**4. Results**

### **4.1 Numerical settings**

The box/EWB model and each of the 6 WSGG models are applied to each of the 2 oxy-fuel environments. As mentioned in subsections 2.1 and 2.2, there are 22 RTEs per direction to resolve the spectrum for the box/EWB approach, and either 4 or 5 RTEs per direction to resolve the spectrum for the WSGG approach. We use the finite-volume method for the both the spatial and directional discretizations. As mentioned earlier in section 3, the enclosure is discretized into 59 778 cells. In each cell, the 3D angular space of 4*π* is divided into 128 angular divisions. A coarse representation of such angular discretization for a hemisphere (angular space of 2*π*) is shown in Figure 3. We have performed sensitivity analyses to check the suitability of both linear and angular resolutions by comparing a solution obtained using the aforementioned ones with a solution obtained using a finer linear resolution (33×33×110 = 119 790 cells) while keeping the angular resolution unchanged; and with a solutions obtained using a finer angular resolution (200 divisions) while keeping the linear resolution unchanged. In both situations, the solutions are nearly identical, and thus the adopted resolutions are considered sufficient. The nongray radiation simulation is performed iteratively using the computational fluid dynamics software ANSYS FLUENT 13.0 (30). None of the radiative-property models described here are available in the standard release. We implemented each method through a user-defined function that is complied and linked to the software for run-time access.

Fig. 3. Sketch illustrating the angular finite-volume discretization in a hemisphere

#### **4.2 Order of presentation**

In the four subsequent subsections, the solutions of the radiative solution for the two oxy-fuel environments are presented. We first start in subsection 4.3 with 2D flooded contours of the 8 Numerical Simulations / Book 1

The geometry of both problems is a large rectangular enclosure, with dimensions 12×12×40 m. The medium temperature is 1 500 K. The temperature of the walls is kept at 750 K, with an emissivity of 0.725. This configuration was proposed by Krishnamoorthy et al. (25) to roughly represent the dimensions of a full-scale 300 MW front-wall-fired, pulverized-coal, utility boiler (29). The domain is discretized with a uniform mesh of

The box/EWB model and each of the 6 WSGG models are applied to each of the 2 oxy-fuel environments. As mentioned in subsections 2.1 and 2.2, there are 22 RTEs per direction to resolve the spectrum for the box/EWB approach, and either 4 or 5 RTEs per direction to resolve the spectrum for the WSGG approach. We use the finite-volume method for the both the spatial and directional discretizations. As mentioned earlier in section 3, the enclosure is discretized into 59 778 cells. In each cell, the 3D angular space of 4*π* is divided into 128 angular divisions. A coarse representation of such angular discretization for a hemisphere (angular space of 2*π*) is shown in Figure 3. We have performed sensitivity analyses to check the suitability of both linear and angular resolutions by comparing a solution obtained using the aforementioned ones with a solution obtained using a finer linear resolution (33×33×110 = 119 790 cells) while keeping the angular resolution unchanged; and with a solutions obtained using a finer angular resolution (200 divisions) while keeping the linear resolution unchanged. In both situations, the solutions are nearly identical, and thus the adopted resolutions are considered sufficient. The nongray radiation simulation is performed iteratively using the computational fluid dynamics software ANSYS FLUENT 13.0 (30). None of the radiative-property models described here are available in the standard release. We implemented each method through a user-defined function that is complied and linked to the

Fig. 3. Sketch illustrating the angular finite-volume discretization in a hemisphere

In the four subsequent subsections, the solutions of the radiative solution for the two oxy-fuel environments are presented. We first start in subsection 4.3 with 2D flooded contours of the

27×27×82 cells, resulting in a total of 59 778 hexahedral cells.

**4. Results**

**4.1 Numerical settings**

software for run-time access.

**4.2 Order of presentation**

radiative source term (in kW/m3) along the 12×40 vertical midplane (the symmetry plane midway between the two vertical side walls separated by a distance of 12 m). Due to the symmetry of the problem, this plane should be identical to the horizontal symmetry plane. Next, the 1D profiles of this radiative source term along the centerline of the enclosure (i.e., the 40-m longitudinal line passing through the geometric center of the 12×12 cross-section of the enclosure) are presented in subsection 4.4. In these profiles, we also include published results (25) using the SLW approach.

The SLW approach (originally proposed by Denison and Webb (31)) is a more-rigorous implementation of the WSGGM. The individual gray gases now have a physical meaning and direct mathematical relationship with the absorption spectrum (in terms of the absorption cross-section, whose SI unit is m2/mol). The range of the absorption coefficient is divided into segments, each of which represents an absorbing/emitting gray gas. In addition, there is one clear gas (as in the WSGG approach). The segmentation of the range of absorption coefficient is typically done such that their logarithmic values are equally spaced. For each segment (i.e., each absorbing/emitting gray gas), a logarithmic average absorption cross-section *Ci* is assigned, and the corresponding blackbody weight *ai* is evaluated to be the fraction of the Planck function that belongs to the range of absorption coefficient of the segment represented by the *i th* gray gas. The linear absorption coefficient for a species is related to the absorption cross-section by the species molar concentration (its SI units is kmol/m3). The exact implementation of this method would require the processing of a high-resolution spectrum (which incurs the processing of millions of spectral data points at high combustion or flue temperatures), the computations are highly simplified by utilizing a fitted hyperbolic tangent function for the cumulative distribution of the absorption cross-section, which is known as the absorption-line blackbody distribution function (ALBDF) (32; 33).

Several different approaches have been developed to apply the SLW method to multicomponent gas mixtures. The approaches are derived using different assumptions and vary in computational cost and accuracy. For the SLW solution we include here, the absorption cross-section domains of H2O and CO2 were individually discretized into 20 logarithmically-spaced intervals between 3×10−<sup>5</sup> <sup>m</sup>2/mol and 120 m2/mol for H2O, and between 3×10−<sup>5</sup> <sup>m</sup>2/mol and 600 m2/mol for CO2. The analytical expressions for the absorption-line blackbody distribution functions of H2O (32) and CO2 (33) were used to compute the blackbody weights of each gray gas. The multiplication method (34) was used to handle the presence of a mixture. Implied in this method, is the assumption that the absorption cross-sections of H2O and CO2 are statistically independent. The number of RTEs per direction was 21 (one RTE per each of the 20 gray gases plus an RTE for the clear gas). The SLW calculations were performed using the T4 angular quadrature (35), and using the same spatial resolution we employed for the other two approaches (namely 27×27×82) and using a similar angular resolution (128 directions).

Unlike the box/EWB and WSGG solutions, in which we use the angular finite-volume method for treating the angular dependence of radiation, the SLW solutions were obtained using the discrete-ordinate method. Whereas both methods have some similarity, the angular finite-volume method conserves the radiative energy (4) and thus is considered a more accurate method for handling the directional dependence of radiation. In addition, the analytical fits for ALBDF of H2O and CO2 are based on an extension of an old version (1991/1992) of the spectral database HITRAN (36). This database was assembled for a (low) temperature of 296 K and thus when applied at high temperatures the absorption of the

box/EWB

4g, quad

5g, quad

5g, quad, (cont)

4g, linear

5g, cubic

4g, cubic, (air)

solution being the worst.

Fig. 4. Midplane radiative source (left: 65%CO2; right: 90%CO2)

deviation from the box/EWB solution, computed as

the flat portion of the radiative-source curve is smallest in the case of the box/EWB solution. As suggested from the 2D contours in the previous subsection, the air-fuel and the 5-gas/cubic WSGGM solutions show noticeable overprediction of the radiative source, with the air-fuel

Tables 5 and 6 list the values of the radiative source at the middle of the profiles (which corresponds to the centroid of the 3D enclosure) for the various solutions, with the relative

box/EWB <sup>×</sup> 100% (13)

Percent error <sup>=</sup> SLW/WSGG <sup>−</sup> box/EWB

**Source [kw/m<sup>3</sup>**

**]: -100 -90 -80 -70 -60 -50 -40 -30 -20 -10 Source [kw/m<sup>3</sup>**

Nongray EWB and WSGG Radiation Modeling in Oxy-Fuel Environments 503

**]: -100 -90 -80 -70 -60 -50 -40 -30 -20 -10**

medium will be underpredicted because many *hot lines* (i.e., transitions from excited vibration levels) are missing (4; 32; 37). Whereas a procedure (37) was followed to extend the original database by generating *hot-line* estimates from *cold-lines* (i.e., transitions from the ground level), Modest (4) showed that these analytical expressions result in nontrivial deviations from LBL calculations at 2 000 K. However, at 1 000 K, they are in good agreement with the LBL solution. On the other hand, the SLW approach does not require the specification of a pathlength as the EWB approach.

In the legends, the different WSGG models are designated by the total number of radiating and clear gases (either 4 or 5) and the temperature-polynomial order (linear, quadratic, or cubic). Since the WSGGM in reference (22) and the one in reference (24) have the same number of gray gases and the same polynomial order (5 gases and quadratic polynomials), we add a suffix *(cont)* to the WSGGM in reference (24) to highlight that its parameters are continuous functions of the H2O/CO2 molar ratio. Also, the air-fuel WSGGM here (28) is further distinguished by adding the suffix *(air)* to its legend entry.

The following subsection, number 4.5, shows also 1D profiles, but for the to-wall radiative flux (in kW/m2) along the 40-m longitudinal midline of the top 12×40 wall of the enclosure. Due to the symmetry of the problem, this should be identical to any midline on the other three 12×40 walls.

The final subsection, number 4.6, is dedicated to the area-integrated radiative heat transfer rate to the walls (in MW). This subsection provides a quantitative measure of the variation among the different solutions with regard to the total radiative heat transfer rate (in MW) to the walls of the enclosure. The area-integrated heat transfer is an important quantity when we are concerned about the operation of the furnace unit within the boiler, as this effects steam generation rate. The average radiative heat flux is calculated from this quantity by dividing area-integrated heat transfer rate by the total surface area of the walls (2 208 m2), and is also included in the comparison tables. This quantity provides a geometry-independent measure of the radiative heat load in oxy-fuel furnaces. The deviations from the benchmark box/EWB solution are also included. One table is provided per oxy-fuel environment.

#### **4.3 Radiative-source contours**

Slices of the radiative source term along the 12×40 plane of symmetry are shown in Figure 4. Each figure corresponds to a different model, with a plot for each of the two oxy-fuel environments. The number and values of the contour levels are the same for the plots. The double-symmetric pattern in all plots is expected. The calculated negative value of the radiative heat source would *drive* the temperature field to lower values in a coupled simulation. The radiative source is smallest near the colder-than-medium walls; it increases steeply and becomes nearly flat over a large portion of the plane. Notice that this value is very similar for both environments. The box/EWB solution exhibits a smaller decrease of the radiative source near the walls than the WSGG solutions. For both environments, the air-fuel WSGGM (28) and the 5-gas/cubic WSGGM (27) show noticeable overprediction of the radiative source, which indicates a weaker influence of radiation on the thermal field.

#### **4.4 Radiative-source profiles**

Profiles of the radiative source term along the longitudinal centerline of the 3D enclosure are compared in Figure 5 for the two oxy-fuel environments. As mentioned in subsection 4.2, we also add published profiles (25) predicted using the SLW approach. For both environments, 10 Numerical Simulations / Book 1

medium will be underpredicted because many *hot lines* (i.e., transitions from excited vibration levels) are missing (4; 32; 37). Whereas a procedure (37) was followed to extend the original database by generating *hot-line* estimates from *cold-lines* (i.e., transitions from the ground level), Modest (4) showed that these analytical expressions result in nontrivial deviations from LBL calculations at 2 000 K. However, at 1 000 K, they are in good agreement with the LBL solution. On the other hand, the SLW approach does not require the specification of a

In the legends, the different WSGG models are designated by the total number of radiating and clear gases (either 4 or 5) and the temperature-polynomial order (linear, quadratic, or cubic). Since the WSGGM in reference (22) and the one in reference (24) have the same number of gray gases and the same polynomial order (5 gases and quadratic polynomials), we add a suffix *(cont)* to the WSGGM in reference (24) to highlight that its parameters are continuous functions of the H2O/CO2 molar ratio. Also, the air-fuel WSGGM here (28) is

The following subsection, number 4.5, shows also 1D profiles, but for the to-wall radiative flux (in kW/m2) along the 40-m longitudinal midline of the top 12×40 wall of the enclosure. Due to the symmetry of the problem, this should be identical to any midline on the other three

The final subsection, number 4.6, is dedicated to the area-integrated radiative heat transfer rate to the walls (in MW). This subsection provides a quantitative measure of the variation among the different solutions with regard to the total radiative heat transfer rate (in MW) to the walls of the enclosure. The area-integrated heat transfer is an important quantity when we are concerned about the operation of the furnace unit within the boiler, as this effects steam generation rate. The average radiative heat flux is calculated from this quantity by dividing area-integrated heat transfer rate by the total surface area of the walls (2 208 m2), and is also included in the comparison tables. This quantity provides a geometry-independent measure of the radiative heat load in oxy-fuel furnaces. The deviations from the benchmark box/EWB

Slices of the radiative source term along the 12×40 plane of symmetry are shown in Figure 4. Each figure corresponds to a different model, with a plot for each of the two oxy-fuel environments. The number and values of the contour levels are the same for the plots. The double-symmetric pattern in all plots is expected. The calculated negative value of the radiative heat source would *drive* the temperature field to lower values in a coupled simulation. The radiative source is smallest near the colder-than-medium walls; it increases steeply and becomes nearly flat over a large portion of the plane. Notice that this value is very similar for both environments. The box/EWB solution exhibits a smaller decrease of the radiative source near the walls than the WSGG solutions. For both environments, the air-fuel WSGGM (28) and the 5-gas/cubic WSGGM (27) show noticeable overprediction of the radiative source, which indicates a weaker influence of radiation on the thermal field.

Profiles of the radiative source term along the longitudinal centerline of the 3D enclosure are compared in Figure 5 for the two oxy-fuel environments. As mentioned in subsection 4.2, we also add published profiles (25) predicted using the SLW approach. For both environments,

further distinguished by adding the suffix *(air)* to its legend entry.

solution are also included. One table is provided per oxy-fuel environment.

pathlength as the EWB approach.

**4.3 Radiative-source contours**

**4.4 Radiative-source profiles**

12×40 walls.


Fig. 4. Midplane radiative source (left: 65%CO2; right: 90%CO2)

the flat portion of the radiative-source curve is smallest in the case of the box/EWB solution. As suggested from the 2D contours in the previous subsection, the air-fuel and the 5-gas/cubic WSGGM solutions show noticeable overprediction of the radiative source, with the air-fuel solution being the worst.

Tables 5 and 6 list the values of the radiative source at the middle of the profiles (which corresponds to the centroid of the 3D enclosure) for the various solutions, with the relative deviation from the box/EWB solution, computed as

$$\text{Percent error} = \frac{\text{SLW/WSGG} - \text{box/EWB}}{\text{box/EWB}} \times 100\% \tag{13}$$

**4.5 Radiative-flux profiles**

this underprediction jumps to 33.9%.

**4.6 Wall radiative heat transfer**

Solution method Wall-center's radiative flux

Solution method Wall-center's radiative flux

(kW/m2)

box/EWB 97.22 0.00% 4g, quadratic 99.67 + 2.52% 5g, quadratic 95.83 −1.43% 5g, quadratic, (cont) 94.37 −2.93% 4g, linear 103.11 + 6.05% 5g, cubic 94.63 −2.67% 4g, cubic, (air) 64.30 −33.87%

(kW/m2)

Table 5. Radiative flux at top-wall center for the oxy-fuel environment with 65%CO2

Table 6. Radiative flux at top-wall center for the oxy-fuel environment with 90%CO2

The area-integrated wall radiative heat flux results are compared for all the solutions in Table 7 for the wet-recycle environment and in Table 8 for the dry-recycle environment. Consistent with the profiles in the preceding subsection, the air-fuel WSGGM underpredicts the heat

box/EWB 113.98 0.00% 4g, quadratic 119.94 + 5.23% 5g, quadratic 119.96 + 5.25% 5g, quadratic, (cont) 113.85 −0.12% 4g, linear 116.33 + 2.06% 5g, cubic 113.19 − 0.70% 4g, cubic, (air) 91.32 −19.88%

%Error

%Error

(relative to box/EWB)

(relative to box/EWB)

The profiles of the radiative flux along the symmetry line of the 12×40 top wall for the two oxy-fuel environments are shown in Figure 6. We notice that the wall radiative flux is significantly more sensitive to the change in mixture composition than the centerline radiative source (see Figure 5). When the CO2 content increased, the radiative flux decreased. This is consistent with the decrease in total emissivity (18) and the changes in the idealized spectra of the linear absorption coefficient shown in Figure 1. We also notice that the relative deviations between the various WSGG models and the box/EWB predictions for the radiative flux differ from the deviations reported for the centerline radiative source. In particular, the 5-gas/cubic WSGGM (27) that showed noticeable error in the radiative source, has excellent agreement (-0.70%) with the box/EWB solution in the wet-recycle oxy-fuel environment and good agreement (-2.67%) in the dry-recycle environment. The radiative flux at the center point (*Z*=20 m) of the profiles in Figure 6 and their relative errors with respect to the box/EWB are compared in Tables 5 and 6 for the wet-recycle and dry-recycle oxy-fuel environments, respectively. All the WSGGM solutions are within 6.1% error (some underpredict and others overpredict) for both oxy-fuel environments, whereas the air-fuel WSGGM exhibits underprediction of 19.9% for the wet-recycle environment. For the dry-recycle environment,

Nongray EWB and WSGG Radiation Modeling in Oxy-Fuel Environments 505

For the SLW and oxy-fuel WSGGM, the errors have decreased for the high-CO2-fraction case, whereas this error increased in the case of the air-fuel solution. For the air-fuel solution, the errors are very large, being around 80%. Further, if the solutions are ranked by error, we get the same ordering for both oxy-fuel environments.


Table 3. Radiative source term at the centroid for the oxy-fuel environment with 65%CO2


Table 4. Radiative source term at the centroid for the oxy-fuel environment with 90%CO2

Fig. 5. Centerline radiative source for 2 oxy-fuel environments

### **4.5 Radiative-flux profiles**

12 Numerical Simulations / Book 1

For the SLW and oxy-fuel WSGGM, the errors have decreased for the high-CO2-fraction case, whereas this error increased in the case of the air-fuel solution. For the air-fuel solution, the errors are very large, being around 80%. Further, if the solutions are ranked by error, we get

%Error

%Error

(relative to box/EWB)

(relative to box/EWB)

**Z [m]**

**EWB SLW 4g, quad 5g, quad 5g, quad, (cont) 4g, linear 5g, cubic 4g, cubic, (air)**

**10% H2O + 90% CO2**

**<sup>0</sup> <sup>5</sup> <sup>10</sup> <sup>15</sup> <sup>20</sup> <sup>25</sup> <sup>30</sup> <sup>35</sup> <sup>40</sup> -50**

centroid (kW/m3)

Table 3. Radiative source term at the centroid for the oxy-fuel environment with 65%CO2

Table 4. Radiative source term at the centroid for the oxy-fuel environment with 90%CO2

**Source [kW/m3** 

**-40**

**-30**

**-20**

**-10**

**0**

**]**

centroid (kW/m3)

the same ordering for both oxy-fuel environments.

Solution method Radiative source at the

Solution method Radiative source at the

**Z [m]**

**EWB SLW 4g, quad 5g, quad 5g, quad, (cont) 4g, linear 5g, cubic 4g, cubic, (air)**

Fig. 5. Centerline radiative source for 2 oxy-fuel environments

**<sup>0</sup> <sup>5</sup> <sup>10</sup> <sup>15</sup> <sup>20</sup> <sup>25</sup> <sup>30</sup> <sup>35</sup> <sup>40</sup> -50**

**Source [kW/m3 ]**

**-40**

**-30**

**-20**

**-10**

**0**

box/EWB -15.15 0.00% SLW -13.53 +10.75% 4g, quadratic -14.64 + 3.37% 5g, quadratic -11.05 +27.09% 5g, quadratic, (cont) -11.47 +24.32% 4g, linear -11.61 +23.37% 5g, cubic - 7.62 +49.71% 4g, cubic, (air) - 2.52 +83.40%

**35% H2O + 65% CO2**

box/EWB -15.91 0.00% SLW -13.24 +16.80 % 4g, quadratic -14.67 + 7.80% 5g, quadratic -10.70 +32.73% 5g, quadratic, (cont) -10.96 +31.09% 4g, linear -11.95 +24.88% 5g, cubic - 7.53 +52.66% 4g, cubic, (air) - 3.22 +79.73%

The profiles of the radiative flux along the symmetry line of the 12×40 top wall for the two oxy-fuel environments are shown in Figure 6. We notice that the wall radiative flux is significantly more sensitive to the change in mixture composition than the centerline radiative source (see Figure 5). When the CO2 content increased, the radiative flux decreased. This is consistent with the decrease in total emissivity (18) and the changes in the idealized spectra of the linear absorption coefficient shown in Figure 1. We also notice that the relative deviations between the various WSGG models and the box/EWB predictions for the radiative flux differ from the deviations reported for the centerline radiative source. In particular, the 5-gas/cubic WSGGM (27) that showed noticeable error in the radiative source, has excellent agreement (-0.70%) with the box/EWB solution in the wet-recycle oxy-fuel environment and good agreement (-2.67%) in the dry-recycle environment. The radiative flux at the center point (*Z*=20 m) of the profiles in Figure 6 and their relative errors with respect to the box/EWB are compared in Tables 5 and 6 for the wet-recycle and dry-recycle oxy-fuel environments, respectively. All the WSGGM solutions are within 6.1% error (some underpredict and others overpredict) for both oxy-fuel environments, whereas the air-fuel WSGGM exhibits underprediction of 19.9% for the wet-recycle environment. For the dry-recycle environment, this underprediction jumps to 33.9%.


Table 5. Radiative flux at top-wall center for the oxy-fuel environment with 65%CO2


Table 6. Radiative flux at top-wall center for the oxy-fuel environment with 90%CO2

#### **4.6 Wall radiative heat transfer**

The area-integrated wall radiative heat flux results are compared for all the solutions in Table 7 for the wet-recycle environment and in Table 8 for the dry-recycle environment. Consistent with the profiles in the preceding subsection, the air-fuel WSGGM underpredicts the heat

Solution method Wall radiative heat

**6. Acknowledgments**

**7. Appendix**

in Equation (9).

Denmark) in implementing the EWBM.

**A. Idealized spectra for the box/EWB approach**

transfer (MW)

box/EWB 190.54 86.30 0.00% 4g, quadratic 200.62 90.86 + 5.29% 5g, quadratic 194.48 88.08 + 2.06% 5g, quadratic, (cont) 191.72 86.83 + 0.62% 4g, linear 210.34 95.26 +10.39% 5g, cubic 194.76 88.21 + 2.21% 4g, cubic, (air) 134.70 61.00 −29.31% Table 8. Wall radiative heat transfer for the oxy-fuel environment with 90%CO2

operating regime of the target system and the regime of the training data.

**B. WSGG linear absorption coefficients and blackbody weights**

Average wall radiative flux %Error

(relative to box/EWB)

(kW/m2)

Nongray EWB and WSGG Radiation Modeling in Oxy-Fuel Environments 507

different qualitative and quantitative radiative characteristics from the obtained solutions, we see that significant improvements in predictive capability can be obtained using an oxy-WSGGM. Using the air-fuel model would result in appreciable underprediction of the local and area-integrated radiative heat flux to the wall, and in an overprediction of temperatures due to the underprediction of the heat loss due to radiation. The errors become more pronounced for the high-CO2-concentration case, which is relevant to dry-recycle oxy-fuel combustion. The radiative heat flux was much more sensitive to the gas composition than the radiative source term. For the oxy-fuel WSGG models, no particular model was clearly superior. This suggests that the model used for a particular combustion problem should be selected based on the simplicity of the model and the consistency between the

This technical effort was performed in support of the National Energy Technology Laboratory's ongoing research in CO2 Capture in the Existing Plants Emissions and Capture (EPEC) Technology Program. Dr. Marzouk activities were funded under the RES contract DE-FE0004000. The authors appreciate the help of Dr. Chungen Yin (Aalborg University,

This appendix presents numerically the idealized spectra of the linear absorption coefficients *ki* and the corresponding blackbody weights *ai* that were computed from the EWB approach for each of the two oxy-fuel environments. The values are used when solving the RTEs given

Analogous to the tabulation in Appendix 7, the computed linear absorption coefficients and the corresponding weights for the gray gases are given in this appendix for all the 6 WSGG models for each of the two oxy-fuel environments. These values are used when solving the

Fig. 6. Radiative flux along the midline of the 12×40 top wall

transfer for both environments. Although the relative error with respect to the box/EWB is smaller than the error recorded for the 1D flux profile, the relative error for the dry-recycle environment is larger than the relative error for the wet-recycle environment. All the oxy-fuel WSGG models overpredict the heat transfer, but the error is within 10.4%.


Table 7. Wall radiative heat transfer for the oxy-fuel environment with 65%CO2
