I n d . Eng. Chem. Res. 1988, 27, 24-30
24
insight into the reaction behavior. The results show that the progress of reaction can be clearly divided into two parts. In the first part, simultaneous conversion of nitrobenzene to azobenzene, azoxybenzene, and aniline occurs. This is accompanied by further hydrogenation of azoxybenzene to azobenzene. In the second part, hydrogenation of azoxy- and azobenzene to hydrazobenzene occurs. The ultimate selectivity to hydrazobenzene, which is the principal consideration in these types of reactions, is governed by the selectivity to azoxy- and azobenzene in the first part of the reaction. The results lead to the following specific conclusions regarding the reaction behavior. (i) Platinum-on-carbon gives significantly higher selectivity for hydrazobenzene as compared to palladiumon-carbon. (ii) An increase in hydrogen pressure increases the rates of all the reactions without affecting the selectivity. (iii) Addition of DMSO improves the selectivity to hydrazobenzene; however, the rates reduce very significantly. (iv) The presence of electron-releasing substituents in the ring results in a decrease in rate as well as selectivity to the partially hydrogenated products.
Table 11. Effect of Substituent on the Selectivity to Partially Hydrogenated Productsa combined selectivity to azoxy and azo compds, nitro comDd mol % 58 o-nitrotoluene 25 o-nitroanisole a Reaction conditions: temperature, 363 K; pressure, 18 atm; impeller speed, 33 rps; nitroaromatic concentration in methanol, 8% (w/w); sodium hydroxide concentration, 2.23% (w/w); catalyst loading, 0.23 (w/w).
to azoxy- and azobenzene increased from 83.3% to 90.2% by addition of DMSO (compare with Figure 1). The rates of the reactions were, however, considerably decreased due to the poisoning effect of DMSO. The reaction was also studied using 2% palladium-oncarbon catalyst. The results are shown in Figure 6. As was expected, the selectivity to hydrazobenzene was observed to be considerably lower. Several substituted hydrazobenzenes are also of practical importance. To investigate the effect of substituents on rate and selectivity, the reaction was carried out employing o-nitrotoluene and o-nitroanisole as reactants. The reaction rates were lower at least by a factor of 2 as compared to nitrobenzene. The selectivity to the partially hydrogenated products was also lower. The selectivities to azoxy and azo compounds a t complete conversion of nitro compounds are compared in Table 11. In the case of onitrotoluene, the azo and azoxy compounds were quantitatively converted to hydrazo compounds, while for onitroanisole no further hydrogenation to hydrazo compound could be observed. The decrease in rate and selectivity with electron-releasing substituents is also observed in partial hydrogenation of nitroaromatics to the corresponding hydroxylamines (Karwa and Rajadhyaksha, 1987). This may be attributed to stronger bonding of the nitrogen atom with the catalyst with an increase in the electron density on the aromatic ring.
Registry No. DMSO, 67-68-5; Pt, 7440-06-4; Pd, 7440-05-3; NaOH, 1310-73-2; KOH, 1310-58-3; CH30Na, 124-41-4; P h N 0 2 , 98-95-3; 2-02NCsHkCH3, 88-72-2; 2-02NCGHdOCH3, 91-23-6; PhNHNHPh,122-66-7; PhNH2,62-53-3; PhN=N(O)Ph,495-48-7; P h N G N P h , 103-33-3.
Literature Cited Ackermann, 0.;Bonsel, P.; Neff, U. Chem. Abstr. 215052. Groggins, P. H. Unit Processes in Organic Synthesis, 5th ed.; McGraw-Hill: Kogakusha, Tokyo, 1958. Karwa, S. L.; Rajadhyaksha, R. A. Ind. Eng. Chem. Res. 1987, 26, 1746-1750. Lubs, H. A. The Chemistry of Synthetic Dyes and Pigments; Robert E. Krieger Publishing: New York, 1955. Russel, G. A.; Geels, E. J.; Smentowski, F. J.; Chang, K. Y.; Reynolds, J.; Kaupp, G. J . Am. Chem. SOC.1967,89(15),3821-3828. Rylander, P. N.; Karpenko, I. M.; Pond, G. R. Ann. N . Y . Acad. Sei. 1970, 172(9),266-275. Stratz, A. M. In Catalysis of Organic Reactions; Kosak, J. R., Ed.; Marcel Dekker: New York. 1984.
Concluding Remarks The paper, which is perhaps the most elaborate study to be reported on this class of reactions, gives considerable
Received f o r review December 23, 1986 Accepted August 11, 1987
A Method for Improving Load Turndown in Fluidized Bed Combustors Robert C. B r o w n * and James E. Foley Department of Mechanical Engineering, Iowa State University, Ames, Iowa 5001 1
We are investigating a new concept in the design of fluidized bed combustors that improves load turndown capability. This design is based on control of heat-transfer rate independent of combustion rate. T h e configuration under investigation consists of a central combustion bed surrounded by an annular fluidized bed that establishes overall heat-transfer rate from the inner combustion bed. The two beds are fluidized independently by separate air plenums. Experiments were performed to evaluate the usefulness of this device in improving load turndown. For a given combustor firing rate and air-to-fuel ratio, a wide variation in combustion temperature could be achieved by controlling the flow of fluidization air to the annular fluidized bed. In other tests, a load turndown ratio of 8.7 was achieved while combustion temperature was held constant. 1. Introduction
Use of fluidized bed combustion (FBC) has increased with the realization that it can burn coal and low-grade fuels in an environmentally acceptable manner. Long fuel-residence times in fluidized beds provide fuel flexibility, while use of inexpensive sorbents for bed material and staged firing reduce emissions of sulfur dioxide and 0888-5885/88/2627-0024$01.50/0
nitrogen oxides, respectively. Unfortunately, technical problems remain that must be overcome before wider markets are developed. Prominent among these difficulties is the poor load turndown capability of fluidized bed combustors; this capability is defined here as the ratio of maximum to minimum firing rates. Even modest load turndowns are frequently accompanied by degradations 0 1988 American Chemical Society
Ind. Eng. Chem. Res., Vol. 27, No. 1, 1988 25
1
EXHAUST
n-n
HEAT TRANSFER COMBUSTION BED
"mf
SUPERFICIAL G A S VELOCITY
Figure 1. Dependence of combustion and heat-transfer rates on fluidization velocity in a fluidized bed.
in combustion efficiency and pollutant emissions. Because load turndown capability is especially important for fluidized beds targeted for coal-fired, gas-turbine power systems and small-scale boilers and furnaces, we are investigating a new concept in fluidized bed design that improves load turndown capability. Load turndown in FBC is a function of the heat-transfer rate from the fluidized bed. Heat transfer from a fluidized bed to a water jacket or water tubes is determined by three factors: (1)the temperature gradient between bed and water, (2) the heat-transfer area, and (3) the overall heat-transfer coefficient between bed and water. Boiler application usually sets the water-side temperature; attempts to control load turndown with temperature gradients require large variations in bed temperature. However, even small variations in bed temperature from optimal design values will greatly degrade both sorbent utilization (Roberts et al., 1975) and combustion efficiency (Anson, 1976). Reduction of heat-transfer area has been suggested as a method for reducing loads in FBC. This condition can be accomplished by either reducing fluidization velocity, which contracts bed volume, or discharging bed material. The former approach is of little practical value because bed contraction is limited to about 30%; thus, the corresponding load turndown is modest a t best. Discharging, storing, and reinjecting hot particles are fraught with many technical difficulties and offer few real advantages as a method for load turndown. In addition to the abovementioned difficulties, both methods for reducing heattransfer area will expose tubes to erosion when they are in the splash zone of the bed. Another method for reducing heat-transfer area requires the air distributor to be partitioned; this allows zones of the bed to be independently fluidized. Load turndown is achieved by selectively slumping part of the bed. The heat-transfer area in defluidized zones is effectively zero. This partitioning technique has some undesirable effects on combustion including fuel smoldering and agglomeration in the slumped regions. Although bed slumping is frequently employed in commercial FBC units, turndown capability is rather modest. Variation in the overall heat-transfer coefficient between bed and tubes can also be employed for load turndown control. Heat-transfer coefficients in fluidized beds show large variations with fluidization velocity; in principle, turndown ratios exceeding 10 can be achieved by reducing fluidization velocity from its maximum heat-transfer value to the minimum fluidization condition. However, as Figure 1 illustrates, the dependence of heat-transfer rate on
-
HEAT TRANSFER BED FLUIDIZATION AIR
t
COMBUSTION BED FLUIDIZATION AIR
Figure 2. Schematic of bed geometry for improving load turndown in fluidized bed combustors.
fluidization velocity is strongly nonlinear above the minimum fluidization velocity, Umf(Kunii and Levenspiel, 1969). Since combustion rate is proportional to fluidization velocity, a match between heat release and heat transfer is difficult to achieve. Horio et al. (1985) have developed a baffled heat-transfer tube in an attempt to achieve a linear response in average heat-transfer coefficient with changes in fluidization velocity. Although they were successful in obtaining a linear response in the velocity range of 0.3-0.5 m/s, this achievement represents only a modest turndown ratio. It is far from evident that a sufficiently linear response can be achieved over larger velocity intervals. A more promising approach to improved load turndown is control of heat-transfer rate independent of combustion rate (Brown and Buttermore, 1986). The device as described here can be employed in fluidized beds that remove heat around the perimeter of the bed, that is, water jackets or water wall construction; however, the same principle can also be applied to vertical water tube designs. Independent control of heat-transfer rate and combustion rate is accomplished by surrounding the fluidized bed in which fuel is burned (combustion bed) by another fluidized bed (heat-transfer bed) that establishes the overall heattransfer rate from the inner combustion bed. The two beds, physically divided by a wall, are fluidized independently by separate air plenums. Figure 2 illustrates a schematic of the device employed in a water-jacketed, cylindrical fluidized bed. The central combustion bed is provided with fluidization air through a circular distributor plate from an air plenum that is designed to give even distribution of air through the bed. Coal or other fuel is fed into the combustion bed at a rate determined by the desired heat generation rate, while air flow into this bed is set at a level consistent with good combustion. The heat-transfer bed is the annular fluidized bed surrounding the combusion bed. The wall separating the beds is constructed of heat-resistant material of reasonably high thermal conductivity. The heat-transfer bed is supplied with fluidization air from a plenum separate from the combustion bed plenum. The heat-transfer bed is surrounded by a water jacket that removes heat from the combustor in the form of hot water or steam.
26 Ind. Eng. Chem. Res., Vol. 27, No. 1, 1988
Overall heat-transfer rate from the combustion bed to the water jacket is determined by the heat-transfer coefficients associated with the boundary layer of the combustion bed, the conductivity of the wall separating the beds, the boundary layers at the inner and outer diameters of the heat-transfer bed, the conductivity of the water jacket wall, and the boundary layer of the water in the jacket. Control of the overall heat-transfer rate is accomplished by changing the fluidization velocity of air entering the annular bed. The combustion bed can be operated at air flow rates consistent with good combustion and independent of heat-transfer considerations. The heat-transfer bed can be operated in region A of Figure 1 where large variations in heat-transfer rate can be achieved. If no air is passed through the heat-transfer bed, then it has the poor heat-transfer characteristics of packed granular beds. If sufficient air is passed through the heat-transfer bed to achieve minimum fluidization, then increased heat transfer due to bed expansion and convection occurs. Heat transfer continues to increase as air flow increases until heattransfer characteristic of bubbling fluidized beds is reached; the result is a continuous and large variation in heat-transfer rate that is controlled independently of combustion rate. We have constructed an annular heat-transfer bed for testing in a 25.4-cm-diameter fluidized bed combustor. Experiments were performed to evaluate the usefulness of this device for improving load turndown in fluidized bed combustors.
2. Theoretical Performance The performance of the heat-transfer bed depends on such factors as fluidization velocity, bed material composition and particle size, width of the bed, and construction of the wall separating the two beds. A simple analysis provides an estimate of the load turndown capability of this device. Let Q be the heat-transfer rate from the combustion bed through the retaining walls of the bed. Load turndown can be approximated as the ratio of Q for full fluidization of the heat-transfer bed to Q for the slumped heat-transfer bed. For steady-state operation of the combustor, the maximum heat transfer from the combustion bed can be approximated by
where hc is the heat-transfer coefficient for the combustion bed, hH is the heat-transfer coefficient for the heat-transfer bed, hw is the heat-transfer coefficient for the water, AF is the heat-transfer area in the fluidized heat-transfer bed, Tc is the temperature in the combustion bed, and T,. is the temperature in the water jacket. In deriving this equation, it is assumed that the combustion and heat-transfer beds are uniform in temperature when fluidized because of rapid mixing. In addition, the heat-transfer bed is assumed to be deep compared to its radial dimension; hence, heat loss associated with energy convected out of the heat-transfer bed with the fluidization air is relatively small. This simplification produces an approximate 20% underestimate in the turndown ratio for the calculations that follow. In the case of minimum heat transfer from the combustor, the heat-transfer bed is completely defluidized:
I
I
I
Figure 3. Experimental apparatus.
where k is the thermal conductivity of bed material, Ax is the width of the heat-transfer bed, and A M is the heat-transfer area in the unfluidized heat-transfer bed. Further simplification is obtained if it is assumed, for the maximum heat-transfer condition, that the beds are equally fluidized and employ identical bed material. If boiling heat transfer is employed in the water jacket, then hH = hc
