the throat of the injector to the inner diameter of the reactor. Multiple converging pieces, numbered one to seven, were made to allow varying momentum flux ratios between the air and fuel by changing the area of the radial injection locations. The number five injector plate was sufficient for all experiments conducted in the current study and has inlets each in (0.397 mm) by in (2.381 mm). Each of the seven converging diverging injection elements has 3 rectangular inlets for a total of 21. The injector s mixing characteristics were experimentally examined with an Acetone PLIF study that is described by Santoro

27 16 Figure 2.3. Pictures of radial injector. Figure 2.4. Radial injector model with cross section view. Air flows from left to right. [6]. The study found that the injector had good mixing within 2.0 in (50.8 mm) of the injector exit, if the momentum flux ratio between the fuel and air stream was above 20. The different injection size holes allow this momentum flux ratio to be maintained for a range of flow rates. Pressure drop across the injector is also important and is essentially a measure of the velocity of the gas that exits the injector. If the pressure becomes too large the injector could be choked and lead to the upstream metering orifices becoming unchoked, in which case the flow rate would be unknown.

28 Axial Injector The axial injector, shown in Figure 2.5, has an internal fuel manifold which distributed fuel to seven 1.25 in (31.8 mm) long, 0.08 in (2.0 mm) inner diameter tubes, and 0.02 in (0.5 mm) wall thickness. The injector was machined in two pieces. The seven tubes were brazed to one of the pieces and were then brazed together into a single piece (Figure 2.5). Mixing occurs with this injector through turbulence created by the high velocity fuel jet entraining the slower moving air coflow. Seven distributed jets were used to increase mixing. Coflow jets or injectors have been extensively examined in the literature. Based on similarity analysis and empirical data, correlations have been developed for the entrainment length of coflow jets. Equation 2.1, from Ricou and Spalding, was solved to give an approximate length that a coflow jet needs to become fully entrained [17]. m m o = 0.32 x d o ( ρ1 ρ o ) 1 2 (2.1) where x is the entrainment length, m o is the jet mass flow rate, and m is the total mass flow rate. The total area of the axial injector jets was chosen to be similar to that of total area of the radial injector jets. With this area, the calculated entrainment length is significantly less than 2 in (50.8 mm). Based on this calculation and the fact that there are seven jets the axial injector should provide good mixing. The results are later compared with autoignition results using the radial injector under the same conditions. This is further discussed in the results, Chapter 4. Time constraints limited validation of the mixing performance of the axial injector design through direct experimental or computational methods. However, it will be shown that under suitable flow conditions, the radial and axial injectors have similar autoignition times. This observation provides an indirect validation of the mixing performance of the axial injector Heated Air System A two stage compressor supplied high pressure air to the two large air tanks. Typically the air pressure in the tanks was between 500 and 700 psi (3.4 and 4.8

29 18 Figure 2.5. Model of axial injector. MPa) during testing. The tanks were always maintained at a pressure above that required for choking of the critical orifices, which varies based on the orifice size and test condition. The ambient air entering the apparatus was measured with a thermocouple just upstream of the air metering venturi. The approximate water concentration of the air was on the order of 100 ppm (determined from the saturation pressure of water at ambient temperature). The air in the tanks was known to be saturated due to the accumulation of water in the bottom of the tanks. The air was filtered with a Hankinson water and particulate filter located inside the test cell. The flow rate was metered prior to entering the electric heater using critical orifices and a dome loaded regulator. The electric heater was rated to heat 1 lbm/sec (0.45 kg/sec) of air to 1100 F (866 K) and can withstand pressures up to 500 psi (3.4 MPa). The heated air was then routed to the autoignition tube with insulated flexible tubing. The electric heater control unit uses temperature measurements at the outlet of the heater as the main control for achieving a constant operating temperature. There can be a relatively small temperature difference between the set temperature and the reactor inlet temperature because of heat loss through the approximately six feet of insulated flexible tubing. Typically, the temperature drop was less than 20 F (11 K), but varied based on the flow rate and temperature. The temperature at the inlet of the reactor was measured and therefore the degree of heat loss can be determined.

30 Preburner and Fuel Heating System The preburner was a cylindrical chamber that burns hydrogen and oxygen to provide energy to heat the fuel that was used in the autoignition experiment. The purpose of heating the fuel was to achieve an overall mixture temperature between 800 and 900 K. Heating of the fuel eliminated the temperature drop that occurred when mixing room temperature fuel with the heated air. This temperature drop was particularly large for hydrogen, which has a large specific heat compared to hydrocarbon fuels. The preburner products mixed with the fuel in a cylindrical mixing section prior to being injected through the main injector. The preburner, as shown in Figure 2.6, consisted of modular sections and was originally designed for use in rocket experiments. It can be split into three functional sections: injector section, combustion section, and dilution section. Two additional sections referred to as the coupling adapter and mixing section were used to introduce the fuel and mix the fuel with the hot preburner products. An adapter at the end of the mixing section constricted the gases to a in (9.53 mm) outer diameter stainless steel tube that was connected to the main flow reactor injector. The preburner and mixing section had circular cross sections of 1 in (25.4 mm) and 1.6 in (40.6 mm) respectively. The mixing section was 10 in (254 mm) in length allowing enough volume for thorough mixing of the fuel and combusted propellants. The preburner was a heat sink design made from Inconel with a wall thickness of 1.5 in (38.1 mm). The coupling adapter and mixing section were made of stainless steel. The preburner outer wall temperature was heated with resistance bar heaters and electric heating tape to approximately 300 F (422 K) before each experiment. Aside from reducing the overall heat loss, this also improved consistency in igniting the preburner propellants and in the temperature of the preburner products. Heat loss from the tube connecting the mixing section to the main injector was minimized with insulation on the portion that were not in the clamshell heater. The propellants used for the preburner were injected through an impinging type injector. Gaseous oxygen was injected through a single center post and hydrogen was introduced through five angled pathways that impinge on the oxygen flow. These flows were ignited in the combustion section by the products of a hydrogen oxygen torch igniter. The torch igniter was a separate chamber that connected to

31 20 the combustion section of the preburner. A spark plug in the torch igniter chamber ignited a fuel rich mixture of hydrogen and oxygen (oxygen to fuel ratio by mass of approximately 5.7). These gases then flowed into the preburner where they ignited the stoichiometric propellant mixture. The igniter flows were ceased and the main preburner flows continued to burn leaving water as the sole combustion product. This product of combustion was diluted with argon in the dilution section to reduce the combustion temperature of the stoichiometric hydrogen/oxygen mixture, which had an adiabatic flame temperature of approximately 3400K. This high temperature was of concern due to possible melting of the hardware. Specifically, the tube connecting the preburner to the main flow reactor injector was of concern because it was made from stainless steel with an inner diameter of 0.2 in (5.0 mm) and wall thickness of 0.08 in (2.0 mm). The main injector could also be damaged from the hot gases. The connecting tube and main injector were also heated by the clamshell heater (Section2.1.8); this further increased the potential for melting. Dilution to 2200 F ( 1478 K), measured near the wall of the mixing section was selected as the operating condition. This temperature was measured when only the propellant and dilution were flowing since the preburner was allowed to run for three seconds before fuel was injected into the mixing section. The measured heated fuel temperatures were typically between 1030 F and 1400 F (830 and 1030 K) depending on the fuel flow rate. Argon was chosen instead of nitrogen as the diluent due to the potential formation of thermal NO x, which was known to catalyze autoignition [18]. The argon was introduced through several holes spaced axially and at multiple angular positions in the dilution section. The diluted products from the preburner, water and argon, expanded into the 90 degree coupling adapter that was connected to a mixing section as seen in Figure 2.7. The main fuel was injected through an orifice at the end of the 90 degree bend along the axis of the mixing chamber. The heated fuel was then injected through the main injector in the flow reactor.

32 21 Figure 2.6. Cross section model of preburner with labels. Igniter chamber not shown Valves Pneumatic and electrically activated solenoid valves were used in this experiment. Control of all gases other than the fuel was controlled using a MODICON system, which controls all pneumatic valves. High speed solenoid valves were used to provide fast opening times for fuel injection and were controlled through a LABVIEW program. The pneumatic valves had an opening time on the order of 100 msec and the high speed solenoid valves had an opening time of approximately 1 msec. The high speed valves had an orifice diameter of 0.04 in (10.16 mm), which was too small to accommodate the fuel flow rate. Therefore, valves mounted in parallel were used to achieve the desired flow rates. Fuel vent solenoid valves were also used that opened prior to the opening of the valves to the fuel inlet. Simultaneously, the fuel vent valves closed and the fuel valves to the fuel inlet opened. The purpose of the vent valves was to cause the pressure just upstream of the fuel inlet to increase to the steady value. Figure 2.8 shows the pressure trace downstream of the metering orifice upstream of the high speed solenoid valves. The pressure initially read the dome loader regulated pressure. When the fuel vent opened the pressure decreased to a low value because the vents opened to essentially atmospheric pressure. The pressure rose when the valves switched because the back pressure from the preburner was above atmospheric. After the initial pressure transient, fuel flowed through the fuel inlet at a steady rate and the pressure equilibrated to

33 22 Figure 2.7. Preburner and fuel mixing chamber. an intermediate pressure that was equal to the preburner chamber pressure plus the pressure drop due to the fuel inlet, valves, and pipe friction losses. The valve timing scheme described above was critical in preventing the initial flow rate of fuel from being higher than the required value. With this valve timing scheme, the initial flow rate was lower and thus the initial transient gases should exit flow reactor before ignition occurs when the steady flow gases is achieved, which equated to the desired equivalence ratio. This initial higher flow rate of the fuel would be a result of the high pressure established between high speed valve and the upstream critical orifice if the fuel vent was closed. By having the fuel vent open first prior to letting the fuel flow through the high speed valve, the pressure between the high speed valve would increase from ambient to the steady state value. In this approach the initial fuel flow was always smaller than the desired flow rate. The summarized valve timings can be found in Section

34 x 106 Fuel Vent Opens pressure (Pa) Valve Closes Due to Autoignition Abort Fuel Pressure Transient Simultaneously Fuel Vent Closes and Fuel Injection Valve Opens time (sec) Figure 2.8. Characteristic fuel pressure trace taken downstream of orifice Nozzle The nozzle, Figure 2.9, was located at the end of the flow reactor test section. The purpose of the nozzle was to control the residence time of the flow reactor. Based on compressible flow theory the Mach number is determined by the area ratio between the constant diameter reactor and the area of the nozzle throat. Since the area ratio was fixed and the temperature only varied by a small amount, the velocity in the tube was constant. The nozzle throat was choked for all conditions considered, since the test pressures were all above 10 atm and the reactor exhausts to the atmosphere (the pressure drop was much greater than a factor of two). The specific relationship from which the flow reactor bulk velocity and residence times were calculated will be presented in Section 2.2. The nozzle was water cooled and also injects water through an annular slit at the throat to prevent damage to the nozzle by quenching the combustion gases. In table 2.3, a list of the currently available nozzles is presented and corresponding

35 24 Figure 2.9. Cross section view of nozzle showing water passages. Table 2.3. Table of available nozzles with corresponding residence times for air at 800 K. Nozzle Diameter (in) Residence time (ms) residence times for air at 800 K (with a tube length of 84.3 in (2.14 m)) Critical Orifices Mass flow rates were controlled using critical orifices. Table 2.4 lists the orifices used with their corresponding discharge coefficient (C D ). The discharge coefficient takes into account non ideal fluid mechanic effects that cause a different flow rate than would be calculated using the actual area of the orifice. Sharp-edged and Venturi type orifices were used. The Venturi type critical orifice had downstream pressure recovery of approximately 80% of the upstream stagnation pressure. This excellent pressure recovery was due to the converging-diverging geometry of the orifice. A sharp-edged orifice type was essentially a hole in a plate that required the downstream pressure be no greater than approximately 50% of the upstream

36 25 Table 2.4. List of available orifices with calibration. Diameter (in) C D (ms) Venturi Type Orifice Type pressure to choke the orifice. The calibration of the orifices had been previously conducted. The C D value was calibrated based on the designed diameter of the orifice and therefore C D can be greater than one Clamshell Heater A 10 kw Split Tube Furnace (Series 3210), referred to as the clamshell heater and manufactured by Applied Test Systems, was used to maintain the outer wall of the flow reactor at a constant temperature. The primary purpose of the clamshell heater was to minimize the axial temperature gradient by creating an adiabatic wall condition. The wall temperature was set to the same temperature as the air heater temperature. The fuel temperature usually was hotter than the air

37 26 temperature and therefore the mixture temperature was typically between 10 to 50 K hotter than the wall temperature. The clamshell heater had a control unit with six zones of control. The input to the control unit was a thermocouple that was held against the outside of the reactor tube by a spring mechanism. A small nitrogen purge was used to purge the gap between the clamshell heating elements and the tube to prevent any buildup of combustible gases from possible leaks. The clamshell heaters element temperature limits how hot it can maintain the wall of the flow reactor. For typical flow conditions the maximum wall temperature achievable was approximately 1100 F (866 K). 2.2 Experimental Variables The following is a discussion on how the experimental conditions were achieved and controlled using the apparatus previously discussed with respect to the test conditions considered in the experiment. Chapter 3 examines how the experiments were analyzed. This section is intended to identify how variables in a parametric study would be setup. The experiments conducted were limited to primarily eight conditions. The data will show small variations in the controlled variables when changing equivalence ratio, pressure, and temperature. The following analysis estimates the percent difference these changes have. The primary autoignition variables were pressure, temperature, equivalence ratio (composition), and residence time. The two main assumptions for this analysis were isentropic flow and ideal gas behavior. The testing conditions for the present studies were limited to primarily, two nozzle diameters, two equivalence ratios, and two pressures. The conclusion of the following analysis was that the temperature can be adjusted using the air heater only while the composition and pressure were held constant. Residence time was primarily controlled by the physical dimensions of the tube and nozzle. The residence time is related to the bulk velocity and the length of the flow reactor tube. The applicable isentropic relationship between the nozzle

38 27 geometry and the residence time is the area ratio equation: A A = ( γ ) (γ+1) 2(γ 1) ( 1 + γ 1 2 M 2) γ+1 2(γ 1) M (2.2) This equation can be solved for the Mach number in the flow reactor. There will be two solutions however; the solution of interest will be the subsonic one. The area ratio between the reactor area, A, and the nozzle throat area, A *, along with the gas property γ, which is the ratio of specific heats c p /c v, determine the Mach number in the tube as long as the throat is choked. The nozzle will be choked if the pressure ratio between the stagnation pressure and the nozzle exit pressure is above approximately two, according to isentropic compressible flow for air. Thus the nozzle was always choked for the high reactor pressures considered in this experiment. The Mach number is related to the velocity, v, by the speed of sound. The speed of sound can be calculated assuming ideal gas from the relation: c = γrt (2.3) R = R U MW mixture (2.4) Where R is the specific gas constant, calculated from Equation 2.4, T is the mixture temperature, R U is the universal gas constant, and MW mixture is the molecular weight of the mixture. The bulk velocity is then calculated from Equation 2.5 and the mean residence time, τ residence, is related to the length of the reactor, L, by Equation 2.6. M = v c (2.5) τ residence = L v (2.6) The overall temperature range that was considered is relatively small, mixture temperature ranged between 800 and 900 K. Specific series of runs had a smaller temperature range that was generally less than 50 K. This leads to less than a 5% change in the speed of sound. Propagating this variation to the residence time leads to less than a 5% change in residence time since the Mach number in the reactor was fixed due to the area ratio between the reactor and nozzle. The ratio

39 28 of specific heats for this mixture of fuel and air does not vary greatly from 1.4 in this temperature range. Also, the ratio of specific heats, gamma, is primarily a function of temperature. However, the temperature range is small and therefore gamma does not vary significantly for fixed composition. Thus, the residence time is primarily dependent on the area of the nozzle throat (nozzle diameter) and the molecular weight of the mixture. The mixture molecular weights varies based on the equivalence ratio. For the two equivalence ratios studied (0.375 and 0.75) the nominal molecular weights were 28 and 24 kg/kmol respectively. This propagates to approximately an 8% change in residence time when varying the equivalence ratio. The two remaining variables are pressure and temperature. The pressure is directly proportional with the mass flow rate in the reactor. Thus, the pressure can be controlled by proportionally increasing or decreasing the flow rates into the reactor. This relationship is evident in Equation 2.7 where a subscript t denotes a stagnation property. ṁ = Ap ( t γ Tt R M 1 + γ 1 2 M 2 ) (γ+1) 2(γ 1) (2.7) Equation 2.7 can be evaluated in the tube or at the throat of the nozzle. Evaluating this equation at the throat of the nozzle, where the Mach number is one, reduces the equation to Equation 2.8. ṁ = Ap t γ Tt R ( γ ) (γ+1) 2(γ 1) (2.8) All variables in this equation except mass flow rate and pressure are constant for a given nozzle diameter, mixture composition, and mixture temperature. Therefore, the pressure for a given nozzle can be changed by multiplying all flow rates by the pressure ratio as related by Equation 2.9. Since the Mach number in the reactor is very small, below 0.1, the stagnation pressure and static pressure are essentially the same. ṁ 1 ṁ 2 = p t1 p t2 (2.9) The mass flow rate is also directly proportional to the area of the nozzle. Therefore,

40 29 the mass flow rate required to achieve the same pressure in a different nozzle can be calculated by multiplying the flow rates by the area ratio as related by Equation ṁ 1 = A 1 = D (2.10) ṁ 2 A 2 D 2 The mass fraction of the fuel is low, between one and two percent of the total mixture. Equivalence ratio can therefore be changed by changing only the mass flow rate of fuel with minimal impact on the overall reactor pressure. The change in molecular weight due to a change in equivalence ratio from to was significant and as stated earlier varied from 28 to 24 kg/kmol respectively. This propagates to a change in pressure of less than 10%. A residence time and pressure have been shown to be adjustable to any combination of composition and temperature under consideration. Also, it has been shown that the equivalence ratio can be changed with minimal impact on the pressure and residence time. The mixture temperature was more difficult to adjust since it is determined not only by setting the temperature on the air heater but by the combustion products of the preburner and associated heat loss in the preburner. Heat transfer is significant in the resulting fuel temperature. The purpose of the current experiments was to conduct experiments at higher temperatures than available with only the air heater. The preburner was used to heat the fuel thus eliminating the temperature drop that is associated with mixing unheated fuel with heated air and also increasing the maximum overall mixture temperature. The mass flow rates that were used were obtained by conducting a first law analysis. A First Law analysis was used to determine how much preburner propellants were required to achieve the desired overall mixture temperature for the test conditions. Recognizing that for any individual experiment that the pressure and residence time are constant, a First Law analysis can be solved for arbitrary flow rates and scaled to the correct pressure. The residence time, as discussed above, is controlled by the diameter of the nozzle for this constant diameter flow reactor tube. Secondly, the temperature of the preburner products will be determined by the ratio of propellants to the diluent since the propellants, hydrogen and oxygen, are at a stoichiometric ratio. The temperature of the mixture of fuel and preburner

41 30 Figure Schematic of the preburner indicating locations of temperatures discussed. products will be determined by the ratio of propellants including dilution with argon to the fuel. The overall mixture temperature of the heated air and heated fuel will be determined by the equivalence ratio once the two ratios above are set and along with the air temperature. Since the fuel flow rate changes, either the flow rate of the propellants would have to change to maintain a constant temperature of the preburner products and fuel mixture, or the propellant flow rates could be constant and the temperature of the preburner products and fuel mixture could vary. The later was chosen because it prevents time consuming adjustments to multiple flow rates. Thus, the temperature of the mixture of fuel and preburner products was not maintained constant, but was allowed to vary based on the fuel flow rate. The preburner products temperature was constant for all conditions except for temperature variations caused from heat loss dependent on the magnitude of the flow rates. The adiabatic temperature of combustion is essentially determined by the equivalence ratio of the propellants and the ratio of diluent to propellants, both of which were fixed. The process of heating the fuel and the specific reasons for adding Ar as a diluent were discussed in Section The most operationally efficient procedure for varying the mixture temperature in the flow reactor was to only change the air temperature. Ignition delay time