• Energy Research

Abstract There is a renewed interest in oxy-fuel combustion of natural gas for reduction of greenhouse gas emissions. This has necessitated various experimental and numerical studies of oxy-fuel combustion. In the numerical combustion study, the radiation model and combustion chemistry are critical for accurate numerical predictions of oxy-fuel combustion characteristics. In this study, three global reaction mechanisms: Westbrook-Dryer (3 equations), Jones-Lindstedt (5 equations) and Jones-Lindstedt (7 equations) for oxy-methane combustion were combined with different weighted sum of gray gas radiation models (WSGGM) available in the literature to determine the most accurate combination for oxy-methane combustion modeling and simulation. Experiments were conducted in a non-premixed swirl stabilized model gas turbine combustor at a firing rate of 4 MW/m3-bar while varying the percentage of CO2 in the oxidizer (O2/CO2) mixture. Numerical model developed using ANSYS FLUENT 17 code was validated against the experimental results. The combinations of Jones-Lindstedt (5 equations) reaction mechanism and WSGGM model proposed by Bordbar gave the closest approximation of the flame temperature with an average deviation of 5.52%. The model combination also predicted the flame attachment to the fuel nozzle and flame lift-off at a high CO2 percentage in the oxidizer mixture. The results of the parametric study on the effect of CO2 percentage in the oxidizer mixture, combustor energy level and equivalence ratio on the combustion characteristics and CO emissions were also reported. The CO emissions monotonically increases with increasing percentage of CO2 in the oxidizer due to decreased residence time and the reduction in the flame temperature. While the CO emission increases with the energy level in the combustor up to 3.5 MW/m3 and decreases thereafter. The optimum equivalence ratio for minimum CO emission is 0.98 with approximately 2 PPM at 40% CO2 in the oxidizer.

Abstract Radiation modeling and combustion chemistry are critical to accurate numerical predictions of oxy-combustion characteristics. The presence of carbon dioxide (CO2) in oxy-fuel combustion acts as diluents and significantly changes the radiative properties of the combustion gases in comparison to air combustion. In this regard, three global reaction mechanisms: Westbrook-Dryer (3 equations), Jones-Lindstedt (5 equations) and Jones-Lindstedt (7 equations) for oxy-fuel combustion of methane were combined with different weighted sum of gray gas radiation models (WSGGM) available in literature to determine the most accurate combination for methane-oxy-fuel combustion modeling and simulation. This study was carried out in a non-premixed swirl stabilized model gas turbine combustor at a firing rate of 4MW/m3-bar. The modified Westbrook-Dryer (WD-oxy) mechanism could not predict the flame attachment to the fuel nozzle at 35% CO2 addition. The combinations of Jones-Lindstedt (5 equations) reaction mechanism and WSGGM model proposed by Bordbar gave the closest approximation to our experimental observations and predicted the flame attachment to the fuel nozzle.

Carbon dioxide (CO2) emission forms the biggest portion of greenhouse gas emissions known to cause global warming, which can lead to climate change. One of the most widely recommended means of tack...

The implementation of reduced syngas combustion mechanisms in numerical combustion studies has become inevitable in order to reduce the computational cost without compromising the predictions' accuracy. In this regard, the present study evaluates the predictive capabilities of selected detailed, reduced, and global syngas chemical mechanisms by comparing the numerical results with experimental laminar flame speed (LFS) values of lean premixed (LPM) syngas flames. The comparisons are carried out at varying equivalence ratios, syngas compositions, operating pressures, and preheat temperatures to represent a range of operating conditions of modern fuel flexible combustion systems. NOx emissions predicted by the detailed mechanism, GRI-Mech. 3.0, are also used to study the accuracy of the selected mechanisms under these operating conditions. Moreover, the selected mechanisms' accuracy in predicting the laminar flame thickness (LFT), species concentrations of the reactants, and OH profiles at different equivalence ratios and syngas compositions are investigated as well. The LFS is generally observed to increase with increasing equivalence ratio, hydrogen content in the syngas, and preheat temperature, while it is decreased with increasing operating pressure. This trend is followed by all mechanisms understudy. The global mechanisms of Watanabe–Otaka and Jones–Lindstedt for syngas are consistently observed to over-predict and under-predict the LFS up to an average of 60% and 80%, respectively. The reduced mechanism of Slavinskaya has an average error of less than 20%, which is comparable to the average error of the GRI-Mech. 3.0. It however over-predicts the flame thickness by up to 30% when compared to GRI-Mech. 3.0. The NO prediction by Li mechanism and the reduced mechanisms are observed to be within 10% prediction range of the GRI-Mech. 3.0 at intermediate equivalence ratio (φ=0.74) up to stoichiometry. Moving toward more lean conditions, there is significant difference between the GRI-Mech. 3.0 NO prediction and those of the reduced mechanisms due to relative importance of the prompt NOx at lower temperature compared to thermal NOx that is only accounted for by the GRI-Mech. 3.0.

AbstractEvaporative cooling technology has a potential to serve as a substitute to conventional vapor compression cooling. Direct evaporative cooling however usually introduces more moisture to the cooling space. In this study, the performance of a modified direct evaporative cooling system that combines a cooling pad and a removable dehumidifying pad has been experimentally evaluated for space cooling. The cooling pad is made of luffa fiber lagged with charcoal, while the dehumidifying pad is made of activated carbon derived from tamarind seed. Results for two experimental days, which span from 8:30 am to 5:30 pm each day are reported in this work. The peak cooling load requirement of the room was evaluated as 4.53 kW. On the first experimental day, in which the dehumidifying pad was removed from the system, results indicated a minimum room temperature of 24oC was achieved, which resulted in a maximum temperature drop of 11oC from ambient temperature. However, indoor relative humidity increased to a maximum of 84%, while outdoor relative humidity was 30%. The dehumidifying pad was used on the second experimental day. Results from the second experimental day showed a minimum room temperature of 26.5oC was achieved, resulting in a maximum temperature drop of 10oC from ambient. Maximum indoor relative humidity recorded was 49%, while the outdoor relative humidity was 34%, an indication that the dehumidifying pad was able to absorb moisture from the cooled air. Maximum cooling capacity, efficiency, and COP of 3.84 kW, 84.6% and 16.1 respectively were achieved by the system without the dehumidifying pad. Corresponding values of 3.2 kW, 71.4% and 13.4 respectively were recorded when the system was operated with the dehumidifying pad.

An ultra lean mixture (ϕ ≤ 0.5) of methane–hydrogen–air was experimentally investigated to explore the effect of fuel flexibility on the flame stability and emission of a nonpremixed swirl stabilized combustor. In order to isolate the effect of hydrogen addition to methane, experiments were carried out at fixed fuel energy input to the combustor while increasing the hydrogen content from 0% up to 50% in the methane–hydrogen mixture on volume basis. The combustor fuel energy was then increased up to the range of typical gas turbine combustors. Equivalence ratio sweep was carried out to determine the lean stability limit of the combustor. Results show that the hydrogen content in the fuel mixture and fuel energy input have a coupled effect on the combustor lean blow off velocity (LBV), temperature and emissions. The LBV increases by ∼103% with the addition of 30% H2. On the other hand, the LBV increases by ∼20% as the fuel energy increases from 1.83 MW/m3 to 2.75 MW/m3. Burning under ultra lean condition serves two purposes. (1) The excess air supplied reduces the overall combustor temperature with its ensuing effect on low NOx formation. (2) It increases the overall combustor volume flow rate which reduces the residence time for NOx formation. The axial temperature profile presented along with the emission data can serve as basis for the validation of numerical models. This would give more insight onto the effect of hydrogen on the turbulence level and how it would improve the localized extinction of methane in a cost-effective way.

Abstract Emission of greenhouse gases like CO 2 is associated with global warming and ultimately climate change. Oxy-fuel combustion as a carbon capture technique is among the most widely recommended means of tackling CO 2 emission. The use of oxygen/carbon dioxide (O 2 /CO 2 ) mixture as oxidizer in place of air will result in distinct combustion characteristics. Flame blowout is one of such characteristics and is one of the critical operational issues in real combustion systems. In this study, blowout characteristics of air and oxyfuel combustion of propane in a non-premixed, swirl-stabilized combustor were studied. The effect of the equivalence ratio on flame blow off for three oxidizer mixtures (air, oxyfuel I, and oxyfuel II) was first studied. Furthermore, the effect of CO 2 addition and swirl number on oxyfuel flame blowout was also studied. Results show that the propane–air flame transits from: attached flame - lifted flame - no flame regimes with the lifted flame regime occurring at a critical velocity ratio (ratio of fuel velocity to oxidizer velocity). The oxyfuel I and oxyfuel II flames, however, transit directly from attached flame to no flame regime at all firing rates below 4.5MW/m 3 bar. This is due to the fact that oxyfuel I and oxyfuel II could not reach the critical velocity ratio for oxy-flames (Critical velocity ratio observed = 3) before flame blowout. We observed that the critical velocity ratio for propane oxyfuel combustion is six times more than that of propane–air combustion (about 0.5). At 5 MW firing rate, the amount of CO 2 required in the oxidizer at the attached flame - lifted flame transition point decreases from 76% to 74% as the equivalence ratio decreases from 1 to 0.9. The velocity ratios at this operating condition are 3.1 and 3.2, respectively and are more than critical velocity ratio for propane oxyfuel combustion. This indicated that attached flame - lifted flame transition is strongly dependent on the critical velocity ratio irrespective of the equivalence ratio. Further studies showed that oxyfuel flames are more stable at stoichiometry conditions and swirl number of unity.

Abstract In this paper, we report results from an experimental investigation on transitions in the average flame shape (or microstructure) under acoustically coupled and uncoupled conditions in a 50 kW swirl stabilized combustor. The combustor burns CH4/H2 mixtures at atmospheric pressure and temperature for a fixed Reynolds number of 20,000 and fixed swirl angle. For both cases, essentially four different flame shapes are observed, with the transition between flame shapes occurring at the same equivalence ratio (for the same fuel mixture) irrespective of whether the combustor is acoustically coupled or uncoupled. The transition equivalence ratio depends on the fuel mixture. For the baseline case of pure methane, the combustor is stable close to the blowoff limit and the average flame in this case is stabilized inside the inner recirculation zone. As the equivalence ratio is raised, the combustor transitions to periodic oscillations at a critical equivalence ratio of ϕ = 0.65 . If hydrogen is added to the mixture, the same transition occurs at lower equivalence ratios. For all cases that we investigated, flame shapes captured using chemiluminescence imaging show that the transition to harmonic oscillations in the acoustically coupled cases is preceded by the appearance of the flame in the outer recirculation zone. We examine the mechanism associated with the transition of the flame between different shapes and, ultimately, the propagation of the flame into the outer recirculation zone as the equivalence ratio is raised. Using the extinction strain rates for each mixture at different equivalence ratios, we show that these transitions in the flame shape and in the instability (in the coupled case) for different fuel mixtures collapse as a function of a normalized strain rate : κ e x t D U ∞ . We show that the results as consistent with a mechanism in which the flame must overcome higher strains prevailing in the outer recirculation zone, in order to stabilize there.