The Oxy-Fuel Burner: emissions and stability analysis

UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner Burner Configuration # 3: Oxy-Fuel Burner Emissions and Stability Performance CEC Agreement No. 500-13-004 Prepared by: Gorka Fuentes Lejarza, Andrés Colorado and Vincent McDonell1 Version1: 15 Jul 2015 1 949 824 5950 x11121; [email protected] UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner Table of Contents Executive summary ........................................................................................................................................ i 1 Introduction .......................................................................................................................................... 1 1.1 Motivations to use an oxy-fuel burner ......................................................................................... 1 1.2 Alternative fuels ............................................................................................................................ 2 2 Key features .......................................................................................................................................... 3 2.1 Operating conditions .................................................................................................................... 4 3 Virtual AeroThermal Field - Computational fluid dynamics (CFD) ........................................................ 4 3.1 Introduction .................................................................................................................................. 4 3.2 Geometry ...................................................................................................................................... 5 3.3 Mesh ............................................................................................................................................. 6 3.3.1 Approach ............................................................................................................................... 6 3.3.2 Mesh sensitivity analysis ....................................................................................................... 7 3.4 Models .......................................................................................................................................... 9 3.5 Boundary conditions ................................................................................................................... 10 3.6 Solution Methods ........................................................................................................................ 10 3.7 Convergence criteria ................................................................................................................... 10 4 Chemical reactor network (CRN) ........................................................................................................ 12 4.1 Zonal distribution ........................................................................................................................ 12 4.2 Reactor network sensitivity analysis ........................................................................................... 13 5 Results and discussion ........................................................................................................................ 15 5.1 Fuels type 1: Hydrogen enriched natural gas ............................................................................. 15 5.1.1 CFD results .......................................................................................................................... 15 5.1.2 CRN results .......................................................................................................................... 20 5.2 Fuels type 2: Landfill/digester biogas ......................................................................................... 26 5.2.1 CFD results .......................................................................................................................... 26 5.2.2 CRN results .......................................................................................................................... 32 5.3 Fuels type 3: Mixtures with heavier hydrocarbons .................................................................... 38 5.3.1 CFD results .......................................................................................................................... 38 5.3.2 CRN results .......................................................................................................................... 43 6 Conclusions ......................................................................................................................................... 48 Appendix A: Details of the domain modeled .............................................................................................. 50 Domain dimensions ................................................................................................................................ 50 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner Boundary conditions ............................................................................................................................... 51 Appendix B: Reactor network sensitivity analysis ...................................................................................... 61 Fuel type 1: Hydrogen mixtures .............................................................................................................. 61 Fuel type 2: Landfill/digester biogas ....................................................................................................... 65 Fuel type 3: Mixtures with heavier hydrocarbons .................................................................................. 69 Appendix C: Equilibrium calculations .......................................................................................................... 73 N2 effect .................................................................................................................................................. 73 Fuel type 1: Hydrogen mixtures .............................................................................................................. 74 Fuel type 2: Landfill/digester biogas ....................................................................................................... 76 Fuel type 3: Mixtures with heavier hydrocarbons .................................................................................. 77 Appendix D: Interchangeability analysis ..................................................................................................... 79 Interchangeability under the criterion of Flashback (AGA) .................................................................... 80 Interchangeability under the criterion of Lifting (AGA) .......................................................................... 81 Interchangeability under the criterion of yellow tipping (AGA) ............................................................. 82 References .................................................................................................................................................. 83 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner List of Figures Figure 1: Fossil fuel energy consumption as a percentage of the total consumption, extracted from [1]. . 1 Figure 2: Burner draft and assembled burner, extracted from [8]. .............................................................. 3 Figure 3: Volume modeled on the left and a section plane with named parts on the right. ....................... 5 Figure 4: ¼ of the Fluid domain is modeled taking advantage of the symmetry. ........................................ 6 Figure 5: Fluid domain divided into blocks. .................................................................................................. 6 Figure 6: Temperature profile of the centerline for each mesh. .................................................................. 8 Figure 7: Mesh M3. ....................................................................................................................................... 9 Figure 8: Residuals plot. .............................................................................................................................. 11 Figure 9: NO and CO mass fraction vs. Number of reactors for 100% CH4. ................................................ 13 Figure 10: Computational time needed to solve the reactor network for 100% CH4, using 3 cores of the Intel i5-4570 processor. .............................................................................................................................. 14 Figure 11: Maximum temperature vs. % of CH4 in volume, for type 1 mixtures. ....................................... 15 Figure 12: 100% CH4 temperature contour. ............................................................................................... 16 Figure 13: 80% CH4-20% H2 temperature contour. ..................................................................................... 17 Figure 14: 50% CH4-50% H2 temperature contour. ..................................................................................... 17 Figure 15: 40% CH4-60% H2 temperature contour. ..................................................................................... 18 Figure 16: 30% CH4-70% H2 temperature contour. ..................................................................................... 18 Figure 17: Maximum temperature vs. equivalence ratio for fuel type 1 mixtures. ................................... 19 Figure 18: Heat transferred to the brick walls vs. equivalence ratio for all type 1 mixtures. .................... 19 Figure 19: H2O mole fraction vs equivalence ratio for all type 1 mixtures. ................................................ 20 Figure 20: CO2 mole fraction vs equivalence ratio for all type 1 mixtures. ................................................ 20 Figure 21: NOx emissions in ppmdv @ 0% O2 vs. equivalence ratio for all type 1 mixtures. ..................... 21 Figure 22: EINOx in gNOx/kgFuel vs. equivalence ratio for all type 1 mixtures. ........................................ 22 Figure 23: EINOx in mgNOx/kWh vs. equivalence ratio for all type 1 mixtures. ........................................ 22 Figure 24: CO emissions in ppmdv @ 0% O2 vs. equivalence ratio for all type 1 mixtures. ....................... 24 Figure 25: EICO in gNOx/kgFuel vs. equivalence ratio for all type 1 mixtures. .......................................... 24 Figure 26: EINOx in mgNOx/kWh vs. equivalence ratio for all type 1 mixtures. ........................................ 25 Figure 27: Maximum temperature vs. % of CH4 in volume, for type 2 mixtures. ....................................... 26 Figure 28: 100% CH4 temperature contour. ............................................................................................... 27 Figure 29: 96% CH4-4% CO2 temperature contour. .................................................................................... 27 Figure 30: 60% CH4-40% CO2 temperature contour. .................................................................................. 28 Figure 31: 55% CH4-45% CO2 temperature contour. .................................................................................. 28 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner Figure 32: 45% CH4-55% CO2 temperature contour. .................................................................................. 29 Figure 33: OH mass fraction contour for the 100% CH4 case. .................................................................... 30 Figure 34: OH mass fraction contour for the 60% CH4-40% CO2 case. ....................................................... 30 Figure 35: Flue gases temperature vs. equivalence ratio for all type 2 mixtures. ...................................... 31 Figure 36: Heat transferred to the brick walls vs. equivalence ratio for all type 2 mixtures. .................... 32 Figure 37: H2O mole fraction vs equivalence ratio for all type 2 mixtures. ................................................ 33 Figure 38: CO2 mole fraction vs equivalence ratio for all type 2 mixtures. ................................................ 33 Figure 39: NOx emissions in ppmdv @ 0% O2 vs. equivalence ratio for all type 2 mixtures. ..................... 34 Figure 40: EINOx in gNOx/kgFuel vs. equivalence ratio for all type 2 mixtures. ........................................ 34 Figure 41: EINOx in mgNOx/kWh vs. equivalence ratio for all type 2 mixtures. ........................................ 35 Figure 42: CO emissions in ppmdv @ 0% O2 vs. equivalence ratio for all type 2 mixtures. ....................... 36 Figure 43: EICO in gCO/kgFuel vs. equivalence ratio for all type 2 mixtures. ............................................. 36 Figure 44: EICO in mgCO/kWh vs. equivalence ratio for all type 2 mixtures.............................................. 37 Figure 45: EICO in mgCO/kWh vs. equivalence ratio for mixtures with high content of CO2..................... 37 Figure 46: Maximum temperature vs. % of CH4 in volume, for type 3 mixtures. ....................................... 38 Figure 47: 100% CH4 temperature contour. ............................................................................................... 39 Figure 48: 85% CH4-15% C2H6 temperature contour. ................................................................................. 39 Figure 49: 75% CH4-25% C2H6 temperature contour. ................................................................................. 40 Figure 50: 75% CH4-25% C3H8 temperature contour. ................................................................................. 40 Figure 51: 75% CH4-15% C2H6-10% C3H8 temperature contour. ................................................................. 41 Figure 52: Flue gases temperature vs. equivalence ratio for all type 3 mixtures. ...................................... 42 Figure 53: Heat transferred to the brick walls vs. equivalence ratio for all type 3 mixtures. .................... 42 Figure 54: H2O mole fraction vs equivalence ratio for all type 3 mixtures. ................................................ 43 Figure 55: CO2 mole fraction vs equivalence ratio for all type 3 mixtures. ................................................ 43 Figure 56: NOx emissions in ppmdv @ 0% O2 vs. equivalence ratio for type III mixtures. ......................... 44 Figure 57: EINOx in gNOx/kgFuel vs. equivalence ratio for all type 3 mixtures. ........................................ 45 Figure 58: EINOx in mgNOx/kWh vs. equivalence ratio for all type III mixtures. ....................................... 45 Figure 59: CO emissions in ppmdv @ 0% O2 vs. equivalence ratio for all type 3 mixtures. ....................... 46 Figure 60: EICO in gNOx/kgFuel vs. equivalence ratio for all type 3 mixtures. .......................................... 46 Figure 61: EICO in mgNOx/kWhl vs. equivalence ratio for all type 3 mixtures. ......................................... 47 Figure 62: Detailed views and dimensions of the modeled domain. ......................................................... 50 Figure 63: Fluid domain. ............................................................................................................................. 51 Figure 64: Fuel inlet. ................................................................................................................................... 52 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner Figure 65: Oxygen inlet. .............................................................................................................................. 53 Figure 66: Outlet. ........................................................................................................................................ 54 Figure 67: Chimney walls. ........................................................................................................................... 55 Figure 68: Brick walls. ................................................................................................................................. 56 Figure 69: Pre-combustor walls. ................................................................................................................. 57 Figure 70: Fuel walls .................................................................................................................................... 58 Figure 71: Inner oxygen walls ..................................................................................................................... 58 Figure 72: Outer oxygen walls. ................................................................................................................... 59 Figure 73: Symmetric planes. ...................................................................................................................... 60 Figure 74: NO mass fraction vs. Number of reactors for 80% CH4-20% H2. ................................................ 61 Figure 75: CO mass fraction vs. Number of reactors for 80% CH4-20% H2. ................................................ 61 Figure 76: NO mass fraction vs. Number of reactors for 50% CH4-50% H2. ................................................ 62 Figure 77: CO mass fraction vs. Number of reactors for 50% CH4-50% H2. ................................................ 62 Figure 78: NO mass fraction vs. Number of reactors for 40% CH4-60% H2. ................................................ 63 Figure 79: CO mass fraction vs. Number of reactors for 40% CH4-60% H2. ................................................ 63 Figure 80: NO mass fraction vs. Number of reactors for 30% CH4-40% H2. ................................................ 64 Figure 81: CO mass fraction vs. Number of reactors for 30% CH4-70% H2. ................................................ 64 Figure 82: NO mass fraction vs. Number of reactors for 96% CH4-4% CO2. ............................................... 65 Figure 83: CO mass fraction vs. Number of reactors for 96% CH4-4% CO2. ................................................ 65 Figure 84: NO mass fraction vs. Number of reactors for 60% CH4-40% CO2. ............................................. 66 Figure 85: CO mass fraction vs. Number of reactors for 60% CH4-40% CO2. .............................................. 66 Figure 86: NO mass fraction vs. Number of reactors for 55% CH4-45% CO2. ............................................. 67 Figure 87: CO mass fraction vs. Number of reactors for 55% CH4-45% CO2. .............................................. 67 Figure 88: NO mass fraction vs. Number of reactors for 45% CH4-55% CO2. ............................................. 68 Figure 89: CO mass fraction vs. Number of reactors for 45% CH4-55% CO2. ............................................ 68 Figure 90: NO mass fraction vs. Number of reactors for 85% CH4-15% C2H6. ............................................ 69 Figure 91: CO mass fraction vs. Number of reactors for 85% CH4-15% C2H6. ............................................. 69 Figure 92: NO mass fraction vs. Number of reactors for 75% CH4-25% C2H6. ............................................ 70 Figure 93: CO mass fraction vs. Number of reactors for 75% CH4-25% C2H6. ............................................. 70 Figure 94: NO mass fraction vs. Number of reactors for 75% CH4-25% C3H8. ............................................ 71 Figure 95: CO mass fraction vs. Number of reactors for 75% CH4-25% C3H8. ............................................. 71 Figure 96: NO mass fraction vs. Number of reactors for 75% CH4-15% C2H6-10% C3H8. ............................ 72 Figure 97: CO mass fraction vs. Number of reactors for 75% CH4-15% C2H6-10% C3H8. ............................ 72 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner Figure 98: Equilibrium temperature [K] vs. equivalence ratio for the different oxidizer mixtures. ........... 73 Figure 99: Equilibrium NO [mass fraction] vs. equivalence ratio for the different oxidizer mixtures. ....... 73 Figure 100: Equilibrium CO [mass fraction] vs. equivalence ratio for the different oxidizer mixtures. ..... 74 Figure 101: Equilibrium temperature [K] vs. equivalence ratio for the different type 1 mixtures. ........... 74 Figure 102: Equilibrium NO [mass fraction] vs. equivalence ratio for the different type 1 mixtures. ....... 75 Figure 103: Equilibrium CO [mass fraction] vs. equivalence ratio for the different type 1 mixtures. ........ 75 Figure 104 Equilibrium temperature [K] vs. equivalence ratio for the different type 2 mixtures. ............. 76 Figure 105: Equilibrium NO [mass fraction] vs. equivalence ratio for the different type 2 mixtures. ....... 76 Figure 106: Equilibrium CO [mass fraction] vs. equivalence ratio for the different type 2 mixtures. ........ 77 Figure 107 Equilibrium temperature [K] vs. equivalence ratio for the different type 3 mixtures. ............. 77 Figure 108: Equilibrium NO [mass fraction] vs. equivalence ratio for the different type 3 mixtures. ....... 78 Figure 109: Equilibrium CO [mass fraction] vs. equivalence ratio for the different type 3 mixtures. ........ 78 Figure 110. AGA Flashback index. ............................................................................................................... 80 Figure 111. AGA lifting index. ..................................................................................................................... 81 Figure 112. AGA Yellow tipping index. Analysis of interchangeability for mixtures CH4-H2 and CO-H2. .... 82 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner List of Tables Table 1: Summary of emissions and stability/1. ............................................................................................. i Table 2: Operating conditions. ...................................................................................................................... 4 Table 3: Summarized information of the different meshes. ........................................................................ 7 Table 4: Models employed. ........................................................................................................................... 9 Table 5: Boundary conditions created and its type. ................................................................................... 10 Table 6: Mass flow rate report .................................................................................................................... 11 Table 7: Sensible transfer heat flux report. ................................................................................................ 12 Table 8: Studied fuel mixtures. ................................................................................................................... 15 Table 9: blow off limit for mixtures with high content of CO2. ................................................................... 31 Table 10: Summary of emissions for interchangeable mixtures (AGA). ..................................................... 49 Table 11: Pressure outlet parameters. ....................................................................................................... 54 Table 12: Brick walls properties [28]........................................................................................................... 56 Table 13: Pre-combustor walls properties [28]. ......................................................................................... 57 Table 14: Outer oxygen walls properties [28] ............................................................................................. 59 Table 15: AGA specification ranges............................................................................................................. 79 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner List of Equations Equation 1: CH4-H2 stoichiometric balance. ................................................................................................ 21 Equation 2: CH4-CO2 stoichiometric balance. ............................................................................................. 33 Equation 3: CH4-C2H6-C3H8 stoichiometric balance. .................................................................................... 44 Equation 4: Beam lentgh expression. ......................................................................................................... 51 Equation 5: Hydraulic diameter formula. ................................................................................................... 52 Equation 6: Inlet turbulent intensity formula [14]. .................................................................................... 52 Equation 7: Expression for the equivalence ratio [27]. .............................................................................. 53 Equation 8: Mass stoichiometric ratio [27]. ................................................................................................ 53 Equation 9: General combustion reaction. ................................................................................................. 54 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 6 | Page Figure 4: ¼ of the Fluid domain is modeled taking advantage of the symmetry. 3.3 Mesh To proceed and solve the problem it is necessary to divide the fluid domain into small cells, also referred as elements, with the purpose of treat each element as a single control volume where the suitable set of equations are solved. The solution of each cell, as well as the errors, are transferred to the neighboring elements in each iteration, until the convergence criteria is met. Obviously, the amount of either computational time or computational resources needed to reach the solution increases with the number of cells, but also a more accurate solution is achieved, thus the importance of creating an efficient and reliable mesh. 3.3.1 Approach In order to achieve the goals posed before, it has been proceeded to divide the fluid domain into different blocks, each one representing different zones of interest, enabling then to have high levels of refinement on those zones where smaller elements are needed to reach enough accuracy. Also, the subsequent blocks created are prisms, allowing then the sweep method of the ANSYS Workbench Mesher, which means that the mesh will be composed of hexahedral cells, which are quite more efficient than the otherwise employed tetrahedral cells [13]. The divided domain is shown in Figure 5, which precedes the consequent brief description of each block. Figure 5: Fluid domain divided into blocks. The block 1 represents the burner nozzle, followed by the pre-combustor chamber showed as block 2. The centerline of the pre-combustor, which is the zone where the flame is going to exist, is represented UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 7 | Page by the blocks number 3 and 4, being number 3 the core of the flame and number 4 the bulk of the combustion productions. Blocks 5 and 6 are intended as the surrounding of blocks 3 and 4, respectively. As it may seem logic, blocks numbered from 1 to 4 are those with a smaller element size and therefore the refined zones, whilst a larger element size is used on the remaining blocks. 3.3.2 Mesh sensitivity analysis To confirm that a converged model solution is mesh independent, it is necessary to carry out a mesh sensitivity analysis. It consists on running the same models and boundary conditions within more refined and coarser meshes, then compare the results and finally select the mesh with the least amount of elements that provides good results.. 3.3.2.1 Mesh quality The mesh quality plays a significant role in the accuracy and stability of the numerical solution, hence the importance of checking the quality of the mesh before running the solver. To achieve a fair sensitivity analysis it is necessary to make the comparison between meshes within reasonable quality intervals. The quality indicators checked during this analysis are the following: • Orthogonal quality: the worst cells will have an orthogonal quality closer to 0, with the best cells closer to 1. The minimum orthogonal quality for all types of cells should be more than 0.01, with an average value that is significantly higher. • Aspect ratio: is a measure of the stretching of the cell. Generally, it is best to avoid sudden and large changes in cell aspect ratios in areas where the flow field exhibit large changes or strong gradients. The values provided by Fluent 15.0 are therefore the minim value for the orthogonal quality and the maximum for the aspect ratio. More information regarding the equations used to compute the value of the indicators can be found in the Fluent User’s Guide [14].Table 3 summarizes the important features for the different meshes that have been created and tested: the number of elements, the minim orthogonal quality and the maximum aspect ratios. The differences between the quality indicators for the different meshes are insignificant, being thus the solution only dependent on the level of refinement. Name of the Mesh Number of elements Minimum orthogonal quality Maximum aspect ratio M1 850741 0.3045 13.79 M2 1130452 0.3045 13.79 M3 1305806 0.3399 16.47 M4 1697303 0.3399 16.47 M5 2271395 0.2695 17.02 Table 3: Summarized information of the different meshes. Meshes M1 and M2 have the same level of refinement for the blocks that compose the centerline of the domain (blocks number 1 to 4), differenced only by the element size of the surrounding blocks (blocks 5 and 6), and here lies the reason why they have the same quality indicators. This same characteristic is UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 8 | Page found in meshes M3 and M4, whose centerline is more refined compared to the previous meshes. Finally, mesh M5 is the one that presents the higher refinement on the centerline blocks and the surroundings. The models, boundary conditions and the approach employed in this analysis are presented in further sections of this Chapter 3, since they have been also used for the rest of the studied cases. In order to compare the results, the temperature profile for the z axis has been used, since it represents the centerline of the fluid domain and the zone where the flame lies. 3.3.2.2 Sensitivity analysis results Figure 6 shows the centerline temperature profile of the converged solution for each mesh. Figure 6: Temperature profile of the centerline for each mesh. As it can be observed, meshes with the same level of refinement of the centerline blocks present basically the same profile, regardless of the refinement of the surrounding blocks. Meshes with the poorest levels of refinement (M1 and M2) yield to a solution that differs significantly when compared to the more refined meshes, whereas the difference between meshes M3, M4 and M5 are considered to be negligible, since the maxim temperature and the position where it is reached, and the outlet temperature are essentially equal for the three cases, proving then that a mesh independent solution has been reached. Thus, the selected mesh has been the one named as M3, since it is the one that provides an accurate, stable and reliable solution, in a more efficient manner (referred to the computational time needed to reach the solution). Figure 7 shows an isometric view of the M3 mesh. 0 500 1000 1500 2000 2500 3000 0 0.5 1 1.5 2 2.5 3 3.5 Temperature [K] z position [m] M1 M2 M3 M4 M5 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 9 | Page Figure 7: Mesh M3. 3.4 Models Table 4 is an overview of the solver set up and the models used for the cases studied.. Information about the governing equations and its parameters can be found in the ANSYS FLUENT User’s Guide [14]. Table 4: Models employed. Solver Steady Pressure based 3D case Pressure-velocity coupling Simple Energy equation On Viscous model Realizable κ-ε Enhanced wall treatment Radiation model Discrete Ordinates (DO) Species model Partially Premixed Model The Realizable κ-ε is an update of the Standard κ-ε to model flow turbulence; it provides more accurate solutions than Standard κ-ε. To model the radiation heat transfer, the Discrete Ordinates (DO) model has been selected over the P-1 model. The latter tends to over predict the radiation heat transfer rate, especially in problems with high temperatures, which is the case of the oxy-gas flame. Since the oxidizer and fuel streams are not premixed but there is some level of premixing in the pre-combustor section; for the species model the Partially Premixed Model has been selected and coupled with the Zimont model [15] to calculate the turbulent flame speed. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 10 | Page 3.5 Boundary conditions A list of the boundary conditions and their type can be found in Table 5. A more detailed description of each boundary can be found in the Appendix A. Table 5: Boundary conditions created and its type. Boundary name Boundary type Value [unit] Fluid domain Interior ---- Fuel inlet Mass flow rate inlet Variable=f(heat_input, fuel class) Oxidizer inlet Mass flow rate inlet Variable=f(heat_input, fuel class, excess of oxidant) Outlet Pressure outlet P= 1 atm Chimney wall Wall Adiabatic Brick walls Wall Non-adiabatic Pre-combustor Wall Non-adiabatic Outer oxygen walls Wall Non-adiabatic Inner oxygen walls Wall Adiabatic Fuel walls Wall Adiabatic Symmetric planes Symmetry ____________---- 3.6 Solution Methods The procedure followed for all the simulations is presented below: • For the first 1000 iterations, the default methods were used; the Green Gauss Node was chosen to evaluate the gradients and derivatives. This is due to the increased accuracy of the method relative to the less computationally expensive Green-Gauss cell based method as discussed by ANSYS INC. [14]. • For the following 9000 iterations, the second order upwind discretization method has been use to solve all the equations. Second order discretization is better able to resolve curvature and gradients than the first order discretization scheme. This approach has always lead to a solution that meets all the requirements to accept it as a converged solution. This requirements are explained in the next section. 3.7 Convergence criteria The following conditions are required to consider a solution to be converged: • Stabilized residual plot: all the residuals must be stabilized at the end of the simulation at reasonable low values. • The unbalanced mass flow rate of the system, which is given by the difference between the sum of the inlet mass flow rates and the outlet mass flow rated, must be less than 1% of the total input (sum of the oxidizer mass flow rate and the fuel mass flow rate). UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 11 | Page • The unbalanced heat transfer rate, which is calculated by the difference between the heat of the reaction source and the heat lost through the outlet and the walls, must be less than 1% of the total input, which is the thermal power with a fixed value of 586 kW. Figure 8 along with Tables 6 and 7 serve as an example of how the residuals plot should look like and examples of acceptable mass and heat flux reports, respectively. All the examples given correspond to the reference case: fuel mixture of 100% CH4 and stoichiometric conditions. Figure 8: Residuals plot. Table 6: Mass flow rate report Fuel inlet 0.002927 kg/s Oxygen inlet 0.011782 kg/s Outlet -0.014709111 kg/s Net - 1.1315569e-07 kg/s % of the total input 0.00077 % UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 12 | Page Table 7: Sensible transfer heat flux report. Brick walls -104942.4 W Chimney walls 0 W Fuel inlet 11.899372 W Fuel walls 0 W Inner oxygen walls 0 W Outer oxygen walls -19.47994 W Outlet -40272.797 W Oxygen inlet 20.00363 W Pre-combustor walls -15.639146 W Heat of reaction source 145301.83 W Net 83.416728 W % of heat of reaction 0.057 % 4 Chemical reactor network (CRN) The emissions from the burner are subject to strict regulations, and more stringent regulations are foreseen. These reasons make the capability of predicting pollutant emissions a valuable tool during the design phase or to predict in the field the impact of the fuel composition on emissions and stability, and make decisions accordingly. Instead of building prototypes, this strategy allows to make informed decisions previous to the construction of the prototype and thus avoiding the waste of economic resources. Obtaining a detailed prediction of minor species, such as pollutant or harmful species (like NOx or CO), requires a detailed chemistry set; i.e. GRI 3.0 [16], UC San Diego mech [17], USC mech II [18]... , which are not included in the CFD simulations ; in general, the computation cost of the chemical reactions is more expensive than the computation of the fluid dynamics equations. However, main flame properties like temperatures, heat release, velocities and main species can be well predicted using a simplified kinetics mechanism. The output solution of the CFD simulation is used afterwards to construct an equivalent reactor network of ideal chemical reactors, and the concentration of minor species is calculated using a detailed chemical scheme. Therefore, the equivalent reactor network becomes the basis for high-fidelity kinetics simulation, while maintaining spatial information derived from the CFD model. For all the cases the reaction mechanism employed has been the GRI 3.0 [16]. 4.1 Zonal distribution Several methods have been proposed with the objective to achieve an equivalent reactor from the CFD solution obtained with a simplified chemical scheme [10], [19]. The new release ANSYS FLUENT 15.0 incorporates an option that allows the user to automatically generate an equivalent reactor network from a converged CFD solution. The user can select the number of reactors to fill the fluid domain. Being the grid of CFD and its solution the starting point, the first step consists in agglomerating the CFD cells into the specified number of reactors. For optimal performance, CFD cells grouped together in each reactor UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 13 | Page have temperatures and species mass fractions that are similar. For the partially premixed model, the cell temperature and the cell mixture fraction are employed to cluster the cells in different reactors, until the final number of reactors is reached. Finally, the species concentration in each reactor is calculated using the detailed chemical scheme, thus providing more information regarding the minor species [20]. Since the CRN is a postprocessor strategy, a precise solution of the pollutant species requires an accurate CFD solution. The CRN results are only good if the solution obtained with the CFD strategy alone is accurate. 4.2 Reactor network sensitivity analysis The equivalent reactor network generated consists of the defined number of reactors defined by the user, where every reactor is modeled as a Perfectly Stirred Reactor (PSR) that works at fixed temperature, which is computed when solving the reactor network. This fact becomes of capital importance when selecting the number of reactors for the network. In order to reach a solution independent from the number of reactors, a sensitivity analysis is carried out. The idea is to increase the number of reactors in the network and then check the results independence to this variable. For the sensitivity analysis the mass fraction of NO and CO were checked. Once the minimum number of reactors necessary to reach an accurate and mesh independent solution has been decided, that number is fixed for the different runs and only excess of oxidant is varied. The default settings for the calculation have been left default; only the maximum number of iterations has been increased to guarantee similar residual values. The reactor pressure is fixed and determined as the mass averaged pressure of the CFD cells conforming the reactor [20]. Figure 9 show the sensitivity NO and CO results for 100% CH4 and operating at stoichiometric conditions. The mass fractions of the species are measured at the stack outlet. For 550 reactors or more, the results become insensitive to this variable. Figure 9: NO and CO mass fraction vs. Number of reactors for 100% CH4. As it can be observed, a high number of reactors are required in order to reach an independent. A large number of reactors is required to provide a spatial resolution of the chemical reactions. Analyzing the results of Figure 9, it can be seen that for the CO mass fraction the solution seems to be stabilized when using more than 200 reactors, whilst the NO mass fraction needs more than 500 reactors 0 0.001 0.002 0.003 0.004 0.005 0.006 60 120 180 240 300 360 420 480 540 600 660 710 0 0.00005 0.0001 0.00015 0.0002 0.00025 0.0003 0.00035 CO [mas fraction] Number of reactors NO [mass fraction] NO CO UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 14 | Page to reach a stable value. For some cases an upper limit of 700 reactors was found. For those cases, it was not possible to create a network of 700 reactors or more due a recurrent error during the iterative process to solve the network equations. The sensitivity analysis for other fuel types are collected in Appendix B. Another remarkable point to account for in this analysis is the time required to reach the solution. Figure 10 shows the increase of the computational time needed when increasing the number of reactors of the network. Figure 10: Computational time needed to solve the reactor network for 100% CH4, using 3 cores of the Intel i5-4570 processor. As it is shown, the computational time increases in a considerably manner when increasing the number of reactors of the network. Here lays the importance to select the proper number of reactors with the goal to maintain the overall efficiency of the simulation (referred to computational time needed) at a reasonable level. 0 5 10 15 20 25 30 35 40 0 100 200 300 400 500 600 700 800 Time [min] Number of reactors UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 15 | Page 5 Results and discussion The fuel mixtures studied in this project are summarized in Table 8. The baseline fuel is 100% CH4. Table 8: Studied fuel mixtures. Fuel type 1 (hydrogen enriched natural gas) Fuel type 2 (Biogases) Fuel type 3 ( mixtures with heavier Alkanes) 80% CH4-20% H2 96% CH4-4% CO2 85% CH4-15% C2H6 50% CH4-50% H2 60% CH4-40% CO2 75% CH4-25% C2H6 40% CH4-60% H2 55% CH4-45% CO2 75% CH4-25% C3H8 30% CH4-70% H2 45% CH4-55% CO2 75% CH4-15% C2H6-10% C3H8 For each fuel mixtures, the global equivalence ratio is varied within the following values: 1, 0.9, 0.8 and 0.5, with the goal to study the effect of operating in fuel lean conditions. In sharp contrast with the premixed flames cases, a stability analysis is not needed due to the high stability of the diffusion flames. Thus in the results obtained at very lean conditions (equivalence ratio equal to 0.5) no blow off instabilities have been observed, with the exception of the fuel type 2 mixtures, whose limits will be explained in its own section. Finally, the upcoming sections present first the results obtained directly from the CFD simulations and afterwards the results provided by the reactor network solution. 5.1 Fuels type 1: Hydrogen enriched natural gas 5.1.1 CFD results Figure 11 presents the maximum temperature (at stoichiometric conditions) as function of the volumetric percentage of H2in the fuel mixture. Figure 11: Maximum temperature vs. % of CH4 in volume, for type 1 mixtures. 2750 2800 2850 2900 2950 3000 3050 3100 0 20 40 60 80 100 120 Temperature [K] CH4[% volume] UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 22 | Page Figure 22: EINOx in gNOx/kgFuel vs. equivalence ratio for all type 1 mixtures. Figure 23: EINOx in mgNOx/kWh vs. equivalence ratio for all type 1 mixtures. The following key points summarize the behavior of NOx emissions for the different type I fuel mixtures: • For a fixed equivalence ratio, the addition of hydrogen yields to an increase of NOx concentration, mainly due to the higher temperatures that are reached with mixtures with high content of hydrogen. The most important route to form NOx in oxy-fuel combustion is the thermal or Zeldovich route, whereas the Fenimore or prompt route becomes important due to the negative net production rates of NOx [23], [24]. Also the addition of H2 enhances the formation of intermediate species like OH and H, which promotes NOx formation through the Zeldovich, NNH and N2O pathways. o The thermal route is enhanced by the addition of H2 since the mechanism is comprised of the three following reactions: N2 + O = N + NO; N + O2 = NO + O; and N + OH = NO + H. 0 5 10 15 20 25 30 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 EINOx [gNOx/kgFuel] Equivalence Ratio 100% CH4 80% CH4-20%H2 50%CH4-50%H2 40%CH4-60%H2 30%CH4-70%H2 0 200 400 600 800 1000 1200 1400 1600 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 EINOx [mgNOx/kWh] Equivalence Ratio 100% CH4 80% CH4-20%H2 50%CH4-50%H2 40%CH4-60%H2 30%CH4-70%H2 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 23 | Page o The NNH reactions enhanced include NH + N = N2 +H; NH + NO = N2 + OH; NNH + M = N2 + H + M; NNH + O2 = HO2 + N2; NNH + O = OH + N2; NNH + H = H2 + N2; NNH + OH = H2O + N2; and NNH + CH3 = CH4 + N2. o The reactions that explain the effect of hydrogen addition on the N2O intermediate mechanism are the following: O + N2 + M = N2O + M; N2O + H = NO + NH; O + N2O = NO + NO. o Also the high availability of O2 in the oxidant increases the concentration of O-atoms, which also participates in all NOx formation pathways. • At a fixed equivalence ratio the addition of hydrogen affects the flame length and width, the flame becomes shorter and narrower with the addition of hydrogen to the fuel mixture. Shorter flames yield to less recirculation of the product gases into the reaction zone, which also explains the increase in NOx concentration. For all the fuel mixtures, operating at fuel lean conditions leads to a considerable increase of the NOx production, conversely to what happens in air-fired flames. As it is shown by the equilibrium calculations regarding the equivalence ratio effect for oxygen and air cases (see Appendix C), the excess of oxygen does not decrease the flame temperature as it happens with the excess of air. Increasing then the amount oxygen used means increasing the amount of N2 entering the chamber, since the oxidizer is assumed to have 1% of N2. High temperatures together with higher amount of N2 lead directly to a considerable increase of NOx production. The results discussed before are based on NOx concentration by volume, while the emission index (EINOx) sheds light on the NOx production on a mass basis. When comparing the EINOx for the different mixtures, the differences become less pronounced, but still exist. This is due to the increased amount of water produced when adding hydrogen to the combustion (it can be seen in Figure 23), once water is drop out of the system the dry emissions are shifted upwards virtually indicating a higher concentration (ppm_NOx_dry basis>ppm_NOx_wet_basis). For hydrogen containing fuels, the virtual shift caused by drying the sample, is significant. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 24 | Page 5.1.2.2 CO emissions The CO predictions, using the same units of measure used for NOx, are shown in Figures 24, 25 and 26. The trends have been compared with the results of the equilibrium calculation (see Appendix C) and have proved to be in good agreement. The only experimental data found addresses the natural gas case, fired at slightly lean conditions [22]. This measured value is marked in Figure 24 with a green point. Figure 24: CO emissions in ppmdv @ 0% O2 vs. equivalence ratio for all type 1 mixtures. Figure 25: EICO in gNOx/kgFuel vs. equivalence ratio for all type 1 mixtures. - 2,000 4,000 6,000 8,000 10,000 12,000 14,000 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 CO [ppmdv @0% O2] Equivalence Ratio 100%CH4 80%CH4-20%H2 50%CH4-50%H2 40%CH4-60%H2 30%CH4-70%H2 Experimental 0 5 10 15 20 25 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 EICO [gCO/kgFuel] Equivalence Ratio 100%CH4 80%CH4-20%H2 50%CH4-50%H2 40%CH4-60%H2 30%CH4-70%H2 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 25 | Page Figure 26: EINOx in mgNOx/kWh vs. equivalence ratio for all type 1 mixtures. The following key points serve to explain the trends and behavior of the predicted CO: • High levels of CO are obtained due to the high temperatures reached in oxy-fuel combustion, since it contributes to the dissociation of CO2 molecules into CO and O-atoms. • Due to its own nature, diffusion flames present local zones where the combustion is produced at overly fuel-lean or overly fuel-rich conditions, being both situations responsible of a not complete combustion [2], [24]. • For all the fuel mixtures, all the CO indices are triggered to very high values near at the stoichiometric point (equivalence ratio = 1). These high levels of CO can be explained by combining the fact that CO is generally produced by the incomplete combustion of carbon containing fuels and the high temperatures (that also promote the CO production) reached in oxy-fuel cases. When working at lean condition, the excess of oxygen helps to complete combustion process and thus lowering considerably the CO emissions. • The addition of hydrogen to the fuel reduces the emission of carbon monoxide, as it is better represented by Figure 26. Again, the higher reactivity of hydrogen helps to complete the oxidation reactions avoiding incomplete combustion products. Moreover, the additional hydrogen acts displacing carbon from the fuel, so inherently less CO can be formed. 0 200 400 600 800 1000 1200 1400 1600 1800 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 EICO [mgCO/kWh] Equivalence Ratio 100%CH4 80%CH4-20%H2 50%CH4-50%H2 40%CH4-60%H2 30%CH4-70%H2 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 26 | Page 5.2 Fuels type 2: Landfill/digester biogas 5.2.1 CFD results Figure 27 illustrates the effect that the addition of CO2 causes on the maximum temperature reached in the flame, which has been obtained from the CFD simulation, for the stoichiometric point. Figure 27: Maximum temperature vs. % of CH4 in volume, for type 2 mixtures. The maximum temperature decreases drastically with the addition of CO2 to the fuel mixture, as shown in Figure 27. This is due to the role that CO2 plays in the fuel: it acts only diluting the mixture. Therefore, the amount of CO2 needs to be heated up, and the flame temperature will be lowered consequently. This drop of the flame temperature will affect the pollutant emissions, as it will be discussed in the next subsection. The same trend has been obtained with the equilibrium calculation (see Appendix C) [21]. Figures 28 to 32 show the temperature contour for the different concentrations of CO2 in the fuel mixtures presented on Table 8, fired at stoichiometric conditions. 2660 2680 2700 2720 2740 2760 2780 2800 2820 0 20 40 60 80 100 120 Temperature [K] CH4 [% of volume] UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 27 | Page Figure 28: 100% CH4 temperature contour. Figure 29: 96% CH4-4% CO2 temperature contour. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 28 | Page Figure 30: 60% CH4-40% CO2 temperature contour. Figure 31: 55% CH4-45% CO2 temperature contour. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 29 | Page Figure 32: 45% CH4-55% CO2 temperature contour. Fuel bends with high content of CO2 tend to behave like premixed flames. Figures 33 and 34, which show the OH mass fraction of the horizontal plane for the 100% CH4 and 60%CH4-40%CO2 cases, are useful to compare the difference between the two cases: the 100% CH4 case presents a larger zone with higher gradients of OH as seen in diffusion flames, whereas the 60%CH4-40%CO2 case has a smaller zone with more uniform contour of OH typical from the premixed flames. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 30 | Page Figure 33: OH mass fraction contour for the 100% CH4 case. Figure 34: OH mass fraction contour for the 60% CH4-40% CO2 case. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 31 | Page Because of the premixed-flame behavior of the fuel mixtures with higher levels of CO2, a lean limit has been detected due to the irregular residual plot that the CFD solution presented, and also because it was found impossible to solve the reactor network. Table 9 indicates the lean blow off limit for these mixtures, in terms of the leanest equivalence ratio reachable. Table 9: blow off limit for mixtures with high content of CO2. 60% CH4-40% CO2 0.75 5% CH4-45% CO2 0.8 45% CH4-55% CO2 0.95 The variation of the exhaust temperature and the heat transferred to the bricks are presented through Figures 35 and 36, respectively. Figure 35: Flue gases temperature vs. equivalence ratio for all type 2 mixtures. 1550 1600 1650 1700 1750 1800 1850 1900 1950 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 Temperature [K] Equivalence Ratio 100%CH4 96%CH4-4%CO2 60%CH4-40%CO2 55%CH4-45%CO2 45%CH4-55%CO2 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 38 | Page • Also, since the addition of CO2 leads to the obtainment of pre-mixed flames, the local zones that present overly rich conditions, which promote the CO production are eliminated an thus the CO emissions are lowered [2]. • For all the fuel mixtures, the CO levels are triggered when reaching the stoichiometric point. Working in slightly lean conditions helps to complete the combustion reactions and therefore the CO is reduced since its formation is related to the incomplete combustion of the carbon containing fuels [2]. 5.3 Fuels type 3: Mixtures with heavier hydrocarbons 5.3.1 CFD results Figure 46 gathers the effect of the fuel composition on the maximum temperature of the solution obtained with the CFD simulations (at stoichiometric conditions). Each line represents the different type 3 fuel mixtures, whereas the horizontal axis corresponds to the content of percentage of CH4 by volume in the fuel blend. Figure 46: Maximum temperature vs. % of CH4 in volume, for type 3 mixtures. It can be seen that slightly higher temperatures are obtained when increasing the content of heavier hydrocarbons in the fuel mixture. Specially, the highest temperature has been reached with mixtures with larger amount of C3H8 present on the fuel mixture. Figures 47 to 51 show the temperature profile of the different mixtures provided by the CFD, at the stoichiometric point. 2800 2820 2840 2860 2880 2900 2920 2940 60 70 80 90 100 110 Temperature [K] CH4 [% of volume] CH4-C2H6 mixtures CH4-C3H8 mixtures CH4-C2H6-C3H8 mixtures UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 39 | Page Figure 47: 100% CH4 temperature contour. Figure 48: 85% CH4-15% C2H6 temperature contour. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 40 | Page Figure 49: 75% CH4-25% C2H6 temperature contour. Figure 50: 75% CH4-25% C3H8 temperature contour. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 41 | Page Figure 51: 75% CH4-15% C2H6-10% C3H8 temperature contour. No significant changes on the flame shape and characteristics have been detected. Only a slight decrease of the flame shape when increasing the content of heavier hydrocarbons. It can be concluded that comparable temperatures and profiles have been obtained due to the similar properties of the molecules that conform the different mixtures. The effect that operating at lean conditions has on the exhaust temperatures and the heat transferred to the brick walls is presented through Figures 52 and 53. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 42 | Page Figure 52: Flue gases temperature vs. equivalence ratio for all type 3 mixtures. Figure 53: Heat transferred to the brick walls vs. equivalence ratio for all type 3 mixtures. The excess of oxygen acts a heat sink of cooling the gases down and reducing the heat transfer to the bricks. The effect of the fuel composition on the heat transfer is negligible, mainly due to the similar properties that the different molecules composing the fuel and atmosphere in the combustion chamber. 1700 1750 1800 1850 1900 1950 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 Temperatire [K] Equivalence Ratio 100%CH4 75%CH4-25%C2H6 85%CH4-15%C2H6 75%CH4-25%C3H8 75%CH4-15%C2H6-10%C3H8 90000 92000 94000 96000 98000 100000 102000 104000 106000 108000 110000 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 Heat [W] Equivalence Ratio 100%CH4 75%CH4-25%C2H6 85%CH4-15%C2H6 75%CH4-25%C3H8 75%CH4-15%C2H6-10%C3H8 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 43 | Page 5.3.2 CRN results The results obtained for the main products of the combustion reaction, H2O and CO2, are presented in Figures 54 and 55, in which the molar fraction of these species can be found plotted against the equivalence ratio, for all the type 3 mixtures. Figure 54: H2O mole fraction vs equivalence ratio for all type 3 mixtures. Figure 55: CO2 mole fraction vs equivalence ratio for all type 3 mixtures. As it has been discussed in the other sections regarding the different types of fuels, the excess of oxygen characteristic of lean combustion acts diluting the concentration of both species. Figures 54 and 55 also show an increase of the CO2 concentration and a decrease of the H2O molar fraction that comes with the mixtures that have higher content of heavier hydrocarbons. This fact can be explained with the stoichiometric balance of the reaction, see Equation 3. 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 H2O [mole fraction] Equivalence Ratio 100%CH4 75%CH4-25%C2H6 85%CH4-15%C2H6 75%CH4-25%C3H8 75%CH4-15%C2H6-10%C3H8 0.15 0.20 0.25 0.30 0.35 0.40 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 CO2[mole fraction] Equivalence Ratio 100%CH4 75%CH4-25%C2H6 85%CH4-15%C2H6 75%CH4-25%C3H8 75%CH4-15%C2H6-10%C3H8 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 44 | Page 𝛼𝛼 ∙ 𝐶𝐶𝐻𝐻 4 +𝛽𝛽 ∙ 𝐶𝐶 2 𝐻𝐻 6 +𝛾𝛾 ∙ 𝐶𝐶 3 𝐻𝐻 8 +(4𝛼𝛼+ 7𝛽𝛽+10𝛾𝛾) 2∙ 𝑂𝑂 2 →(𝛼𝛼+ 2𝛽𝛽+ 3𝛾𝛾)∙ 𝐶𝐶𝑂𝑂 2 + (2𝛼𝛼+ 3𝛽𝛽+ 4𝛾𝛾)∙ 𝐻𝐻 2 𝑂𝑂 Equation 3: CH4-C2H6-C3H8 stoichiometric balance. 5.3.2.1 NOx emissions Figures 56, 57 and 58 present the results obtained for the NOx in the form of concentration in ppm (dry basis and corrected at 0% O2), emission index in gNOx/kgFuel and the emission index expressed in mgNOx/kWh, respectively. The only experimental data found addresses the natural gas case, fired at slightly lean conditions [22]. This measured value is marked in Figure 56 with a green point. Furthermore, the trends have been compared with the results of the equilibrium calculations (Appendix C) Figure 56: NOx emissions in ppmdv @ 0% O2 vs. equivalence ratio for type III mixtures. 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 NOx [ppmdv @0% O2] Equivalence Ratio 100%CH4 75%CH4-25%C2H6 85%CH4-15%C2H6 75%CH4-25%C3H8 75%CH4-15%C2H6-10%C3H8 Experimental UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 45 | Page Figure 57: EINOx in gNOx/kgFuel vs. equivalence ratio for all type 3 mixtures. Figure 58: EINOx in mgNOx/kWh vs. equivalence ratio for all type III mixtures. The following key points summarize the behavior of NOx emissions for the different type III fuel mixtures: • As it can be observed in all the Figures, the results for the different mixtures are collapsed around the same values, again due to the similarities of the different fuels. However, the NOx emissions of the 75% CH425% C3H8 seem to be slightly higher, thus indicating the temperature dependence of the NOx formation. • It is clear that the excess of oxygen affects increasing the NOx emissions, as it has been found for all the fuel types. The increased amount of the oxidizer mass flow entering the chamber does not lower the temperatures enough to counteract the larger amount of N2 that comes with it, thus resulting the NOx emissions triggered at lean conditions. 0 2 4 6 8 10 12 14 16 18 20 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 EINOx [gNOx/kgFuel] Equivalence Ratio 100%CH4 75%CH4-25%C2H6 85%CH4-15%C2H6 75%CH4-25%C3H8 75%CH4-15%C2H6-10%C3H8 0 200 400 600 800 1000 1200 1400 1600 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 EINOx [mgNOx/kWh] Equivalence Ratio 100%CH4 75%CH4-25%C2H6 85%CH4-15%C2H6 75%CH4-25%C3H8 75%CH4-15%C2H6-10%C3H8 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 46 | Page 5.3.2.2 CO emissions The CO predictions using the same terms that have been employed before are presented in Figures 59, 60 and 61. The only experimental data found addresses the natural gas case, fired at slightly lean conditions [22]. This measured value is marked in Figure 59 with a green point. Again, the predicted trends are in agreement with the equilibrium calculations, with the exception of the predicted values at the stoichiometric point. Figure 59: CO emissions in ppmdv @ 0% O2 vs. equivalence ratio for all type 3 mixtures. Figure 60: EICO in gNOx/kgFuel vs. equivalence ratio for all type 3 mixtures. 0 2000 4000 6000 8000 10000 12000 14000 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 CO [ppmdv @0% O2] Equivalence Ratio 100%CH4 75%CH4-25%C2H6 85%CH4-15%C2H6 75%CH4-25%C3H8 75%CH4-15%C2H6-10%C3H8 Experimental 0 5 10 15 20 25 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 EICO [gCO/kgFuel] Equivalence Ratio 100%CH4 75%CH4-25%C2H6 85%CH4-15%C2H6 75%CH4-25%C3H8 75%CH4-15%C2H6-10%C3H8 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 47 | Page Figure 61: EICO in mgNOx/kWhl vs. equivalence ratio for all type 3 mixtures. The following key points serve to explain the trends and behavior of the predicted CO: • High values for the CO emissions are obtained due to the nature of diffusion flames and the high temperatures reached in oxy-fuel combustion, as it has been posed before. • The trends predicted for the CO seem to be collapsed when comparing the different fuel mixtures at lean conditions. However, the results obtained for the stoichiometric point differ significantly of the expected trend observed in the equilibrium calculations. Furthermore, Figures 59, 60 and 62 clearly show that mixtures with higher content of heavier hydrocarbons produce less CO than mixtures with larger amount of CH4, at the stoichiometric point. • The values for the CO are again triggered at the stoichiometric point. Since CO is generally produced by the incomplete reaction of carbon containing fuels, and the excess of oxygen helps to complete the fuel combustion. Therefore the produced CO is decreased at equivalence ratios below the stoichiometric one [2]. 0 200 400 600 800 1000 1200 1400 1600 1800 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 EICO mgCO/kWh] Equivalence Ratio 100%CH4 75%CH4-25%C2H6 85%CH4-15%C2H6 75%CH4-25%C3H8 75%CH4-15%C2H6-10%C3H8 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 54 | Page 𝑣𝑣′𝐹𝐹∙𝐹𝐹𝐹𝐹𝑅𝑅𝐹𝐹 +𝑣𝑣′𝑂𝑂∙𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑅𝑅𝑂𝑂 →𝑃𝑃𝑂𝑂𝑃𝑃𝑂𝑂𝐹𝐹𝑃𝑃𝑃𝑃𝑠𝑠 Equation 9: General combustion reaction. The parameters regarding the turbulence information are obtained by following the same procedure used for the fuel inlet. The inlet temperature and pressure are assumed to be 300 K and 1 atm, respectively. Finally, no change is needed for the boundary conditions related to the species model since the default value of 0 for the species variables matches the theoretical conditions at the oxidizer inlet. Outlet Figure 66: Outlet. The outlet of the furnace remains at the atmospheric pressure, which is in agreement with the default value of 0 Pascal for the gauge pressure. Table 11 presents the information of the turbulence and other parameters that have been employed for the pressure outlet boundary condition. Table 11: Pressure outlet parameters. Variable [units] --- Gauge pressure [Pa] 0 Backflow turbulent kinetic energy [m 2 /s 2 ] 1 Backflow turbulent dissipation rate [m 2 /s 3 ] 1 Backflow temperature [K] 1500 Backflow progress variable 1 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 55 | Page The default values for the backflow turbulent kinetic energy and turbulent dissipation rate are used, as these values have also been employed in other studies [26]. The backflow temperature is set to 1500 K as a rough estimation of the outlet temperature. Finally, the backflow progress variable is set to 1 in order to indicate that at the outlet the reaction is completed. Chimney wall Figure 67: Chimney walls. The default properties of this wall have not been modified, being therefore an adiabatic wall. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 56 | Page Brick walls Figure 68: Brick walls. The brick walls represent the furnace walls, which are usually water cooled during experimental runs, in order to simulate the load that is present while operating under normal conditions. Thus convection option has been selected for these walls, but also due to the benefits (in terms of the stability of the numerical simulation) that presents when compared with the fixed temperature or fixed heat transfer rate options. Table 12 gathers the material and thermal properties that are required. Table 12: Brick walls properties [28]. Material name Brick Conductivity [W/m·K] 4 Thickness [mm] 150 Heat transfer coefficient [W/m 2 ·K] 250 Free stream temperature [K] 298 Internal emissivity 0.6 With this set of parameters the solution reached when modelling pure methane combustion yields a wall temperature at around 1800 K, well below the melting point of the ceramic materials usually employed in industrial furnaces, and also a heat flux of approximately 35 kW/m2. These results are in agreement with the experimental data extracted from [29]. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 57 | Page Pre-combustor walls Figure 69: Pre-combustor walls. As its name denotes, the pre-combustor walls represent the delimitating walls of the pre-combustor chamber, which are made of an insulating material in order to prevent high temperatures at the burner nozzle that compromise its proper functioning. Instead of using the adiabatic wall option, the convection one has been preferred. Table 13 summarizes the properties of the material for this wall and the convective properties, which pretend to model the insulating capacity exposed before. Table 13: Pre-combustor walls properties [28]. Boundary name Brick 2 Conductivity [W/m·K] 0.32 Thickness [mm] 200 Heat transfer coefficient [W/m 2 ·K] 250 Free stream temperature [K] 298 Internal emissivity 0.75 As it can be noticed, the thermal resistance derived from the pre-combustor wall properties is much higher than the associated with the brick walls, being then the heat lost through the pre-combustor negligible. UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 58 | Page Fuel and inner oxygen walls Figure 70: Fuel walls Figure 71: Inner oxygen walls UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 59 | Page They both represent the surface of the solid part inside the burner nozzle whose function is to maintain separated the oxygen and fuel streams. The heat lost through one of this walls is gained by the other one, as it can be demonstrated with a simple heat balance. Also, this amount of transferred heat can be considered negligible, since the both streams are at the same temperature when flowing through the burner nozzle. Taking into account these points, it has seem reasonable to leave the adiabatic condition for these walls. Outer oxygen walls Figure 72: Outer oxygen walls. The outer oxygen walls represent the outer walls of the burner nozzle, which surround the oxygen stream. The convection option has been also chosen for this boundary condition, whose properties are shown in Table 14. Table 14: Outer oxygen walls properties [28] Material name Steel Conductivity [W/m·K] 16.27 Thickness [mm] 4 Heat transfer coefficient [W/m2·K] 250 Free stream temperature [K] 298 Internal emissivity 0.85 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 60 | Page Even though the properties presented in Table 14 would yield to a low thermal resistance, the insulating role of the pre-combustor appears, and the solution shows that the heat lost through these outer walls is practically negligible. Symmetric planes Figure 73: Symmetric planes. No modifications on the defaults settings have been done since they are not needed for this type of boundary condition. . UCI Combustion Laboratory Burner Config. #2—Surface Stabilized Combustion 61 | Page Appendix B: Reactor network sensitivity analysis The gathered results of the sensitivity analysis are shown in Figures from 74 to 97. The analysis has been done on the stoichiometric case for each of the mixtures of each type. Fuel type 1: Hydrogen mixtures 80% CH4-20% H2 Figure 74: NO mass fraction vs. Number of reactors for 80% CH4-20% H2. Figure 75: CO mass fraction vs. Number of reactors for 80% CH4-20% H2. 0.00E+00 5.00E-05 1.00E-04 1.50E-04 2.00E-04 2.50E-04 3.00E-04 3.50E-04 4.00E-04 0 100 200 300 400 500 600 700 800 NO [mass fraction] Number of reactors 0 0.001 0.002 0.003 0.004 0.005 0.006 0 100 200 300 400 500 600 700 800 CO [mass fraction] Number of reactors UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 62 | Page 50% CH4-50% H2 Figure 76: NO mass fraction vs. Number of reactors for 50% CH4-50% H2. Figure 77: CO mass fraction vs. Number of reactors for 50% CH4-50% H2. 0 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006 0 100 200 300 400 500 600 700 800 900 1000 NO [mass fraction] Number of reactors 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0 100 200 300 400 500 600 700 800 900 1000 CO [mass fraction] Number of reactors UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 63 | Page 40% CH4-60% H2 Figure 78: NO mass fraction vs. Number of reactors for 40% CH4-60% H2. Figure 79: CO mass fraction vs. Number of reactors for 40% CH4-60% H2. 0 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006 0 200 400 600 800 1000 1200 NO [mas fraction] Number of reactors 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0 200 400 600 800 1000 1200 CO [mass fraction] Number of reactors UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 70 | Page 75% CH4-25% C2H6 Figure 92: NO mass fraction vs. Number of reactors for 75% CH4-25% C2H6. Figure 93: CO mass fraction vs. Number of reactors for 75% CH4-25% C2H6. 0 0.00005 0.0001 0.00015 0.0002 0.00025 0.0003 0.00035 0.0004 0.00045 0.0005 0 100 200 300 400 500 600 700 800 900 1000 NO [mass fraction] Number of reactors 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0 100 200 300 400 500 600 700 800 900 1000 CO [mass fraction] Number of reactors UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 71 | Page 75% CH4-25% C3H8 Figure 94: NO mass fraction vs. Number of reactors for 75% CH4-25% C3H8. Figure 95: CO mass fraction vs. Number of reactors for 75% CH4-25% C3H8. 0 0.00005 0.0001 0.00015 0.0002 0.00025 0.0003 0.00035 0.0004 0 100 200 300 400 500 600 NO [mass fraction] Number of reactors 0.00096 0.00098 0.001 0.00102 0.00104 0.00106 0.00108 0.0011 0.00112 0.00114 0.00116 0 100 200 300 400 500 600 CO [mass fraction] Number of reactors UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 72 | Page 75% CH4-15% C2H6-10%C3H8 Figure 96: NO mass fraction vs. Number of reactors for 75% CH4-15% C2H6-10% C3H8. Figure 97: CO mass fraction vs. Number of reactors for 75% CH4-15% C2H6-10% C3H8. 0 0.00005 0.0001 0.00015 0.0002 0.00025 0.0003 0.00035 0.0004 0.00045 0 200 400 600 800 1000 1200 NO [mas fraction] Number of reactors 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0 200 400 600 800 1000 1200 CO [mass fraction] Number of reactors UCI Combustion Laboratory Burner Config. #2—Surface Stabilized Combustion 73 | Page Appendix C: Equilibrium calculations The equilibrium model available in the CHEMKIN PRO software has been employed to perform all the simulations that follow this section [21]. N2 effect Figures 98, 99 and 100 show how the amount of nitrogen that the oxidizer stream contains affects the temperature, NO and CO concentrations, respectively, assuming that the equilibrium is reached. Figure 98: Equilibrium temperature [K] vs. equivalence ratio for the different oxidizer mixtures. Figure 99: Equilibrium NO [mass fraction] vs. equivalence ratio for the different oxidizer mixtures. 1000 1500 2000 2500 3000 3500 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 Temperature [K] Equivalence Ratio 99%O2-1%N2 90%O2-10%N2 80%O2-20%N2 70%O2-30%N2 60%O2-40%N2 50%O2-50%N2 40%O2-60%N2 30%O2-70%N2 21%O2-79%N2 0 0.005 0.01 0.015 0.02 0.025 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 NO [mass fraction] Equivalence Ratio 99%O2-1%N2 90%O2-10%N2 80%O2-20%N2 70%O2-30%N2 60%O2-40%N2 50%O2-50%N2 40%O2-60%N2 30%O2-70%N2 21%O2-79%N2 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 74 | Page Figure 100: Equilibrium CO [mass fraction] vs. equivalence ratio for the different oxidizer mixtures. Fuel type 1: Hydrogen mixtures Figures 101, 102 and 103 present the results obtained for the temperature, NO and CO mass fractions, assuming the equilibrium is reached, for the type 1 fuel mixtures. Figure 101: Equilibrium temperature [K] vs. equivalence ratio for the different type 1 mixtures. 0 0.05 0.1 0.15 0.2 0.25 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 CO [mass fraction] Equivalence Ratio 99%O2-1%N2 90%O2-10%N2 80%O2-20%N2 70%O2-30%N2 60%O2-40%N2 50%O2-50%N2 40%O2-60%N2 30%O2-70%N2 21%O2-79%N2 2800 2850 2900 2950 3000 3050 3100 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 Temperature [K] Equivalence Ratio 100%CH4 80%CH4-20%H2 50%CH4-50%H2 40%CH4-60%H2 30%CH4-70%H2 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 75 | Page Figure 102: Equilibrium NO [mass fraction] vs. equivalence ratio for the different type 1 mixtures. Figure 103: Equilibrium CO [mass fraction] vs. equivalence ratio for the different type 1 mixtures. 0.003 0.0035 0.004 0.0045 0.005 0.0055 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 NO [mass fraction] Equivalence Ratio 100%CH4 80%CH4-20%H2 50%CH4-50%H2 40%CH4-60%H2 30%CH4-70%H2 0 0.05 0.1 0.15 0.2 0.25 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 CO [mass fraction] Equivalence Ratio 100%CH4 80%CH4-20%H2 50%CH4-50%H2 40%CH4-60%H2 30%CH4-70%H2 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 76 | Page Fuel type 2: Landfill/digester biogas In Figures 104, 105 and 106 the temperature and the NO and CO mass fractions, respectively, can be found for the different mixtures of the type 2, assuming the equilibrium is reached. Figure 104 Equilibrium temperature [K] vs. equivalence ratio for the different type 2 mixtures. Figure 105: Equilibrium NO [mass fraction] vs. equivalence ratio for the different type 2 mixtures. 2550 2600 2650 2700 2750 2800 2850 2900 2950 3000 3050 3100 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 Temperature [K] Equivalence Ratio 100%CH4 96%CH4-4%CO2 60%CH4-40%CO2 55%CH4-45%CO2 45%CH4-55%CO2 0 0.001 0.002 0.003 0.004 0.005 0.006 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 NO [mass fraction] Equivalence Ratio 100%CH4 96%CH4-4%CO2 60%CH4-40%CO2 55%CH4-45%CO2 45%CH4-55%CO2 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 77 | Page Figure 106: Equilibrium CO [mass fraction] vs. equivalence ratio for the different type 2 mixtures. Fuel type 3: Mixtures with heavier hydrocarbons The temperatures and the NO and CO mass fractions obtained for the different type 3 mixtures, with the equilibrium assumption, are presented in Figures 107, 108 and 109. Figure 107 Equilibrium temperature [K] vs. equivalence ratio for the different type 3 mixtures. 0 0.05 0.1 0.15 0.2 0.25 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 CO [mass fraction] Equivalence Ratio 100%CH4 96%CH4-4%CO2 60%CH4-40%CO2 55%CH4-45%CO2 45%CH4-55%CO2 2800 2850 2900 2950 3000 3050 3100 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 Equilibrium Temperature [K] Equivalence Ratio 100%CH4 75%CH4-25%C3H8 85%CH4-15%CH6 75%CH4-10%C3H8-15%C2H6 75%CH4-25%C2H6 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 78 | Page Figure 108: Equilibrium NO [mass fraction] vs. equivalence ratio for the different type 3 mixtures. Figure 109: Equilibrium CO [mass fraction] vs. equivalence ratio for the different type 3 mixtures. 0.003 0.0032 0.0034 0.0036 0.0038 0.004 0.0042 0.0044 0.0046 0.0048 0.005 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 NO [mass fraction] Equivalence Ratio 100%CH4 75%CH4-25%C3H8 85%CH4-15%CH6 75%CH4-10%C3H8-15%C2H6 75%CH4-25%C2H6 0 0.05 0.1 0.15 0.2 0.25 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 CO [mass fraction] Equivalence Ratio 100%CH4 75%CH4-25%C3H8 85%CH4-15%CH6 75%CH4-10%C3H8-15%C2H6 75%CH4-25%C2H6 UCI Combustion Laboratory Burner Config. #3—The Oxy-Fuel Burner 79 | Page Appendix D: Interchangeability analysis In this section an interchangeability analysis is presented. For the analysis we take into account the Wobbe index and the rule 30. The former regulates the content of other constituents of the natural gas distributed in California. According to rule 30 the gas shall have a minimum Wobbe Number of 1279 Btu/scf and shall not have a maximum Wobbe Number greater than 1385 Btu/scf. The gas shall meet American Gas Association's Lifting Index, Flashback Index and Yellow Tip Index interchangeability indices for high methane gas relative to a typical composition of gas in the Utility system serving the area. Acceptable specification ranges are: Table 15: AGA specification ranges. AGA Indices Acceptable specification ranges Lifting Index (IL) IL <= 1.06 Flashback Index (IF) IF <= 1.2 Yellow Tip Index (IY) IY >= 0.8 Regarding the fuel composition, rule 30 specifies: Carbon Dioxide: The gas shall not have total carbon dioxide content in excess of three percent (3%) by volume. Inerts: The gas shall not contain in excess of four percent (4%) total inerts (the total combined carbon dioxide, nitrogen, oxygen and any other inert compound) by volume. Since the Wobbe Index is only concerned with matching heat release for a given burner, other indices have been developed to assess the interchangeability of other flame properties such as lifting, yellow tipping, flash back, air supply, incomplete combustion, and burner load [1]. In 1946 the American Gas Association (AGA) carried out extensive experimental research on fuel gas interchangeability [30]. AGA Tests were done on a specially developed partial premixing Bunsen-type burner and focused on establishing criteria for blending “supplemental” or “peaking gases” with base load supplies or adjustment gases; those adjustment gases were the three most historically representative natural gases available in the U.S at that time [31]. Based on this experimental work, AGA developed several empirical indices to address the effects of fuel interchange on Yellow Tipping ( Y I ), Flame Lifting ( L I ) and Flash Back ( F I ). The methodology to calculate these indices is presented in the AGA bulletin#36 of 1946 [30]. These indices may not be applicable to the typical complex turbulent premixed flames found in current practical systems. The inaccuracies expected from applying AGA indices to current fuels may be even greater, since current fuels of interest include coal derived syngas, landfill and biomass gases, imported liquefied natural gas (LNG) and hydrogen augmented fuels whose composition is completely different to those natural gases used to set the stability criteria. Nonetheless, these indices were considered as the most advanced methods and effective tools to predict interchangeability in the United States and should be considered as a starting point. Considerations before applying the AGA indices: