Patent Application
20200049343
Fossil fuels produce highly reliable and low-cost energy. Nevertheless, increasing concerns over global warming and the role of anthropogenic carbon dioxide emissions from fossil fuel combustion have increased the demand for low-carbon technologies. Carbon capture, utilization, and sequestration (CCUS) is an important strategy to minimize these emissions by capturing the carbon dioxide produced. CCUS is a relatively low cost option, and it could also have a wide impact on supplying reliable low-cost electricity worldwide. Oxy-fuel combustion is one of the three major strategies for capturing carbon dioxide (CO2) from stationary combustion furnaces and power plants. The most common oxy-fuel concept is to combust a pulverized fuel (coal) in a stream of oxygen that has\ been diluted with a large amount of recycled flue gas (˜70% of the flue gas is recycled). The flue gas, after moisture removal, consists of mainly CO2, which, after purification to the required standard, can be either utilized or sequestered. The recycled flue gas acts as an inert to reduce the temperature in the boiler and thereby control the heat flux to within the constraints of the boiler tube materials. A challenge of existing “first-generation” oxy-combustion technologies is that the plant efficiencies are rather low and the cost of electricity is rather high, relative to power generation without capture. The two system parameters that have the greatest potential for increasing efficiencies and reducing costs are the gas pressure in the boiler and the amount of flue gas recycled.
The inherent requirement of high pressure carbon dioxide for either enhanced oil recovery (EOR) or sequestration makes it possible to pressurize the oxy-combustion process with no intrinsic added cost, since for coal combustion compressing oxygen before combustion requires comparable energy to compressing carbon dioxide after combustion. The dew point of the flue gas moisture increases with pressure, so that at higher pressure, condensation occurs at a higher temperature than at atmospheric pressure. Thus, pressurized oxy-combustion is a well-posed method to increase the efficiency of carbon capture. A significant portion of the latent heat of condensation, extracted at higher temperature, can be utilized effectively in the steam cycle, instead of being wasted, as it is in atmospheric pressure oxy-combustion. A number of approaches to pressurized oxy-combustion have been proposed, and their efficiencies have been independently compared. In some embodiments, staged, pressurized oxy-combustion (SPOC) technology reduces the efficiency penalty for carbon capture from about 10 percentage points for a 1st generation technology to less than about 3 percentage points.
The reduction of flue gas recycle has traditionally been challenging because flue gas recycle has been used in oxy-combustion boilers to create an environment similar to air-fired combustion, i.e., to control the temperature and residence times in the oxy-combustion boiler so that they are similar to those in air-fired combustion. Though the primary need is to control wall heat flux and ensure that it is always within the material constraints of the boiler tubes, most first-generation approaches do this by controlling temperature, and flue gas recycle is the most common approach to controlling temperature, with recycling being typically 70% of the flue gas. Still, there has been interest in trying to reduce the amount of flue gas recycle to increase efficiency and reduce cost. Furthermore, reducing flue gas recycle increases the proportion of heat transfer on the furnace wall that is due to radiative heat transfer over convective heat transfer. Since radiative heat flux from a flame is much higher than convective heat flux in power plants, reducing flue gas recycle could reduce the amount of boiler tube materials required, thereby reducing the capital cost. A few combustor designs have been proposed for reduced recycle and have provided a summary of various efforts towards reduced flue gas recycle designs and demonstrations in. These noted that even though many of the demonstrated designs have shown performance improvements, they were demonstrated for either industrial furnaces or for boilers with low thermal input and low temperature and pressure steam, where much higher heat fluxes can be handled than in a typical utility-scale boiler.
None of these designs, however, have been demonstrated for utility-scale boilers, due to their inability to control wall heat flux within manageable levels. The SPOC process uses very unique approaches of fuel staging combined with radiative trapping to reduce flue gas recycle without increasing the radiative heat flux to the wall. Radiative trapping is a method to control radiative wall heat flux even with very high flame temperatures. In the SPOC process, separate boilers or stages are connected in series, with the fuel and oxidizer being distributed through the stages. The oxidizer is fed to the first stage at a much higher stoichiometry than required for the fuel in that stage. After combustion of the fuel fed to the first stage and subsequent heat transfer to the boiler tubes, the partially cooled flue gas and the excess oxygen are transferred to the next stage, where more fuel and some oxidizer is fed, and this process is repeated in subsequent stages until all the fuel and nearly all the oxygen is consumed. Conceptually, this method of fuel-staging could be used to reduce the amount of flue gas recycle to a very small amount. In some embodiments, the use of fuel staging provides the conditions necessary to successfully incorporate radiative trapping to control heat flux. Using computational fluid dynamics (CFD) simulations, shows a design of a SPOC boiler system where, even at very low flue gas recycle, the wall heat fluxes for all stages were controlled to manageable levels for utility scale applications.
Due to combustion in a high oxygen concentration environment, the first stage in the SPOC process is the most challenging in terms of controlling wall heat flux. In some embodiments of the present disclosure, a potential practical design of the burner and boiler for a SPOC process is disclosed, which provides additional flexibility in terms of operation, and the opportunity to control the oxygen concentration near the boiler tubes. In some embodiments, computational fluid dynamics (CFD) simulations are used, and the effect of various design parameters on the wall heat flux show how these are utilized as a toolbox for the design of a low recycle, pressurized oxy-combustion burner/boiler where wall heat flux is controlled via radiative trapping rather than large amounts of flue gas recycle.
In one aspect, the present disclosure is directed to a staged, pressurized oxy-combustion system, the system comprising a unique design including at least one burner and at least one boiler.
In another aspect, the present disclosure is directed to a method of capturing carbon dioxide, the method comprising using a staged, pressurized oxy-combustion system including a unique design, where the design includes at least one burner and at least one boiler.
In yet another aspect, the present disclosure is directed to a staged, pressurized oxy-combustion system, the system comprising a unique design including at least one burner and at least one boiler, wherein the boiler includes at least one wall ring.
In still another aspect, the present disclosure is directed to a method of capturing carbon dioxide, the method comprising using a staged, pressurized oxy-combustion system including a unique design, where the design includes at least one burner and at least one boiler, wherein the boiler includes at least one wall ring.
The present disclosure is generally directed to a burner and boiler/furnace for pressurized oxy-combustion boilers and furnaces. The disclosure includes a design of a burner and boiler for a staged, pressurized oxy-combustion (SPOC) technology process and designs that affect wall heat flux. The disclosure further includes the introduction of wall rings to increase, for example, advection.
To understand the burner and boiler design for SPOC, it is important to understand the concept of radiative trapping.
Radiative trapping, in the present context, is a means to control the radiative heat flux to the boiler tubes. In some boilers and furnaces, the local radiative emission in the central core could be very high due to the high particle number concentrations and high temperatures there. But, with radiative trapping, in some embodiments, the radiation leaving this region is “trapped” by the “optically dense medium” between the high temperature region and the wall. Depending on the effectiveness of the implementation of radiative trapping, the peak wall heat flux is reduced to a value comparable to or even lower than those in traditional boilers.
An exemplary embodiment of radiative trapping is shown in
τw(x)=∫0xke(x)dx|.
The extinction coefficient is defined as the inverse of the distance that radiation travels in a medium before it is scattered and/or absorbed. τw increases from zero at the wall to a higher thickness when approaching the axis of the boiler. A medium is considered optically dense when τw is much larger than 1, because the radiation penetration depth is much smaller than the physical distance to the wall from this location.
Based on Beer's law, the transmissivity of a medium with an optical thickness, τw, is given by exp(−τw(x)). A Higher optical thickness leads to a lower transmission of radiation through the medium. For example, if τw is 2.3 at a certain location, the transmissivity between this location and the wall is about 10% (about =e−2.3), meaning that of all the incident radiation reaching this location, only 10% passes through to the wall. The rest is “trapped” in the form of back-scattering and absorption within the medium. In such a situation, even if the local emission in the system is very high in some regions, especially in regions where both the temperature and the number of emitters (particles) are very high, the incident radiation at the wall is independent of the high local emissions. Instead, it is mainly dependent on the temperature and medium properties between the wall and the location where τw is 2.3. This region is termed as the “radiation penetration layer” (see, e.g.,
However, in a low recycle, pressurized oxy-combustion system, the particle number concentration is increased significantly due to the combination of high pressure and low dilution gas flow rate, that the medium becomes optically dense. Such a medium then effectively traps a significant portion of the radiation emanating from the core of the boiler. With a proper burner and boiler design to achieve low temperature in the radiation penetration layer, low-recycle, pressurized oxy-combustion boilers utilize radiative trapping to control wall heat flux, instead of flue gas recycle.
Note that, in some embodiments, radiation is also trapped in other pressurized systems, such as gasifiers, where the heat transfer in regions of high extinction coefficients is effectively modeled as a conduction problem rather than radiation.
In some embodiments, parametric analysis of process efficiency showed that for the SPOC process, the boilers should operate at about 16 bar. Since the boiler must be a pressure vessel, a cylindrical vessel was used as the base design. To minimize the total heat transfer surface area required in this boiler, a high, yet controlled and relatively uniform heat flux is desired over a long length of the boiler. Thus, an axial-flow burner is preferred over a swirl-stabilized burner, because the latter relies on high mixing and rapid heat release near the burner, while the former has slower mixing and axially-distributed heat release. While swirling flow is required for flame stabilization in air-fired and first-generation oxy-combustion systems, here high oxygen concentration in the oxidizer flows is relied upon, as well as on specially designed flame holders. An axisymmetric axial-flow burner also helps avoid flame impingement and particle deposition on the boiler tubes (walls).
A down-fired boiler was chosen over an up-fired for a number of reasons, the most obvious being that an up-fired burner would be prone to bottom ash hitting the burner. Nonetheless, buoyancy-induced recirculation in the flame region of a down-fired boiler would push the flame radially outwards, and could possibly lead to high temperatures in the radiation penetration layer and even flame impingement on the walls. To minimize the impact of buoyancy, the initial section of the boiler was designed as the frustum of a cone, as shown in
The design incorporates a central pure oxygen stream surrounded by fuel in an annulus, which is further surrounded by a secondary oxidizer (SO) with an oxygen concentration that can be varied, and in some embodiments is about 35 vol. %. The use of the central pure oxygen stream helps reduce the flue gas recycle, for a given SO oxygen concentration and overall stoichiometric ratio. Flame holders were designed as circular rings on both the surrounding oxidizer side and the central oxygen side of the fuel annulus. These rings trip both the oxidizer flows at the burner exit and cause small localized recirculation zones at the burner to enhance flame stability and attachment for a wide range of flows, including all the flow conditions studied in this disclosure. Furthermore, in some embodiments, the first 2 m of the boiler wall is refractory to further aid flame stabilization and attachment even during minor fluctuations. The design of the flame holder rings allows for uniform flow in the bulk while locally allowing for mixing, which allows the flam to sit stably on the burner.
With the use of two oxidizer streams, two flames are formed: a high temperature inner flame where the fuel and pure oxygen stream react, and a lower temperature outer flame where the fuel reacts with the surrounding oxidizer. The high number density of fuel particles between the high temperature flame and the wall effectively traps the radiation from this flame, avoiding high wall heat flux. Further downstream, with the dispersion of char and ash particles, part of the radiation from the surrounding oxidizer flame is also trapped. It is important to note here that this design with central oxygen is quite different in both intent and implementation from the use of a central oxygen lance by some atmospheric pressure systems. In the atmospheric pressure systems, an oxygen lance stabilizes the flame at the burner tip and improves boiler efficiency. In such systems, the amount of oxygen used is less than 5% of the total oxygen required for the fuel. On the other hand, in SPOC, the amount of oxygen supplied from the central tube can even be more than the oxygen required by the fuel in that stage. Furthermore, as discussed above, the placement of this large stream of pure oxygen in the center, with the fuel in the annulus, reduces the amount of flue gas recycle required while still maintaining a manageable wall heat flux because of radiative trapping by the coal particles in the annulus.
In some embodiments, the burner has a tapered shape. The tapered shape minimalizes the effects of buoyancy so that no particle deposition on the walls occurs and so that the heat flux is controlled. The tapered shape keeps the Richardson number small throughout the burner, which is needed to minimize buoyance. Also, the flows between the various ports are controlled this way to avoid circulation. These two concepts keep a uniform flow through the combustor section.
Furthermore, the use of fuel-staging, with the boilers connected in series, provides a large amount of gas to the stage as compared to that required for complete combustion of the fuel in the stage. This gas creates a “buffer layer” between the flame and the wall, thus maintaining a low temperature inside the radiation penetration layer for a long distance, enhancing the effectiveness of radiative trapping.
In some embodiments, the first stage of a three-stage SPOC system was analyzed because this is the most challenging stage in terms of managing the heat flux. There are several ways in which the flows of fuel and oxidizer in the three stages are distributed. An exemplary distribution provides an understanding of the 3-stage SPOC system, with all the stages connected in series. In some embodiments, fuel is evenly distributed between each of the three stages (385 MWth each). When the surrounding oxidizer's oxygen concentration is about 35 vol. %, at least some flue gas recycle is required in the first stage. For the later stages, the flue gas from the previous stage is used to control the surrounding oxidizer concentration. Approximately 33% of the total flue gas from the exhaust of the third stage will be recycled to the first stage in this disclosure. This is less than half the amount of recycle incorporated in first-generation approaches, which is typically about 70%. The exact amount of recycle is dependent on the oxygen concentration in the secondary oxidizer and the distribution of the fuel among the various stages, as well as the coal proximate and ultimate analyses.
In some embodiments, PRB coal is used, with the analyses provided in Table 1. The coal is carried by a 1.2 kg/s stream of CO2. The central oxidizer was pure oxygen with a flow rate of 21.7 kg/s for the base case, corresponding to 78% of the total oxygen required for the fuel in this stage. The flow rate of the surrounding oxidizer for the base case was 69.44 kg/s.
In some embodiments, the impacts of several different design parameters and system characteristics related to the burner and boiler on the wall heat flux are:
(1) The length of the conical frustum: By varying the length of the conical section for a given cone angle, the onset of buoyancy-induced internal recirculation in the flame region is controlled.
(2) The geometry of the burner: The sizes of the different affect the velocities of the different streams, and thus the gas mixing and the flame shape. Additionally, they also affect the distance between the flame and the wall, thus the temperature inside the radiation penetration layer. For example, a tapered shape for the burner minimalizes the effects of buoyancy and keeps the Richardson number small throughout the burner.
(3) The amount of central oxygen supplied: For a fixed surrounding oxidizer composition and flow rate, the amount of central oxygen changes both the near-field mixing and the stoichiometric ratio of the stage.
(4) Burner and boiler scaling: Different scaling methods are suitable for different systems. The optimal scaling method for SPOC boilers have been determined.
In some embodiments, ANSYS FLUENT version 13.0 was used to simulate different burner designs. The flow field was modeled using the Reynolds-Averaged Navier-Stokes (RANS) equations with the Semi-Implicit Method for Pressure Linked Equations (SIMPLE). Particle trajectories were computed in the Lagrangian frame and were coupled to the gas phase. The sub-models used in this study are presented in Table 2, together with the particle and wall properties. Further details on the selection criteria for the sub-models and the sensitivity studies of the particle and wall properties are discussed throughout this disclosure. Mesh insensitivity studies, including comparison with a 3D mesh consisting of 10 million cells, showed that a 2D axisymmetric mesh with 150,000 cells was sufficient to model the system.
The use of two oxidizer streams produces a flame structure with two flame fronts, as can be seen from the volatile reaction rate contour in
Note that for such an optically dense, axial flow boiler, especially in the radiation penetration layer, the axial gradients are much smaller than the radial gradients. And, the radiation penetration layer effectively acts like a boundary layer for radiation. Similar to the concept of other boundary layers, the radiation penetration layer is analyzed as a one-dimensional problem. Thus, the radial temperature profiles in this region are sufficient for a qualitative discussion on the impact of various design parameters and system characteristics on the wall heat flux in the following sections.
Parametric Study of Important Design Parameters and System Characteristics
As described above, the first part of the boiler was a conical frustum. The cone angle and the frustum length were designed to avoid buoyancy-induced internal recirculation in the flame region. The cone angle was designed to match the increase in area with the increased volumetric flow rate of the gases.
For a given cone angle, a larger frustum length leads a larger boiler diameter, thus a larger surface area per unit length and a shorter boiler. Although a larger conical frustum may be desired from this perspective, it may also lead to buoyancy-induced internal recirculation. A 10 m conical frustum length was designed as an optimal length (base design), and the impact of increasing the conical frustum length on wall heat flux is demonstrated in this section.
q
w=2∫0τ
Three different cases were simulated to understand the impact of the burner geometry, i.e., the sizes of the different tubes, on the wall heat flux. The size of the central oxygen tube was progressively increased while keeping the overall size of the burner and the hydraulic diameter of the fuel tube fixed. The hydraulic diameter of the surrounding oxidizer (SO) was thus reduced. The sizes chosen were such that in one case the inlet velocity of the central oxygen was higher than that of the SO (Case G1), in the second case they were nearly equal (Case G2), and in the third case the inlet velocity of the SO was higher than that of the central oxygen (Case G3).
The details of the three cases are listed in Table 3. Since the momentum ratios for all cases were such that external recirculation due to flow entrainment was never present, the main impact of changing the size of the burner was on the relative velocities of different streams and the distance between the flame and the wall. The difference in the velocities affects the near field mixing and combustion intensity, while the proximity of the flame to the wall increases the temperature in the radiation penetration layer.
Though the peak heat flux for G1 is similar to G2, the heat flux profile of G1 has multiple peaks. Due to the high central oxygen velocity, the mixing between the central oxygen and the fuel stream increases. This leads to increased combustion intensity and a shorter volatile flame (˜9 m for G1 compared to ˜12 m for G2). The higher combustion intensity combined with the high radial mixing results in a higher temperature on average at short distances from the burner. Hence the temperature in the radiation penetration layer and the heat flux are higher in the initial region for G1. The first peak in the heat flux profile is related to this high mixing, short volatile flame.
In the post-flame region, the heat flux for G1 first reduces and then increases to a second peak, and then finally tails off. The first reduction in the heat flux is due to the end of the volatile flame, which results in a reduction of energy generation in the core. At these distances from the burner, the high velocity of the central oxygen stream dissipates and the mixing is similar to G2. The surrounding gas, on the other hand still has significant momentum to maintain a low temperature radiation penetration layer, especially with the reduced energy generation in the core and reduced mixing. Further downstream, the mixing of the cooler buffer gas near the wall with the hot combustion gases increases and the effective radiation temperature in the radiation penetration layer increases. The second peak in the heat flux profile occurs where the mixing between the hot combustion gases and the buffer gas is almost complete. Even further downstream, the mixing of the hot combustion gases with the colder SO gas for all cases is high enough that the temperature in the radiation penetration layer is similar and the heat flux profiles are similar, but due to lack of substantial energy generation in the core, the heat flux profiles tail off.
Thus, from the geometry cases, it is seen that with radiative trapping the peak heat flux is not strongly dependent on the relative velocities, till the flame is moved very close to the wall. Increased velocity difference between the different inlet streams increases turbulence and thus the reaction rates. The high mixing and combustion rates lead to a flatter radial temperature profile in the near burner region, and hence a higher heat flux in the initial region, followed by a decrease in heat flux due to continued trapping beyond the flame region. Further downstream a second peak in heat flux is also obtained due to buffer layer momentum loss with distance. This type of multiple peaks results in challenges in terms of steam integration.
It is clear from these cases that, to obtain a smooth heat flux profile, it is important to create a long flame away from the wall. Furthermore, a cool buffer gas flow is important to extract the most benefits of radiative trapping. For the purposes of this disclosure, case G2, which has the central oxygen inlet velocity comparable to the other inlet velocities, is used as the base case.
The effect of central oxygen flow was evaluated on the basis of the central stoichiometric ratio (SRcentOx), which is defined as the ratio of the oxygen supplied by only the central oxidizer to the stoichiometric amount of oxygen required by the fuel. Four different central oxygen flow rates were considered, corresponding to central stoichiometric ratios of 0.35, 0.78 (base case), 1.28 and 1.78. With a fixed SO oxygen concentration, the amount of oxygen supplied by the SO affects the flue gas recycle ratio and the buffer gas flow next to the wall. To keep the same flue gas recycle ratio and the buffer gas momentum, the SO flow rate was kept constant. The central oxygen flow, on the other hand, can be varied in the first stage by simultaneously changing the direct oxygen supply to the other stages, without affecting the overall oxygen supply to the three-stage process. In all cases, it was ensured that the total oxygen supplied to the first stage was more than required by the fuel in the stage.
In some embodiments, the central oxygen flow rates have central stoichiometric ratios of from about 0.1 to about 10.0, from about 0.2 to about 7.5, from about 0.3 to about 5.0, or from about 0.35 to about 2.0.
The heat flux for SRCentOx 0.35, on the other hand, closely matches that for SRCentOx 0.78 in the initial region. But, SRcentOx 0.35 has a slightly higher heat flux that continues for a much longer distance because of the much longer flame. The higher heat flux for a long distance is due to the higher temperature in the radiation penetration layer caused by a lower stoichiometric ratio and the increased reliance of the volatile flame on the surrounding oxidizer of the SRcentOx 0.35 case. With small differences in the heat flux profiles between the two lower SRcentOx cases, the case with SRcentOx of 0.78 was chosen as more appropriate due to the higher reliance of the fuel on the inner pure oxygen flame, thereby keeping the volatile flame away from the wall.
With a burner and boiler designed at one utility scale, it is important to be able to easily scale the size of the boiler as per the requirement. Understanding the best scaling method to scale between various utility scales without having to go through all the burner and boiler design considerations highlighted is important. There are three major types of scaling for industrial furnaces and boilers, viz., constant combustion intensity scaling, constant residence time scaling, and constant velocity scaling. The combustion intensity scaling itself can be divided into two methods—constant heat release per unit volume (Q/V scaling), and constant heat release per unit wall surface area (Q/S scaling). For the boiler under consideration, the former is equivalent to the constant residence time scaling, whereas the latter is equivalent to the constant velocity scaling. Thus, in some embodiments of the present disclosure, it is sufficient to consider only two different types of scaling, viz., constant velocity (or Q/S) and constant residence time (or Q/V) scaling.
To understand which is a better scaling method for SPOC, scaling the thermal input by a factor of ‘f’ was considered. For a co-axial flame, as in SPOC, with a Q/S scaling, the flame length scales similar to the boiler length, both by a factor off f0.5 With Q/V scaling, the flame length scales as f, while the boiler length scales by a factor of f0.5. The difference in the scaling of flame length and boiler length for the different scaling methods leads to differences in the location of the peak heat flux. On a normalized length scale, the peak heat flux location for the Q/S scaling doesn't change, whereas for the Q/V scaling method, the peak heat flux location shifts downstream when scaling up and upstream when scaling down.
To show this quantitatively, the constant velocity (Q/S) scaling with a constant residence time (Q/V) scaling was done when scaling between two utility scales—385 MWth and 770 MWth (per SPOC stage). These represent a three-stage SPOC power plant with a nominal power output of 450 MWel and 900 MWel, respectively. The purpose of this comparison is to study the sensitivity of varying the size of the boiler unit on the heat flux profile.
With the constant Q/S scaling method, not just the location of the peak heat flux, but also the value of the heat flux everywhere matches the base case almost exactly. This is because for the SPOC boilers, which are optically dense, heat transfer to the wall is mainly dependent on the temperature in the radiation penetration layer, which is very close to the boiler wall, a larger flame does not lead to a higher emission, as in atmospheric pressure boilers.
This shows that a constant velocity based scaling, which also ensures a relatively constant surface area per unit thermal input, results in similar heat fluxes for such axial-flow, optically dense systems. It further shows that the size of the SPOC boilers is able to be modified based on the final requirement, without having to optimize the burner and boiler design again at the new scale.
The present disclosure thus provides a practical and flexible design for the first stage boiler of a 3-stage SPOC power plant. Because the total oxygen supplied to this stage is high, potentially high wall heat fluxes are expected. The burner and boiler designs are based on the philosophy of ensuring a relatively flat and high heat flux profile, below the material constraint limitations of the boiler tubes. For additional flexibility in operation, in some embodiments an axial flow burner with pure oxygen in the center, surrounded by a fuel annulus (coal carried by CO2), which is in turn surrounded by a secondary oxidizer is designed. The oxygen concentration is varied in the surrounding oxidizer, and in some embodiments is about 35% (by vol.). In this design, because of combustion of fuel with pure oxygen in some regions, locally the temperature is very high. Unlike a traditional approach of controlling heat flux by controlling the temperature, SPOC utilizes radiative trapping. Due to the large number concentration of particles from the fuel tube, the use of pure oxygen in the center, surrounded by a fuel annulus, ensured a high optical thickness of the medium between the pure oxygen-volatile flame and the wall. The high optical thickness, combined with a low temperature “buffer gas” near the wall in the region from which radiation penetrates to the wall (radiation penetration layer), is used to control the heat flux. The conical design of the initial section of the boiler ensures a low Richardson number over its length and avoids buoyancy-induced recirculation, which causes a high temperature in the radiation penetration layer or even flame impingement on the wall, and hence a high heat flux. The frustum is sized to avoid such recirculation.
Other burner and boiler parameters are disclosed to understand their impact on the heat flux. The effect of both burner geometry and central oxygen flow rate affects the peak heat flux only slightly. But, with increasing velocity difference due to either reduced central oxygen tube size or increased flow, the combustion intensity and radial mixing both increased. This causes differences in the heat flux profiles in some local regions close to the burner. These differences are explained using temperature contours, and radial profiles of temperature and optical thickness. In cases with increased central oxygen velocity, the heat flux profile has a double peak structure, due to competing effects of high combustion intensity and mixing, and radiative trapping. This poses some challenge to steam integration. To avoid this, conditions with similar velocities of the different streams are used as the base design.
Finally, in some embodiments, using the results from these analyses and the principles for the design of such a burner and boiler, a constant velocity based scaling shows the effectiveness of the principles at various scales of utility boiler units.
Two scales were chosen which correspond to a 420 MWel and 900 MWel unit. Scaling by this method ensures the heat flux profile remains the same at different scales. This is because such a scaling ensures the heat transfer surface area scaled with the thermal input, and hence the length of the radiating region. It also ensures similar near-field aerodynamics, avoiding any non-smoothness in the heat flux profile. A comparison with a constant residence time scaling, further proves this. Thus, the present disclosure provides a new, flexible, staged, pressurized oxy-combustion boiler design. And by the systematic study of several burner and boiler parameters, its resilience to the different firing conditions is shown. The present disclosure further discloses the most ideal conditions for operation to ensure a high and relatively flat heat flux profile. Furthermore, the present disclosure provides a toolbox which is utilized to control the heat flux profile in a pressurized boiler with localized high temperatures. Using scaling analysis, the present disclosure demonstrates the effectiveness of this design when scaled using the constant velocity scaling method.
In some embodiments, ANSYS FLUENT version 13.0 was used to simulate different burner designs.
The flow field was modeled using the Reynolds-Averaged Navier-Stokes (RANS) equations with the Semi-Implicit Method for Pressure Linked Equations (SIMPLE) algorithm to address pressure-velocity coupling. RANS has been used extensively to study the flame behaviors of pulverized fuel. Previously, none of the solid fuel combustion submodels have been validated for pressurized oxy-combustion environment. In the present disclosure, sub-models that have been extensively validated under atmospheric pressure oxy-coal combustion environment are used. It has been shown that RANS is able to provide reasonable agreement between experiments and CFD simulations under atmospheric pressure oxy-combustion conditions. The shear stress transport (SST) k-w model was used in this simulation, which has been shown to be the combination of the best aspects of the two standard models, κ-ε and κ-ω). The finite rate/eddy dissipation model was used for turbulence-chemistry interactions.
A two-step global reaction mechanism was used to capture the gaseous volatile reaction. The first step was volatile oxidation to carbon monoxide, and the second step was oxidation of carbon monoxide to carbon dioxide. More detailed reaction mechanisms together with a more advanced reaction model, i.e., eddy dissipation concept (EDC) model have also been tested. The EDC model takes account of the CO2 and H2O dissociation and thus gives a lower flame temperature and slightly lower peak wall heat flux than the finite rate/eddy dissipation model. In the present disclosure, the more conservative results from the finite rate/eddy dissipation model are presented. Particle trajectories were computed in the Lagrangian frame and were coupled to the gas phase. All discrete particles were assumed to be spherical. The pulverized coal particle size was assumed to follow the Rosin-rammler distribution, with an average diameter of 65 μm. The minimum and maximum particle sizes were 10 μm and 200 μm, respectively, and the spread parameter of the distribution was 3.5. The effect of turbulence on the particle trajectories was accounted for with the Discrete Random Walk (DRW) model.
The devolatilization rate of the coal particles was calculated using the Chemical Percolation Devolatilization (CPD) model. NMR chemical structure parameters were calculated according to known correlations. The char surface oxidation rate with oxygen was modeled by a kinetics/diffusion-limited model. A multiple-surface-reaction char model, which includes the reactions between char particles and oxygen, carbon dioxide, and water vapor, was also considered for comparison. The simulation results for these two models were essentially the same, indicating that char oxidation is dominant in this boiler condition.
The Discrete Ordinates (DO) radiation model was used to solve the radiative transfer equation (RTE). Each octant of the entire angular space 411 at any spatial location was discretized into 5×5 solid angles (i.e. control angles), leading to a total of 200 directions. Further refined discretization did not show any noticeable improvement in terms of the distributions of temperature and incident radiation. Gaseous emission and absorption were considered using the domain-based Weighted Sum of Gray Gases Model (WSGGM). A particle emissivity of 0.6 was assumed for this study. The emissivity of coal particles is approximately 1, while that of fly ash is 0.6. Based on a sensitivity analysis, it was found that, counterintuitively, a lower emissivity leads to a higher wall heat flux due to radiative trapping.
Thus, a conservative value of 0.6 was chosen so that the results reported here will yield the maximum incident radiation at the wall. The particle scattering factor was set to be 0.6. The surface temperature of the boiler wall was assumed to be 700 K. A sensitivity analysis of wall temperature was conducted and showed that a variation from 100 K to 1000 K produced only about an 8% difference in wall heat flux. For wall emissivity, it was assumed that during operation the walls would accumulate a small amount of ash. Depending on the properties of the ash, and whether the deposit is loose, sintered or fused, values between 0.37 to 0.93 have been reported in the literature.
In the present embodiment, a relatively conservative value of 0.8 was chosen for the emissivity of the wall. A high emissivity implies a less reflective wall, and hence a higher net heat flux to the wall. A 2-D axisymmetric model was built based on the burner and boiler geometry, taking advantage of the axisymmetric nature of the boiler. Simulation results for cases with 30,000, 150,000 and 300,000 cells were compared. Negligible differences in velocity, temperature and heat transfer were observed for the last two cases. Moreover, a 3-D model with 10 million cells was constructed, and the difference between the results of the 3-D model and those of the 2-D model with 150,000 cells was negligible. Thus, only results for the 2-D model with 150,000 cells are reported.
SPOC Boiler with Low Flue Gas Recycle—Operational Flexibility
Oxy-fuel combustion is one of the leading strategies for capturing carbon dioxide (CO2) from stationary combustion furnaces and power plants. Compared to existing “first-generation” oxy-combustion technologies, which operate at atmospheric pressure, pressurized oxy-combustion offers a unique advantage. Elevating the gas pressure leads to an increased dew point of flue gas moisture. Thus, the flue gas moisture can be condensed at a higher temperature than in atmospheric pressure systems, and the latent heat of condensation can be utilized in the steam cycle. With a proper design, pressurized oxy-combustion can significantly increase process efficiency. The impact of pressure on the efficiency of various configurations of pressurized oxy-combustion has been studied, and it has been shown that, depending on the actual configuration, the pressure for optimal plant performance ranges from about 10 bar to 16 bar. Also, since the flue gas, which is mainly CO2 after moisture condensation, is usually required at very high pressures (˜150 bar) for transportation, utilization, and storage, increasing the pressure of the combustion process does not impose efficiency or economic costs. On the other hand, due to the higher pressure and hence the smaller volumetric flow rates, the gas handling and cleanup equipment can be smaller. In addition to pressurizing the combustion process, reducing flue gas recycle can also improve plant performance and cost.
In order to maintain boiler tube temperatures within material constraints, the heat flux to the boiler tubes must be controlled. In a typical oxy-combustion design, this is achieved by controlling the flame and flue gas temperature by mixing the inlet oxygen with large amounts (˜70% by volume) of recycled flue gas. Such a high flue gas recycle is costly: There are operating and capital costs for the recycling equipment and an associated performance loss from the fan's power consumption. Use of additional water or steam to control the gas temperature and hence the heat flux has also been shown to worsen the overall performance of the process because not all the latent heat of condensation can be effectively recovered. Even if all the latent heat were to be recovered, performance would suffer from additional water injection into the boiler, due to the difference in exergy between the moisture vaporizing in the boiler (heat transfer with high temperature difference) and the condensation in the recovery unit (heat transfer with lower temperature difference). A comparison of various power plant configurations—air-fired, atmospheric oxy-combustion, and two different pressurized oxy-combustion processes—using energy and exergy analysis shows that the plant configuration can have a significant impact on efficiency, and a low-recycle, dry-coal feeding plant configuration, as in SPOC, would have very high efficiency.
In some embodiments, the SPOC process is a high efficiency, low recycle pressurized oxy-combustion process, with the boilers optimized for operation at from about 10 bar to about 20 bar, from about 12 bar to about 18 bar, or about 16 bar. In some embodiments, it comprises multiple boilers (or stages) connected in series, with the fuel and oxidizer distributed to the different stages. An exemplified process flow diagram is shown in
Apart from cost-effectively capturing carbon dioxide, another major challenge faced by power generation systems today is an increased variability in demand due to the high market penetration of renewable sources. Coal and combined cycle gas power plants, which were historically considered as baseload generators requiring minimal cycling, are now being subjected to significant cycling. New plant designs must focus on being flexible to operate at various reduced plant loads, i.e., turndowns. The impact of flexible operation on carbon capture processes, has been studied from both process and boiler operation perspectives.
For the SPOC process, many plant units such as the air separation unit and CO2 purification unit, require turndown considerations similar to those for other oxy-combustion plants. Some other aspects of the SPOC process, such as low temperature NOx removal (instead of a high temperature SCR/SNCR based process), are advantageous for flexible operation. In the SPOC process, with a proper distribution of oxygen and fuel between the stages, the last two stages can operate under favorable conditions during turndown. For example, if a turndown to 66% thermal input were desired, a single stage system would undergo a big change in the heat flux profile. On the other hand, due to the similarity of the exhaust gas characteristics in stages 2 and 3, the SPOC process could operate by only shutting-down stage 3, without affecting the heat flux profiles in the first two stages and hence the steam integration.
As discussed above, the SPOC system design provides additional flexibility in plant operation and the opportunity to control the oxygen concentration near the boiler tubes. In some embodiments, alternative designs are present compared to those discussed above and provide better control heat flux and boiler exhaust temperatures under varying thermal loads.
The main philosophy of the design discussed above is based on creating a long flame to evenly distribute the wall heat flux, while utilizing radiative trapping to control the heat flux below material limitations, thereby enabling reduced flue gas recycling. Radiative trapping occurs in an optically dense system where the temperature is high in the core of the system but relatively low near the wall. In this case, the radiation emitted from the high temperature region is effectively trapped in the core region. The incident radiation on the wall is mainly determined by the temperature in the radiation penetration layer, which is defined as the region from the wall up to an optical thickness (τw) of 2.3. In some embodiments, a burner is designed to keep the temperature in the radiation penetration layer low, and the medium inside the boiler is optically dense, the wall heat flux is low, irrespective of the temperature in the core of the boiler. This is analogous to the sun, where the core temperature is about 15.7 million K but the radiation temperature of the sun is only about 6000 K. The radiation from the core is trapped because of a high optical thickness, and only the radiation from the photosphere is able to penetrate through and leave the sun.
In the SPOC system, due to high pressure and low recycle, the particle number concentration is very high, leading to a large optical thickness. The low recycle leads to a high flame temperature in the core of the boiler, but with a strategic design of the burner and boiler, the temperature in the radiation penetration layer is kept low, thereby keeping the heat flux to the boiler tubes within material constraints. Furthermore, the burner and boiler design results in a long flame, with radiative trapping controlling the heat flux for a long distance and maintaining a relatively constant and high heat flux, within material constraints. This high heat flux reduces the total heat transfer surface area required and hence the cost.
The SPOC boiler system has many flexibilities in terms of the number of stages, and fuel and oxidizer distributions among different stages. In some embodiments of the present disclosure, a three stage SPOC system is disclosed with an even distribution of fuel between the stages (385 MWth each). Constraints and requirements lead to a design with a down-fired, axisymmetric, cylindrical boiler, with the initial section designed as the frustum of a cone. The schematic of the burner and boiler from is shown in
In some embodiments, an axial-flow burner is chosen to reduce mixing between the fuel and the oxidizer and thus creates a long flame. It also avoids flame impingement and minimizes particle impaction on walls. The burner is made up of a central oxygen tube, surrounded by a fuel annulus, which is itself surrounded by a flow of secondary oxidizer (SO). In the present disclosure, the fuel is carried by a small flow of pure CO2, and the secondary oxidizer composition is about 35% O2 and about 65% CO2 by volume. Flame holders on both SO and central oxygen side are incorporated for flame attachment at the burner. The first 2 m of the conical section is designed as a refractory wall to ensure flame attachment and stabilization, even during minor, unexpected burps or fuel clogs. The benefits are similar to those observed in industrial furnaces which have refractory walls.
With the use of two oxidizer streams, two flames are formed: a high temperature, fuel and pure oxygen inner flame, and a lower temperature, fuel and SO outer flame. The high number density of fuel particles between the high temperature flame and the wall effectively traps the radiation from this high temperature flame before it reaches the wall, avoiding high wall heat flux. Further downstream, with the dispersion of char and ash particles, part of the radiation from the lower temperature flame is also trapped. The central oxidizer (pure oxygen) delivers about 80% of the total oxygen required for the fuel in the first stage. The total amount of oxygen supplied to the first stage is more than required for the fuel in the stage. Since most of the oxygen for the first stage is added as central oxygen, and flue gas recycle is used only in stage 1 to control surrounding oxidizer oxygen concentration, the total flue gas recycle ratio is very low. The exact amount depends on the surrounding oxidizer oxygen concentration, the configuration of the three stages, and the fuel properties. The fuel and oxidizer composition, and flow rates through the various tubes used in this disclosure for the base or designed condition (100% thermal input) are listed in Table 4.
The design and flow conditions described above are considered as the baseline for analyzing the performance of the SPOC boiler under turndown conditions. A number of different turndowns were tested, including a turndown to about 40% thermal load. Other turndowns include, but are not limited to, about 90% thermal load, about 80% thermal load, about 70% thermal load, about 60% thermal load, about 50% thermal load, about 40% thermal load, about 30% thermal load, about 20% thermal load, about 10% thermal load, about 5% thermal load and about 1% thermal load.
ANSYS FLUENT version 13.0 was used to simulate different burner designs. The flow field was modeled using the Reynolds-Averaged Navier-Stokes (RANS) equations with the Semi-Implicit Method for Pressure Linked Equations (SIMPLE). Particle trajectories were computed in the Lagrangian frame and were coupled to the gas phase. The sub-models used in this study are presented in Table 5, together with the particle and wall properties. Further details on the selection criteria for the sub-models and the sensitivity studies of the particle and wall properties are discussed throughout the disclosure. In some embodiments, mesh insensitivity studies, including comparison with a 3D mesh consisting of 10 million cells, showed that a 2D axisymmetric mesh with 150,000 cells was sufficient to model the system.
indicates data missing or illegible when filed
The base case, or 100% thermal input case (T100), was chosen as an exemplary design case. The radiation and total wall heat flux for the base case T100 and the turndown case, T40 (40% thermal input), are shown in
The first 30 m of the temperature contours for the two cases are shown in
Even though properly designed for 100% thermal input, two main aspects of the design need to be adapted for reduced thermal input: avoiding buoyancy induced adverse pressure gradients, and having a cold buffer gas flow over a long distance from the burner. One way to achieve both with the same tool is discussed below.
In some embodiments, to enhance the effect of radiative trapping for reduced thermal input, the high temperature gases must be “focused” to the center (axis) of the reactor, and not allowed to spread very quickly to the walls. Also, advection should be kept high, even at the reduced flow rates corresponding to the turndown conditions. To achieve this, in some embodiments, small obstructions in the form of rings, are added to the wall so that small, localized re-circulations are formed in their wake next to the wall, which effectively pushes the remaining flow away from the wall, keeping the near wall region relatively cold for a longer distance. The displaced flow essentially increases the advection in the main flow, reducing the effect of buoyancy forces, and consequently the bulging of the flame towards the wall the added advantage of such a modification is that the size (diameter) of the boiler is increased further without causing buoyancy-induced internal recirculation in the flame region, as was observed upon increasing the boiler diameter for the base design discussed in alternative embodiments of the present disclosure.
In some embodiments of the present disclosure, the boiler includes rings with a rectangular cross-section on the wall of the boiler, starting at about 2 m downstream of the burner and repeating about every 1 m up to 31 m from the burner. In some embodiments, the wall rings are present from about 0.1 m, about 0.5 m, about 1 m, about 2 m, about 3 m about, 5 m, about 7 m or about 10 m downstream of the burner. In some embodiments, the wall rings repeat from about 0.5 m to about 50 m from the burner, including various lengths in between.
In some embodiments, each wall ring includes a radial projection from about 0.05 m to about 5 m, from about 0.1 m to about 3 m. In some embodiments, each wall ring has a thickness of from about 0.01 to about 0.1 m, from about 0.02 to about 0.05 m. In some embodiments, each wall ring has a radial projection of about 0.1 m and a thickness of about 0.0254 m. In some embodiments, the rings are made of refractory material or water-cooled tubes. In some embodiments, the rings are modeled as adiabatic walls. Since the surface area of the rings is negligible compared with the total surface are of the wall, the boundary conditions of the rings have little impact on the overall heat flux profile. An exemplary schematic of a boiler with these wall rings is shown in
The radiative and total heat flux profiles for the 100% and 40% thermal input cases with and without the wall rings are shown in
For the turndown (T40) case with rings, the total heat flux peak is reduced by more than 100 kW/m2 compared to without rings. The location of the peak heat flux shifts downstream as well, due to the reduced impact of buoyancy. Similarly, the use of rings reduces the radiative heat flux for the T100 case in the flame region. But with increased convection, the total heat fluxes in the T100 cases are similar with and without rings. However, for the T40 case, since the peak heat flux was caused primarily by high temperatures in the radiation penetration layer, the reduction in radiative heat flux surpasses the modest increase in convective heat flux, and the total is lower with rings than without. In the post-flame region, the heat flux is slightly higher, especially for the T40 case, because of different flue gas temperatures caused by different upstream heat transfer rates to the boiler tubes.
The effect of the rings on the heat flux profile is understood by analyzing the way rings affect the local and global fluid dynamics. Locally, the rings create small recirculation zones, which enhance local mixing. Globally, these small recirculation zones push the main flow towards the core (axis) of the boiler since the near-wall gases start at a lower temperature and there is limited spread of energy from the core of the boiler to these near-wall gases, the enhanced mixing creates a thicker buffer layer of “cold” gases compared with the case without rings, leading to a lower heat flux. Radial temperature and optical thickness profiles provide a more quantitative picture of the effect of wall rings on heat flux.
q
w=2∫0τ
The boiler diameter in some embodiments is constrained to 2.83 m by the need to maintain a high Richardson number in order to avoid buoyancy-induced internal recirculation in the flame region. With the addition of wall rings, and the associated increase in advection, the cone length is increased without causing internal recirculation. In some embodiments, the boiler was extended up to about 4 m in diameter by extending the conical frustum length to about 25 m from the burner. This diameter was chosen such that off-site manufacturing and shipping of the boilers on railroads would be possible. In some embodiments, the diameter is changed with evolving capabilities for shipping the boilers.
Because the exhaust flue gas temperature must be reduced to an acceptable value before convective tubes can be added, the length of the radiant boiler is dependent on the total heat transferred per unit length. In some embodiments, increasing the diameter of the boiler helps reduce its overall length. In some embodiments, this is achieved by creating a second expansion downstream of the flame and high temperature region, where the gases are sufficiently mixed. The design with rings enables a continued expansion of the conical frustum and hence a larger diameter boiler is designed with less loss of surface area and a simpler construction. The rings, especially in the flame region, also improve the performance of the boiler under varying thermal loads. The number, size and shape of these rings are not limited by this disclosure. That is, the number, size and shape of these rings vary depending upon the desires and objectives of the user.
The present disclosure discusses the impact of low-thermal load operation of SPOC boilers on wall heat flux. The first stage of the three-stage process was analyzed because it was considered as the most challenging one in terms of controlling the heat flux. The results from the analysis of this stage are utilized in the design of the other stages as well. Since SPOC boilers use radiative trapping to control heat flux, importance was placed on the effect of turndown on the overall fluid dynamics, especially in relation to the impact on the temperature in the radiation penetration layer. The turndown considered in this disclosure was 40% of the designed thermal load, and the flame characteristics and heat flux profiles were studied. By shutting down the last stage boiler, the multi-stage approach of SPOC effectively achieved equivalent turndowns with lower impact on heat flux profiles, and hence reduced difficulties in steam integration, compared to a single stage process.
In terms of boiler design, the restriction on the length of the conical section due to buoyancy-induced recirculation, and the excessive peak heat flux during turndown operations were addressed with the use of wall rings. The wall rings enhanced the effectiveness of radiative trapping in the flame region, and also enhanced advection. The increased advection allowed increasing the size of the conical frustum section from about 10 m to about 25 m, so that the boiler diameter was about 4 m, without causing an increase in heat flux due to buoyancy-induced recirculation. The increased boiler diameter, on the other hand, reduced the peak heat flux under the designed firing conditions. With a 40% turndown, even though the buoyancy-induced internal recirculation occurs, it was far downstream of the volatile flame, and had no impact on the shape of the high temperature core. Thus, the temperature in the radiation penetration layer did not increase and the heat flux in the flame region was similar with and without expansion. With the final boiler design, incorporating wall rings and a longer conical frustum section, the resulting heat flux profile under both thermal inputs presented here was less than 500 kW/m2, thereby enabling the use of standard boiler tube material everywhere.
In view of the above, it will be seen that the several advantages of the disclosure are achieved and other advantageous results attained. As various changes could be made in the above processes and composites without departing from the scope of the disclosure, it is intended that all matter contained in the above description and shown in the accompanying drawings shall be interpreted as illustrative and not in a limiting sense.
When introducing elements of the present disclosure or the various versions, embodiment(s) or aspects thereof, the articles “a”, “an”, “the” and “said” are intended to mean that there are one or more of the elements. It is also noted that the terms “comprising”, “including”, “having” or “containing” are intended to be open and permits the inclusion of additional elements or steps.
This application claims the benefit of U.S. Provisional Patent Application Ser. No. 62/464,159, filed Feb. 27, 2017, the entire contents of which are incorporated herein by reference.
This invention was made with government support under Grant # DE-FE0009702, awarded by the Department of Energy. The Government has certain rights in the invention.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/US2018/019942 | 2/27/2018 | WO | 00 |
| Number | Date | Country | |
|---|---|---|---|
| 62464159 | Feb 2017 | US |
