The Stability of Turbulent Hydrogen Diffusion Jet Flames

Transcription

1 The Stability of Turbulent Hydrogen Diffusion Jet Flames Wu, Y. 1, Al-Rahbi, I. S. 1 and Kalghatgi, G. T. 2 1 Department of Chemical and Process Engineering, University of Sheffield, Sheffield, S1 3JD, UK 2 Shell Global Solutions (UK), Cheshire Innovation Park, P.O. Box 1, Chester CH1 3SH. UK Abstract A most probable scenario resulting from an accidental release of hydrogen from a highpressure source is hydrogen diffusion jet flames. Therefore hydrogen diffusion jet flames posse serious safety hazards for hydrogen storage vessels and transport pipelines. The lift-off velocity, lift height and blow-off velocity have significant meaning in the diffusion flame modelling and also have strong implication in safety consideration and hazard assessment. This study is aimed at investigation of the factors affecting the stability of turbulent hydrogen jet flames. Carbon dioxide and propane have the same density but different chemical kinetics. By adding carbon dioxide or propane into hydrogen jet, the present work studied the effect of mixing and kinetics on the flame stability. The experiments were carried out using a concentric burner with mm inner diameter and 8 mm outer diameter. Carbon dioxide or propane was added into hydrogen in three ways. The first was to premix with hydrogen and to form a jet flame. The second was to be added as an annular jet around the hydrogen flame. The third was to injected into the centre of the hydrogen flame. The flame structures were recorded by direct filming and also by a schlieren apparatus. A wide range of stream velocities and Reynolds number were investigated. The experiments showed that propane and carbon dioxide affected the stability of the hydrogen very different. Although carbon dioxide had quenching effect on reaction, but the propane appeared to be more effective to produce a lifted flame. The experiments also showed that hydrogen flame can be effectively lifted or blow out by adding the carbon dioxide or propane as annular flow around the hydrogen flame. CFD simulations were carried out to reveal details of jet flow mixing process. 1. Introduction Hydrogen has very wide flammability limits and a very high burning velocity. At atmosphere pressure and ambient temperature, the flammability limits for hydrogen-air mixture are 4.1% to 74.8%. The burning velocity varies from 26 to 32 cm/s. The lift-off velocity is greater than m/s. Consequently, a pure hydrogen leak from a rupture over 2mm on a hydrogen storage vessels, which operate at a pressure about 2 MPa, will likely produce a stable attached flame. It is very unlikely the exit velocity could reach the lift-off and blow-out velocity. The implication is that the hydrogen jet flame has to be quenched by adding other gaseous or liquid substances into the jet. This work concerns the examination of the characteristics of the stability of hydrogen flame, and the effect of adding inert gases or other hydrocarbon fuels into the jet. The inert gases used are carbon dioxide and argon. Propane was selected to represent hydrocarbon fuel. Propane has the same density as carbon dioxide but different chemical kinetics. By adding carbon dioxide or propane into hydrogen jet, the present work studied the effect of mixing and kinetics on the flame stability. 2. The Experimental Study The experiments were carried out using a concentric burner with a.3 mm or 2 mm inner diameter and a 8 mm outer diameter. Carbon dioxide or propane was added to hydrogen in three ways. The first was to premix with hydrogen and to form a jet flame. The second was to be added as an annular jet around the hydrogen flame. The third was to be injected into the centre of the hydrogen flame. Both the central flow stream and the annual flow stream have flow settling chamber and flow straightening device. The gasses were introduced from

2 compressed gas bottles through flow meters and were mixed before channelling into the settling chambers. The visual characteristics of the flames of pure hydrogen, hydrogen/carbon dioxide mixtures and hydrogen/propane mixtures were very different. The pure hydrogen flames were almost invisible. Hydrogen/carbon dioxide flames were in blue. The hydrogen/propane flames resembled characteristics of propane ones and appeared in blue in the base of the flame, but bright yellow in the main combustion zone. To visualise the flames and establish the lift off height of the flames, both schlieren technology and direct digital photography technology were used to capture the flame images. A positive-negative-grid schlieren system was constructed for the rig and this was mainly used to visualise hydrogen, hydrogen/carbon dioxide flames. All flame images were captured using digital camera and processed using computer graphic packages. 3. Experimental Results Pure Hydrogen Jet Flames The experimental results showed that pure hydrogen jet flame was very stable. Lifted flames were produced only in 2mm-diameter burner. The lift-off velocity measured for pure hydrogen flame was as high as 7 m/s. Figure 1 shows the lift-off height against the jet exit velocity from present study and the measurement from Kalghatgi [1] and Cheng and Chiou [2]. The results showed that the lift-off height increased linearly with the jet velocity. The measured lift-off heights are in good agreement with the results of Kalghatgi [1] and Cheng and Chiou [2] and also in agreement with the values of Brockhinke and co-workers [3] based on volumetric flow rate without pressure correction. Blow out of pure hydrogen jet flame has never been achieved in our experiments. 3 3 Lift-off height (mm) Current study Kalghatgi [1] Cheng and Chiou [2] Jet exit velocity (m/s) Figure 1. Experimental measured lift-off height of pure hydrogen flame. Effect of Adding Propane or Carbon Dioxide on the Lift-off Velocity and Lift-off Height The effect of adding CO 2 or propane into an initial attached hydrogen flame was tested and the jet lift-off velocity against the percentage of CO 2 or propane addition in the 2mm burner was plotted in Figure 2. It appeared that propane gas was more effective in lifting the flame. Adding very small amount propane, the lift-off velocity was reduced significantly. CO 2 was not so effective in produce lifted flame. Lifted hydrogen/co 2 flames were observed only in the conditions when CO 2 addition was less than 6%. When CO 2 addition was more than 6%, the hydrogen/co 2 flame stayed attached until it blew off. Comparing the lift-off height shown in Figure 3, for the same percentage of addition the hydrogen/propane flames produced a higher lift-off height than hydrogen/co 2 flames.

3 Percentage of added gas Adding carbon dixode Adding propane Jet lift-off velocity (m/s) Figure 2. Jet lift-off velocity plotted against the percentage of propane or carbon dioxide addition in 2 mm-diameter burner. Lift-off height (mm) Adding C3H8 Adding CO2 Pure H Jet exit velocity (m/s) Figure 3. Effect of adding propane and carbon dioxide into an initially lifted hydrogen flame. To analyse the experimental results, a detailed study and comprehensive literature review on the effect of propane and CO 2 addition on the burning velocity of hydrogen flames was carried out. Some limited research has been carried out on the effect of additive on the burning velocity of hydrocarbon fuel. The effect of CO 2 addition on the burning velocity of methane and propane has been well studied. Yumlu [4] experimentally studied the effect of additives on the burning speed of fuel-air mixtures using a circular burner. For adding inert gas, an equation for the burning velocity of the mixture,, was derived as: S = (1 α) S exp[ E / R(1/ T 1/ T 2 1/ 2 u, m u b b b u m )]S, Eq. 1 Where α is the mass fraction of the additive, T is the adiabatic flame temperature and E is the activation energy. However the effect of adding combustible gas to the fuel-air mixture was not well understood and there was very limited experimental data available on the laminar burning velocity of hydrogen/propane mixture. Leason [] experimentally studied the effect of various additives including hydrogen on the burning velocity of propane-air mixture. Though this work covered an extended range of equivalence ratio, it was limited to hydrogen concentrations in the H 2 -C 3 H 8 mixtures between and 3%. Hydrogen was the dominant fuel with small amount propane addition; therefore Leason s data were not very useful in our study. Milton and Keck [6] made a further study on the effect of hydrogen addition on the burning velocity of methane-air and propane-air mixtures at the stoichiometric ratio. Figure 4 shows a comparison of the effect of adding CO 2 on the burning velocity based on Yumlu s

4 equation and the effect of adding propane on the burning velocity based on the experimental data from Milton and Keck. It was shown that the reduction on the burning velocity by adding propane was much more significant and greater than by adding CO 2. Adding a small amount of propane reduced the burning velocity significantly. If the percentage of propane was greater than 3%, the burning velocity of the mixture was equal to that of pure propane fuel. The analysis revealed that it was propane, rather than hydrogen that seemed to be the dominant element on the burning velocity of the mixture. It was suspected that the explanation for this phenomenon may lie on the detailed chemical reaction mechanism of the fuels. CO 2 did not involve combustion reaction, therefore the reaction mechanisms remained as hydrogen-air. However hydrogen and propane are very active fuels. It can be concluded that the blended fuel should be treated as separate fuel having its own characteristics. This area has received little attention before. There is clearly increasing demand for further research on the effect of fuel blending and on the overall combustion characteristics of such fuels. 1.2 Ratio of burning velocity, S u / S u,hydrogen Adding propane Adding carbon dioxide Percentage of additive (%vol.) Figure 4. Ratio of laminar burning velocity of the mixture to that of pure hydrogen against the percentage of gas added. Effect of Adding Propane or Carbon Dioxide on the Blow Out Velocity Both CO 2 and propane could blow out hydrogen flames. Figure shows the blow out velocity against the percentage of gas addition for propane, carbon dioxide or argon addition. Although propane required a slightly higher concentration, its effectiveness on blow out the flame could be considered as good as that of CO 2. Blow out velocity (m/s) 16 Adding CO2 14 Adding C3H8 12 Adding Ar Diluent Concentration (% Vol.) Figure. Blow out velocity against percentage of added gas on a 2 mm-diameter burner. Influence of Adding Propane and Carbon Dioxide as Annular Co-flow The experimental tests showed that the hydrogen flame could be lifted easily when propane and CO 2 were added as annular flow around the hydrogen jet. The lift-off height, as shown in Figure 6, was predominantly dependent on the velocity of the annular flow and was very

5 weakly influenced by the central hydrogen velocity. Figure 7 showed that again the lift-off height was greater when propane was in the annular flow. Blow-out could also be achieved easily in this flow arrangement. When CO 2 was used, blow out occurred when the velocity of CO 2 approached the hydrogen central jet velocity. There was a linear relationship between the CO 2 velocity required to blow out the flame and the hydrogen jet velocity as shown in Figure 8. Tests also carried out using Argon. The Argon required a slightly higher blow out velocity. This flow arrangement has great influence on the lift-off and blow-out conditions of the jet. The annular flow around the hydrogen jet directly interfered with the hydrogen jet diffusion process in the air and interacted directly with the reaction zone. On the other hand, when propane and CO 2 were injected as central flow with hydrogen in the annular flow, the flame was as stable as the pure hydrogen flames. The propane and CO 2 mainly changed the combustion characteristics in the main combustion zone higher in the jet. 2 1 Lift-off height (mm) 1 propane in annular flow at 12 m/s propane in annular flow at 18 m/s CO2 in annular flow at 1 m/s CO2 in annular flow at 12 m/s CO2 in annular flow at 18 m/s Hydrogen central jet exit velocity (m/s) Figure 6. Lift-off height versus hydrogen jet velocity when propane and carbon dioxide were in annular flow and hydrogen in central jet. 3 2 propane in annular flow CO2 in annular flow Lift-off height (mm) Annular flow velocity (m/s) Figure 7. Lift-off height versus annular flow velocity. The inner burner diameter was.3 mm and outer diameter of the annular flow was 8 mm.

6 Annular flow velocity (m/s) Argon in annular flow CO2 in annular flow Hydrogen central jet exit velocity (m/s) Figure 8. Blow out conditions when CO 2 or Argon was in the annular flow and hydrogen in the central jet. Inner burner diameter was.3 mm and outer diameter of the annular flow was 8 mm. 4. Analysis of Jet Velocity and Turbulent Burning Velocity based on CFD Simulations Vanquickenborne and van Tiggelen [7] proposed that the position of a lifted flame in a jet of fuel is that at which the turbulence characteristics give rise to a burning velocity on the stoichiometric contour equal to the incident mean velocity. Pitts [8] [9] pointed out that the flame stability properties could be predicted using time-averaged concentration and velocity profiles of the corresponding non-reacting jet flows of fuels into air. In this study, we explored the use of CFD to predict the non-reacting jet, and to calculate the inferred burning velocity using the turbulent properties and species concentration from CFD simulations and the turbulent burning theories by Gülder [1] and Peters [11]. The jet was simulated using a standard k ε model by FLUENT6. As an example, the CFD simulation of hydrogen/co 2 flame shown in Figure 9 with hydrogen in central jet at 4m/s and CO 2 in a annular flow at 18 m/s is discussed. Figure 1 (a), (b) and (c) shows jet velocity profiles and burning velocity profiles at cross-sections 1 mm, 18mm and 2 mm above the burner rim, respectively. The lift-off height was 2 mm and the inner edge and outer edge of the flame base and located.3 mm and 7 mm away from the burner axial, respectively. The actual flame base is marked in Figure 1 (c). Figure 1 showed that the burning velocity profile approached to the velocity profile in Figure 1 (a) & (b). At the flame base, the maximum burning velocity equal to the flow velocity and the entire base was within the flammability and the burning velocity matched with the flow velocity. Lift-off height Figure 9. Image of hydrogen/ CO 2 flame. Hydrogen was in central jet at 4 m/s and CO 2 in annular flow at 18 m/s.

7 3 Burning velocity- using Peters [11] Burning velocity - using Gulder [1] Jet velocity 2 Velocity (m/s) Radial position (mm) 3 2 (a) 1 mm above the burner Velocity (m/s) Radial position (mm) (b) 18 mm above the burner 3 2 Velocity (m/s) Flame base Radial position (mm) (c) 2 mm above the burner Figure 1: Comparison of jet velocity and the turbulent burning velocity at the base of the flame, calculated using CFD simulations and turbulent burning velocity models. Hydrogen was in central jet at 4 m/s and CO 2 was in annular flow at 18 m/s. With the aid of CFD simulation and the turbulent burning velocity models, we have been successfully in predicting the cases of attached flames and blow out flames. For the lifted flames, the difficulties were in predicting the burning velocity. First of all, the experimental value of laminar burning velocity over a range of equivalence ratio was not always available, especially for blended gases. The second was the issues of turbulent burning velocity models and the influence of large-scale structures and associated intermittence

8 . Conclusions This work showed that adding quenching gases co-currently around the hydrogen jet flame could easily lift hydrogen flame and blow out the jet flame. In this arrangement, propane and CO 2 were equally efficient in lifting and blow out the flame. If hydrogen was premixed with propane or CO 2, the hydrogen/propane mixture possessed a lower laminar burning velocity than the hydrogen/co 2 mixture. Propane addition was much more effective in produce lifted flame and had a higher lift-off height than CO 2 addition. However, both mixtures were effective in blowing out the flame. The present work showed that the cold flow simulation could be used to study the conditions for the flame lift-off and predict the lift-off height. The key difficulty here is the availability of the laminar burning velocity and the viability of the turbulent burning velocity models. References 1. Kalghatgi, G. T., Combustion Science and Technology, Vol 41, pp.17-29(1984). 2. Cheng, T. S. and Chiou, C. R., Combustion Science and Technology, Vol 136, pp.81-94(1998). 3. Brockhinke, A., Haufe, S. and Kohse-Hoinghaus, K., Combustion and Flame, Vol 121, pp (2). 4. Yumlu, V. S., Combustion and Flame, Vol 12, pp14-18(1968).. Leason, B. D., the Forth Symposium (International) on Combustion/The Combustion Institute, 192/pp Milton, B. E. and Keck, J. C., Combustion and Flame, Vol 8, pp.13-22(1984). 7. Vanquickenborne, L. and van Tiggelen, A., Combustion and Flame, Vol 1, pp.9-69(1966). 8. Pitts, W. M., the Twenty-second Symposium (International) on Combustion/The Combustion Institute, 1984/pp Pitts, W. M., Combustion and Flame, Vol 76, pp (1989). 1. Gülder, O. L., the Twenty-third Symposium (International) on Combustion/The Combustion Institute, 199/pp Peters, N., Twentieth Symposium (International) on Combustion/The Combustion Institute, 1988/pp