Measurement of Ignition Delay Time For a Syngas Air Mixture Containing Water Vapor
Posted: Saturday, January 03, 2009
by Alexander Barrett
Alexander Barrett
5 th US Combustion Meeting
Organized by the Western States Section of the Combustion Institute
and Hosted by the University of California at San Diego
March 25-28, 2007.
Measurement of ignition delay time for a syngas/air mixture containing water vapor
A. B. Barrett, S. C. Reehal, and E. L. Petersen
University of Central Florida , Orlando , FL , USA
Ignition delay times of syngas mixtures containing water vapor are of interest to the gas turbine industry because water is often a non-negligible constituent of syngas. This paper presents ignition time data from shock-tube experiments obtained for three mixture compositions, two with water, fuel, and air and one with just fuel and air. The experiments were performed in the intermediate temperature range (950-1400 K) and at pressures around 1.5 atmospheres. The fuel-to-air equivalence ratios were 1.0, 0.9 and 1.0 for the no-water, 50/50% fuel/water, and 87/13% fuel/water blends, respectively. The results were then compared with a modern kinetics model. There was good agreement between the model and the data at the pressures and temperatures of the present experiment, particularly at the highest pressures where the kinetics are dominated by H + O 2 = OH + O branching kinetics.
1. Introduction
Recently, there has been increased interest in syngas, or synthetic gas, a by-product of the coal gasification process, for use as a fuel in gas turbine and other applications. Due to the non-uniformity of syngas composition, there is a need for ignition time data on various H 2 /CO blends [1,2]. Such data are useful fundamentally for the refinement of chemical kinetics models and practically for the estimation of autoignition times in premixed systems. Ignition delay times applicable to normal operating conditions for gas turbines are needed for specific temperatures, pressures, compositions, and fuel-to-air equivalence ratios. The current study focuses on syngas blends mixed with water vapor and undiluted air.
Recent research by the authors and their collaborators on syngas ignition at elevated pressures (up to about 40 atm) and low-to-intermediate temperatures has shown that at the highest pressures and lowest temperatures modern chemical kinetics models for CO/H 2 combustion do not always agree with experimental data [3-6]. Although water vapor can be a significant component of syngas fuel blends, relatively few kinetics data are available with water as the collision partner. Hence, it can be expected that further refinement of the kinetics mechanisms may be needed to model combustion behavior in the presence of water vapor. This paper presents some of the first data on the combustion of a water/fuel/air blend at temperatures ranging from 950-1400 K, with a fuel-to-air equivalence ratio of around 1.0 and an average pressure of about 1.5 atmospheres. The following sections detail the shock-tube facility where the experiments were performed, the procedure used to obtain the data, the ignition delay time results, and comparisons to a modern kinetics model.
2. Experimental Setup
The shock-tube facility at the Gas Dynamics Laboratory at the University of Central Florida (UCF), as described by Rotavera et al. [7], was used to conduct all experiments. The UCF shock tube has 1.8-m driver and a 4.3-m driven section separated by a 0.25-mm-thick lexan diaphragm. The driven section is constructed of square stainless steel tubing that measures 15.24 cm per side, and the helium-driven driver is round, stainless steel tubing with an inner diameter of 7.62 cm. The OH* emission was measured at the test section through the side wall port located 1 cm from the endwall through a 10-nm bandpass filter centered at 310 nm. The endwall pressure trace was obtained with the use of a PCB 134A pressure transducer, and a Kistler 603B1 pressure transducer was used to get the side wall trace. The shock velocity was determined by using four fast response ( 1s) pressure transducers (PCB 113A) and three time interval counters (Phillips Fluke PM 6666). The temperature and pressure data behind the reflected shock was derived from the incident shock speed using the standard 1-D shock relations and the Sandia thermodynamic database [8].
A list of the three mixtures is contained in Table 1. Usually the fuel, oxidizer, and diluent are premixed in a mixing tank and then the shock-tube is filled with this mixture prior to the experiment. However, the vapor pressure of water at room temperature (20 torr) is too low to run experiments using this procedure. This is because there will be a tendency for the water to condense when the blend reaches sonic conditions at the choke point where the mixture enters the vacuum of the shock-tube driven section. Therefore, for these sets of experiments, the shock-tube was pumped down using a turbomolecular pump, and then the tube was filled with the proper amount of water vapor by allowing the low pressure within the tube to vaporize the water. Water vapor was added to the test section to the pressure calculated beforehand to produce the desired amount of H 2 O in the final fuel/air/water mixture. The fuel/air blend was then added simultaneously through two different locations on the driven section, and the complete mixture was allowed to mix within the tube by the induced turbulence from the injected flows and mass diffusion. Filling from two different axial locations along the tube increases the turbulent mixing of the gas mixture injected through the ports and the water vapor already present in the tube; it also decreases the distance that the mixture has to diffuse to create a thorough mixture of the fuel/air blend with the water vapor in the tube.
Table 1: Mixtures and Compositions
mixture
blend
phi
X H2O
X H2
X CO
X CH4
X CO2
X Ar
X O2
X N2
1
no water
1.0
-
0.2338
0.0134
0.0116
0.0306
0.0029
0.147
0.5612
2
87/13% fuel/water
1.0
0.0396
0.2223
0.0129
0.0112
0.0292
0.0027
0.1415
0.5406
3
50/50% fuel/water
0.9
0.2314
0.1795
0.0103
0.0089
0.0235
0.0022
0.1128
0.4314
The ignition times were interpreted using the sudden rise in the endwall pressure that occurred at the time of ignition, as in Kalitan et al. [4,5] and Reehal et al. [6]. OH* emission obtained from both the endwall and sidewall locations was used to verify the ignition as determined from the pressure measurement.
3. Results
The temperature range for the current data was between 950 and 1400 K with an average pressure between 1.0 and 1.5 atmospheres. Using Chemkin and the H 2 /CO mechanism of Davis et al. [9], a comparison between the data and the latest kinetics modeling was made for each mixture composition using the same fuel-to-air equivalence ratio and temperature range as the data set and the average pressure for that mixture. Each mixture has been plotted in comparison to the model. No driver-gas tailoring was done for these mixtures to increase the test time, and each mixture was combusted at the lowest possible temperature within the experimental test time, which was less than 2.0 ms.
Figure 1 shows a plot for mixture 1 (no-water). The temperature range for this mixture was 950-1250 K, and the fuel-to-air equivalence ratio was 1.0. The average pressure for this data set was 1.4 atmospheres. The data for this set compares extremely well with the model. The only discrepancy occurs at the one datum at a temperature of 953 K where the experimental ignition time is slightly faster than the theoretical ignition time.
Figure 1: Comparison of the model and the data for the "no-water" fuel mixture
For mixture 2 (87/13% fuel/water), the temperature range was 980-1160 K, with an average pressure of 1.3 atmospheres and a fuel-to-air equivalence ratio of 1.0. Figure 2 shows a plot comparing the data for the second mixture to the model. The data for this set has moderate agreement with the model. Most of the data match the model, but some data near 1000 K had a faster ignition time than the model predicted.
Figure 2: Comparison of the model and the data for the 87/13% fuel/water mixture
The data for mixture 3 (50/50% fuel/water) is contained in Fig. 3. The temperature range for this mixture was 1070-1400 K. The fuel-to-air equivalence ratio was 0.9, and the average pressure was 1.2 atmospheres. The data for this set compares extremely well with the model. The only discrepancy occurs at a temperature of 1071 K where the experimental ignition time is slightly faster than the theoretical.
Figure 3: Comparison of the model and the data for the 50/50% fuel/water mixture
Figure 4: Comparison of the model and the data for the all three mixtures
Figure 4 shows a plot with all three mixtures and the corresponding model for each mixture. From this plot, it can be determined that the water causes the ignition time to be longer for this fuel blend. This plot clearly demonstrates that the 50/50% fuel/water blend causes the ignition to be delayed by a considerable amount. The effect of water on the ignition behavior at lower temperatures and higher pressures is not unexpected and is due to the increased efficiency of water as a collision partner in this region of mixture concentration and temperature. At the higher temperatures, the ignition behavior is between the first and second explosion limit and is dominated by the well-known branching reaction H + O 2 = OH + O. However, above the second explosion limit for H 2 /CO mixtures, the chain termination reaction H + O 2 + M = HO 2 + M competes for H radicals. The increased collision efficiency of the water (M = H 2 O) in turn increases the effectiveness of the termination reaction, leading to longer ignition times.
It can also be established that no ignition will occur below 950 K within the test-time limitation of 1.9 ms because none was observed when experiments were performed for these mixtures at lower temperatures. The 87/13% water/fuel mixture and the no-water mixture show little difference in ignition times in the hot region (i.e. > 1000 K), but in the cooler region the difference can be quite substantial due to the chain termination kinetics mentioned above. When plotted as log t versus 1/T, each data set is fairly linear for temperature above 1100 K.
4. Future Work
A new shock tube is being constructed at UCF that will have a fill tank, fill line, and driven section capable of being heated to 100C or higher. This shock tube will have the capacity to mix these water/fuel blends in the mixing tanks and then deliver them to the driven section without having to wait for complete mixing by mass diffusion, since the gases and the fill line will be at a high enough temperature to prevent condensation. This procedure will also allow higher-pressure experiments to be performed because the current method of making the fuel/water/air mixture directly in the shock tube is limited by the partial pressure of the water at room temperature. The shock tube will also be equipped with a tunable diode laser centered at 1384 nm that will be used as a diagnostic to measure the water vapor concentration in the shock tube using the absorption of the laser light and the known water spectroscopy. The laser used will be the TEC-500 tunable diode laser in the Littman/Metcalf configuration according to Sacher Lasertechnik. This instrument will make it possible to measure the water vapor concentration very accurately.
The new shock tube that is being constructed will also have the capability to reach much higher pressure than the current UCF shock tube. Another unique feature of this shock tube will be a longer driven section, which will enable longer test times since the expansion waves will take longer to reflect off the driver endwall and back down to the test section to end the test time. This will allow lower temperature experiments to be performed.
5. Summary
Ignition delay time data for three separate mixture compositions of syngas fuel, water, and air were obtained. These ignition times represent some of the first data on CO/H 2 /H 2 O mixtures undiluted in air to the authors' knowledge. The results were compared to a modern kinetics model. Overall, the model and the data presented in this paper have considerable agreement. A few discrepancies in the low-temperature region may be an indication of a variance from the model. Further study of these mixtures at higher pressure and lower temperature are required because the impact of water as a third body collider is expected to be more prominent at such extremes of temperature and pressure.
Acknowledgments This work was supported by Siemens Power Generation with Dr. Scott Martin as the technical monitor.
References
[1] T. Lieuwen, V. G. McDonell, E. L. Petersen, D. Santavicca, ASME Paper GT2006-90770 (2006).
[2] G. A. Richards, M. M. McMillan, R. S. Gemmen, W. A. Rogers, S. R. Cully, Progress in Energy and Combustion Science 27 (2001) 141-169.
[3] E. L. Petersen, D. M. Kalitan, A. Barrett, S. C. Reehal, J. D. Mertens, D. J. Beerer, R. Hack, V. McDonell, Combustion and Flame, in press (2007).
[4] D. M Kalitan, E. L. Petersen, AIAA Paper 2005-3767 (2005).
[5] D. M. Kalitan, E. L. Petersen, J. D. Mertens, M. W. Crofton, ASME Paper GT2006-90488 (2006).
[6] S. C. Reehal, D. M. Kalitan, T. Hair, A. Barrett, E. L. Petersen, 5 th U.S. Combustion Meeting, Paper C24 (2007).
[7] Rotavera, B., Amadio, A., Antonovski, V., and Petersen, E., AIAA Paper 2006-4725 (2006).
[8] R. J. Kee, F. M. Rupley, J. A. Miller, J. A., The Chemkin Thermodynamic Database, SAND87-8251B, Sandia National Laboratory, March 1990.
[9] S. G. Davis, A. V. Joshi, H. Wang, and F. Egolfopolous, Proceedings of the Combustion Institute 30 (2005) 1283-1292.
Alexander Barrett, 3502 Pecos St . , Bryan , TX 77801
321-278-5115
Alexanderbb2003@yahoo.com
This Article has been viewed 694 times. (Not updated in real-time.)
Top-level comments on this article: (1 total)wow--This seems to be the kind of technical article that is usually found in a professional journal--it was way beyond me, but I do hope there are readers who find it informative and useful.It was for the 5th US Combustion Meeting, papers submitted to that conference are supposed to be in the style of a technical paper, but not of the length of a journal paper.Correct. The style was in accordance with the conference's regualtions. However, the formatting was changed when transferring it to this site.
We want your comments! If you can read this, you don't have javascript enabled, so you can't use this comment system. Please enable javascript.