Experimental and numerical investigation of vortex-induced flame propagation in a biomass furnace with tangential over fire registers.
The vortex-combustion units have become more attractive in the field of power station firing systems. They have been used extensively throughout the world with wide applications in many types of steam boilers, including both small and high capacity units. Tangentially fired furnaces are units with four, six, eight, or more sides, with fuel and air supplied tangentially from the furnace corners or near them (Romadin, 1973). Furnaces of this type are essentially vortex-combustion units. The limits of ignition are found to depend on the size of the vortex and the inclination angle (Habib et al., 1992). The main advantage of a tangentially fired system lies in reducing the NOx. NOx in tangentially fired units is lower than other firing types. NOx emissions from tangentially fired boilers are about half the values from wall firing systems (El-Mahallawy and El-Din Habik, 2002). Also, the tangential-firing technique is characterized by lower carbon losses, and greater adaptability for the combustion of fuels with low calorific value or low volatile contents. Vortex motion at the furnace centre prevents or minimizes slugging of the furnacewalls, erosion due to impingement and local over-heating.
Tangentially fired systems are best suited for fuels such as coal, oil, municipal waste, and agricultural wastes, because of the rapid contact with fuel and air, flame impingement, and the increased particulate residence time due to vortex motion (Hamilton, 1979; Whaley and Rankin, 1987). The tangentially fired units have a good record in being able to meet emission regulations on NOx as a result of their flexibility and the ability to control the heat release rate (Mundez-Lcanada and Rossano-Roman, 1983).
The drawback with tangentially fired furnaces is burner velocity. Low velocities are not suitable for fuels having high volatile contents, as ignition occurs in or near the burner causing slugging and distortion problems. Very high velocities on the other hand are undesirable as fuel particles can centrifuge out of the main combustion zone as unburnt carbon.
Some of the boilers being operated in power stations are tangentially fired with pulverized coal/oil or gas fuels. Research on a fully three-dimensional numerical model has been developed to predict the gas flow, particle trajectories, combustion, and heat transfer in the furnace. The predicted velocity vectors in the plane of the heated jet had a general clockwise flow pattern set-up. The predicted and measured mixture fractions are in good agreement. In the present furnace, the flow path of the vortex is in an anticlockwise direction.
Research has been conducted on the combustion and radiation heat transfer in tangentially fired furnaces (Knill et al., 1993). The calculations are made with the boiler operation at various operating loads and compared the existing temperature and gas composition data. The results showed that the jet behaviour influenced the circulation in the boiler.
[FIGURE 1 OMITTED]
Various works have been carried by researchers related to the detailed temperature, heat flux, and performance for operating conditions like load, excess air, and compared them with predictions obtained from a three-dimensional heat transfer model (Categen and Richter, 1987).
The ideas of micromixing in tangentially fired furnaces and the effects of turbulent mixing of incompressible fluids on chemical reactions have been investigated (Baldyga and Pohorecki, 1995). They indicated that stretching material elements and vortices in addition to molecular diffusion results in the growth of the mixing zones. A three-dimensional numerical model to predict the temperature and species concentration in a power station boiler was established and found that the measured experimental data were in good agreement with the computational measurements (Coelho and Carvalho, 1996). The flow field, heat transfer and emissions for a large-scale tangentially fired pulverized coal furnace have been numerically modeled and, compared with the existing values in site operation records, are in good agreement (Yin et al., 2002).
Computational fluid dynamics (CFD)-based analysis has gained increasing acceptance in recent years as a tool in the evaluation of combustion systems. Although reacting CFD models involve varying degrees of approximation, they still provide valuable insight into the performance of furnaces. With the development of increasingly affordable and powerful computers, numerical simulation has become an engineering tool for evaluating furnace designs. With the advent of numerical models for combustion, many researchers have employed increasingly more sophisticated models and have demonstrated the potential of such models as engineering tools (Gibson and Morgan, 1970; Lockwood et al., 1980; Fiveland et al., 1984; Boyd and Kent, 1994; Bockelie et al., 1998; Cremer et al., 2002).
The furnace selected for the present research has tangential over fire air registers to aid the suspension burning. The bagasse particles coming out of the spreader are accelerated by the high velocity distributor air jets. The velocity of the distributor air is 35 [ms.sup.-1]. There are five bagasse spreaders on the front wall located at a height of 2.36 m above the grate. The smaller particles carried upwards by the combustion air while larger particles settle on the grate. Here the preheating of the particles will be carried below the tangential over fire air registers plane. Finally, the heated particles enter the vortex and get complete combustion as the vortex moves up. Since the tangential air is coming separately in all the four tangential registers, the problems of high and low velocities are avoided in this furnace. The temperature of the furnace wall was measured 353[degrees]C. The temperature of the combustion air was measured 210[degrees]C. The velocity of the tangential over fire air jet was measured 22.5 [ms.sup.-1] and mass flow rate was 3.75 kg [s.sup.-1]. The operating fuel ratio was 3.79:1. Temperatures were recorded at various levels of the furnace by using a calibrated chromel-alumel thermocouple and a digital indicator with the accuracy of [+ or -]2[degrees]C.
Bagasse normally comes with higher moisture content around 50% by weight and also in different sizes. They have their own influence on the combustion. Certain thermal investigations were carried on sophisticated instruments for the bagasse as a part of the investigation. Since the bagasse contains high moisture, it absorbs energy before its thermal degradation. This is studied by conducting differential scanning calorimetry analysis for various sized particles at lower temperature of 10[degrees]C/min, under the atmosphere of nitrogen. The results obtained clearly explain the difference in the amount of energy absorbed before their degradation. The thermal degradation and rate of degradation of bagasse particles were studied under thermogravimetric and differential thermal analysis for two different sizes. There are five spreaders on the front wall of the furnace with distributor air ducts at the bottom. The tangential over fire registers are four in number and located on the side walls. The angles of inclination of these ducts are 76 and 41[degrees], respectively.
[FIGURE 2 OMITTED]
Figure 1 shows the TG-DTA analysis results for the bagasse particles with size 250 [micro]m, under the atmosphere of air. A sample of 2.590 mg was used for the test, with a heating rate of 40[degrees]C/min up to 1200[degrees]C. There is a loss of 10% in weight from the temperature 34.48 to 100[degrees]C. From 100 to 220[degrees]C, the sample is thermally stable. The thermal decomposition begins at 220[degrees]C and continues up to a peak temperature of 372.57[degrees]C. Later passive pyrolysis follows from 372.57 to 500[degrees]C. From 500[degrees]C onwards, the degradation becomes stable and attributes to ash. Major thermal decomposition of the fuel occurs between 220 and 372.57[degrees]C. The TG-DTA analysis result of 250 [micro]m particles are tabulated in Table 1. The rate of degradation of 250 [micro]m particles is shown in Figure 2. The maximum rate of degradation of -5.889 mg/min occurs at 337.54[degrees]C. Similarly the TG-DTA analyses of larger particles with more than 3000 [micro]m size are shown in Figures 3 and 4.
[FIGURE 3 OMITTED]
The bagasse particles are analyzed by differential scanning calorimetry tests to discover their heat absorption capacity before the exothermic reaction. The particles sieved by the standard sieves were selected for test. The tests were conducted at low temperature of 10[micro]C/min and in the atmosphere of nitrogen. The DSC results for various particles are shown in Figure 5. The details of the analyses are tabulated in the Table 2. It is found that the 250 [micro]m particles absorb around -3.4 mW of energy during endothermic reaction and found -4.6 mW for 2057 [micro]m particles. From the results, it is clear that the larger particles take more time to absorb energy before the exothermic reaction starts.
A three-dimensional computational model of the furnace is developed by using FLUENT package. The segregated implicit solver is used for solving the transport equations. The turbulence is modelled by standard k-[epsilon] model. Radiation is modelled by P1. Species transport is used for combustion and combustible particles are assumed for bagasse particles. Bagasse particles are tracked by Lagrangean method. The grate is modelled as a porous media model including flow through the packed bed. The packed bed is modelled using both permeability and an inertial loss coefficient. The pressure drop across the packed bed may be calculated by the so-called Ergun equation, a semi-empirical correlation applicable over a wide range of Reynolds numbers and for many types of packing as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
where [[V.bar].sub.s] is the superficial velocity, [epsilon] the void fraction, [bar.d] the mean particle diameter, L the depth of bed, [rho] the density of the gas, and [DELTA]p the pressure drop across the bed. The SIMPLE algorithm is used to solve the equations.
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
Figure 6 shows the path lines of the particles. The bagasse particles are accelerated by the jets of distributor air at the spreaders. As the particles travel toward the sloping back wall of the furnace, they are influenced by the under grate air and tangential over fire air system. The larger particles which are heavy fall in front of the spreader on the grate. The majority of the medium particles travel the rear end of the furnace and gets combustion. The inner side of the vertical flame forms a recirculation zone attached to the front wall. This will help in complete combustion of the particles which are swept away by the under grate air.
Figure 7 shows the temperature distribution at various planes along the furnace height. It is clear from the results that the vortex formation will be initiated at the tangential over fire air registers level. Later this vortex is induced by the under grate air. The shape of the vortex formed remains circular up to the neck (from 3.5 to 10 m.) At the neck, size reduces and the temperature increases to maximum of around 1400[degrees]C. The temperature distribution is found maximum along the width of the rear end of the furnace at spreader level due to the combustion of bagasse particles at the rear end. This is due to the delay caused by the moisture content of the particle. At the height of 3.5 m, where tangential over fire registers are located, the jets of air influences the path lines of the particles coming out of the spreaders. At tangential plane, higher temperature is found on the front wall, rear wall, and the side walls. This is due to the effect of tangential air jets. As the flame moves upwards toward the neck, the swirling effect will be accelerated and temperature rises to the maximum near the neck of the furnace, which is then leading to super heaters zone.
[FIGURE 9 OMITTED]
[FIGURE 10 OMITTED]
Figure 8 shows the temperature distribution at a height of 7.5 m above the grate. It shows the higher temperatures on the walls than at the centre.
Figure 9 shows the vortex formation and shape at various levels of the furnace, from the level of 3.5 m above the grate to the neck level of the furnace. At the tangential registers level, the vortex is influenced by the jets of distributor air. There will be a rapid and random movement of the particles at this zone. But slowly as it traverses it gains its own vortex form and traverses along the height of the furnace. A small recirculation zone over the tangential registers at the height of tangential plane is found. These recirculation zones extend up to 5.5m and become negligible size at the neck.
The predicted temperatures at various heights above the grate are tabulated in Table 3. It is found that the temperatures on the walls are higher than at the centre. The vortex formed at the tangential registers plane moves upwards in the furnace, influenced by the under grate air. Due to the combustion of particles in suspension, the temperature keeps on rising as the vortex moves. The temperature attains the maximum value of about 1400[degrees]C at the neck of the furnace, which then leads to a super heater zone. Also, the temperature on the front and rear walls are much higher than the side walls. This is due to the effect of tangential air jets at the tangential plane and also due to the recirculation of the particles toward the front wall. Since bagasse particles are accelerated by the high velocity distributor air jets, the particles reach the sloping back wall where majority of the particles get combustion. This is the reason for the higher temperature on the rear side of the furnace. The details of the temperature measured and predicted at various heights of the furnace are given in Table 4.
Figure 10 shows the mole fraction of the nitrogen at various levels of the furnace. The effect of swirling flame reduces the amount of nitrogen content. From the result, it is clear that the nitrogen content is high at the entry of the tangential registers and at the rear end of the furnace. The amount of nitrogen reduces as the height increases. The computational values are tabulated in the Table 5. The mole fraction of the nitrogen measured at the induced draft fan was found 0.632 mol. Also, other species measured in the flue gas had 0.093 mol of C[O.sub.2], 0.207 mol of [H.sub.2]O, and 0.068 mol of [O.sub.2].
The particle temperature profiles for two selected particles along their path and time are given in Figures 11 and 12. The thermal behaviour of the particles depends on the size as well as their moisture content. The temperature of the smaller particles rises faster than the larger particles due to their small size and larger surface area. The density of the larger particles are more and are influenced by the moisture.
[FIGURE 11 OMITTED]
Figure 13 shows the temperature distribution inside the furnace above the grate at various heights. The temperature near the front wall rises from 950 to 1380[degrees]C, almost linearly. This is due to the recirculation of the combustion air on the front wall and due to convection. On the rear side, up to 7.5 m the temperature is in increasing trend, afterwards decreases. This is due to the convergence of the flue gas towards the neck. The temperatures on the side walls are influenced by the angle of inclination of the tangential air registers. The temperature is higher near the south wall than the north wall. The difference becomes minimum at the level of neck. Similarly the effect of these temperature distributions on the molar concentration of [N.sub.2] is shown in Figure 14.
[FIGURE 12 OMITTED]
[FIGURE 13 OMITTED]
[FIGURE 14 OMITTED]
The simulation results and the measurements show the increase in the temperatures as the flame traverses upwards. The maximum temperature occurs at the neck level. The formation of vortex starts at the tangential over fire air jets. But it gains its full momentumas it traverses up. The vortex will be affected by the distributor air jets in the beginning. Later it is influenced by the under grate air. Also from the results, it is clear that the molar fraction of nitrogen reduces as the vortex travels in the upward direction.
In this furnace, the bagasse particles are preheated by the combustion air coming from the grate as soon as they enter the combustion chamber from the spreaders. The furnaces of this kind overcome the problems of higher velocities or lower velocities affecting the combustion in traditional tangentially fired furnaces. The results of DSC tests show that the particles with larger sizes absorb more energy during endothermic reaction before its thermal decomposition. The thermogravimetric analysis shows that the complete thermal decomposition of the particles occurs by 500[degrees]C. The behaviour of the particles inside the combustion chamber varies according to their size as well as their moisture content.
The author wish to thank Dr. K. Ramachandran, School of Mechanical & Building Sciences, Vellore Institute of Technology, Vellore University, Vellore for the invaluable suggestions extended in preparing the manuscript.
Manuscript received June 16, 2006; revised manuscript received April 27, 2007; accepted for publication June 9, 2007.
Baldyga, J. and R. Pohorecki, "Turbulent Micro Mixing in Chemical Reactors--A Review," Chem. Eng. J. 58, 183-195 (1995).
Bockelie, M. J., B. R. Adams, M. A. Cremer, K. A. Davis, E. G. Eddings, J. R. Valentine, P. J. Smith and M. P. Heap, "Computational Simulations of Industrial Furnaces," Computational Technologies for Fluid/Thermal/Chemical Systems with Industrial Applications, ASME, 1998, pp. 117-124.
Boyd, R. K. and J. H. Kent, "Comparison of Large Scale Boiler Data with Combustion Model Predictions," Energy Fuels 8, 124-130 (1994).
Categen, B. M. and W. Richter, "Heat Transfer Modeling of a Large Coal Fired Utility Boiler and Comparisons with Field Data," ASME, New York, NY, U.S.A. (1987), Vol. 1, pp.225.
Coelho P. J. and M. G. Carvalho, "Evaluation of a Three-Dimensional Mathematical Model of a Power Station Boiler," J. Gas Turbines Power 118, 887-895 (1996).
Cremer, M., B. Adams, J. Valentine, J. J. Letcavits and S. Vierstra, "Use of CFD Modeling to Guide Design and Implementation of Overfire Air for NOx Control in Coal-Fired Boilers," Proceedings of Nineteenth Annual International, Pittsburgh Coal Conference, Pittsburgh, PA, U.S.A. September 23-27, 2002, pp. 1-9.
El-Mahallawy, F. and S. El-Din Habik, "Fundamentals and Technology of Combustion," 1st ed., Elsevier Science Ltd., Oxford, U.K. (2002).
Fiveland, W. A., D. K. Cornelius and W. J. Oberjohn, "COMO: A Numerical Model for Predicting Furnace Performance in Axisymmetric Geometries," ASME Paper No. 84-HT-103, (1984).
Gibson, M. M. and B. B. Morgan, "Mathematical Models of Combustion of Solid Particles in a Turbulent Stream with Recirculation," J. Inst. Fuel 43, 517-523 (1970).
Habib, M. A., F. M. El-Mahallaway, A. Abdel-Hafez and N. Nasseef, "Stability Limits and Temperature Measurements in a Tangentially-Fired Model Furnace," Energy Inst J. 17, 283-294 (1992).
Hamilton, T. J., "Computer Control System for Multifuel Industrial Boilers," Power Eng., Haney Well 79, 56 (1979).
Knill, K. K., E. Chui and P. M. Hughes, "Application of Grid Refinement in Modeling a Tangentially Fired Coal Boiler," FACT, Combustion Modeling, Co-firing and NOx Control, ASME, 1993, pp. 115-121.
Lockwood, F. C., A. P. Salooja and S. A. Syed, "A Prediction Method for Coal-Fired Furnaces," Comb. Flame 38, 1-15 (1980).
Mundez-Lcanada, E. and M. Rossano-Roman, "Proc. of the 5th Power Plant Dynamics," Control and Testing Symposium, Konxville, TN, U.S.A. May 21-23, 1983, 1-901.
Romadin, V. P., "Furnaces with Corner Firing Tangential Burners," Thermal Eng. 20, 79-89 (1973).
Whaley, F. M. and D. M. Rankin, "Coal-Water Fuel Combustion in Boiler," Canada Energy Mines Resourcers, Canada, 88, 75 (1987).
Yin, C., S. Caillat, J. L. Harion, B. Baudoin and E. Perez, "Investigation of Flow, Combustion, Heat Transfer and Emissions from a 609 mW Utility Tangential Fired Pulverized Coal Boiler," Fuel 81, 997-1006 (2002).
K. S. Shanmukharadhya (1) * K. G. Sudhakar (2)
(1.) Department of Mechanical Engineering, Bannari Amman Institute of Technology, Sathyamangalam 638 401, Tamil Nadu, India
(2.) Advanced Studies, Government Tool Room and Training Centre, Rajajinagar, Bangalore 44, Karnataka, India
* Author to whom correspondence may be addressed. E-mail address: kss email@example.com
Table 1. TG-DTA analysis results of bagasse particles Particles 250 [micro]m Larger particles Sample weight 2.481 mg 3.822 mg Atmosphere Air Air Gas flow rate 100 mL/min 100 mL/min Heating rate 50[degrees]C/min 50[degrees]C/min Heating range 40-1200[degrees]C 40-1200[degrees]C Peak temperature 372.57[degrees]C 411.22[degrees]C Area -8549.708 mJ -10769.254 mJ Delta = H -3446.0733 J/g -2817.7012 J/g Maximum particle degradation rate -5.889 mg/min -7.602 mg/min Table 2. DSC analysis result of bagasse particles with different sizes Parameters Large particles 2057 [micro]m Peak 1 Peak 2 Integral (mJ) -601.71 -559.10 -8.86 Normalized ([Jg.sup.-1]) -138.93 -198.69 -3.08 Onset ([degrees]C) 38.51 38.25 126.01 Peak ([degrees]C) 83.21 81.35 130.38 End set ([degrees]C) 121.15 117.49 138.19 Sample weight (mg) 4.3310 2.8140 Heating rate 10[degrees]C/min 10[degrees]C/min Parameters 1003 [micro]m Peak 1 Peak 2 Integral (mJ) -422.93 -8.86 Normalized ([Jg.sup.-1]) -249.37 -5.22 Onset ([degrees]C) 30.12 134.07 Peak ([degrees]C) 73.42 135.94 End set ([degrees]C) 110.09 140.57 Sample weight (mg) 1.6960 Heating rate 10[degrees]C/min Parameters 710 [micro]m Peak 1 Peak 2 Integral (mJ) -381.46 -12.32 Normalized ([Jg.sup.-1]) -212.27 -6.86 Onset ([degrees]C) 37.38 120.13 Peak ([degrees]C) 75.73 124.91 End set ([degrees]C) 111.46 131.40 Sample weight (mg) 1.797 Heating rate 10[degrees]C/min Parameters 250 [micro]m Peak 1 Peak 2 Integral (mJ) -596.65 -11.64 Normalized ([Jg.sup.-1]) -242.44 -4.73 Onset ([degrees]C) 34.05 135.55 Peak ([degrees]C) 78.22 140.92 End set ([degrees]C) 116.59 147.43 Sample weight (mg) 2.461 Heating rate 10[degrees]C/min Table 3. Details of temperature near the furnace walls Furnace height above Temperature near furnace walls in [degrees]C the grate in metre Front wall Rear wall North wall At 3.5 m 1020 1030 950 At 4.5 m 1150 1130 965 At 5 m 1110 1150 1000 At 5.5 m 1110 1180 1000 At 7 m 1250 1360 1200 At 7.5 m 1280 1380 1320 At neck 1400 1120 1380 Furnace height above Temperature near furnace walls in [degrees]C the grate in metre South wall Centre At 3.5 m 962 780 At 4.5 m 1110 818 At 5 m 1110 818 At 5.5 m 1120 820 At 7 m 1280 890 At 7.5 m 1300 900 At neck 1320 1003 Table 4. Details of temperature measured Measurements at the various Measured Predicted heights of the furnace temperature temperature ([degrees]C) ([degrees]C) At 3.5 m above the grate rear 987 1030 At 5.5 m above the grate from rear 1128 1180 At 8.2 m above the grate from rear 1320 1350 At 10.7 m above the grate from front 1364 1400 Table 5. Mole fraction of [N.sub.2] at various heights of the furnace Furnace height Mole fraction of [N.sub.2] at various levels of the above the grate furnace in metre Front wall Rear wall North wall At 3.5 m 0.446 0.751 0.649 At 4.5 m 0.446 0.649 0.446 At 5.5 m 0.406 0.620 0.560 At 7.5 m 0.487 0.550 0.568 At neck 0.527 0.406 0.580 Furnace height Mole fraction of [N.sub.2] at various levels of the above the grate furnace in metre South wall Centre At 3.5 m 0.645 0.203 At 4.5 m 0.530 0.243 At 5.5 m 0.440 0.284 At 7.5 m 0.485 0.284 At neck 0.424 0.284