Co-firing of Rice Husk and Bagasse in a Conical Fluidized-bed Combustor

Transcription

1 Co-firing of Rice Husk and Bagasse in a Conical Fluidized-bed Combustor Kasama Janvijitsakul 1,*, Vladimir I. Kuprianov 1 2 and Watchara Permchart 1 Mechanical Engineering Program, Sirindhorn International Institute of Technology, Thammasat University, Pathumthani, Thailand 2 Department of Agricultural Engineering, King Mongkut s Institute of Technology Ladkrabang, Bangkok, Thailand Abstract: In this work, as-received rice husk and sugar cane bagasse were co-fired in the cone-shaped fluidized-bed combustor with the aim of effective and environmentally friendly utilization of these biomass fuels. Temperatures and volume gas concentrations (O 2, CO and NO) were measured at different location points along the combustor height and in the stack flue gas when co-firing the two biomass fuels at different mass/energy fractions as well as when using rice husk only. Axial temperature and gas concentration profiles in this combustor operated at kg/h fuel feed rates and various values (4 1%) of excess air, EA, for the 1%, 75% and 45% rice husk mass fractions in the fuel blend, are discussed. The axial temperature profiles were found to be almost independent of excess air and strongly affected by the rice husk energy fraction, EF r h, i.e. contribution by rice husk to the heat input when rice husk and bagasse co-firing. The CO emissions, or CO concentrations in the flue gas leaving the combustor, were found to reduce for higher values of EF r h and EA. Meanwhile, the NO concentrations at all the point over the combustor volume as well as in the stack flue gas were increased for higher EF r h and EA values. Co-firing of rice husk and bagasse led to higher combustion efficiency at reduced environmental impacts compared with those in the case of firing rice husk only. Through the co-combustion with rice husk, an effective use of as-received sugar cane bagasse becomes feasible for energy production in fluidized-bed combustion systems. Keywords: Axial Profiles, Temperature, Gas Concentrations, Emissions, Combustion Efficiency. 1. INTRODUCTION Rice husk and sugar cane bagasse are important biomass resources for heat and power generation in Thailand. Annually, about 2 million tons of rice and 5 million tones of sugar cane are produced in this country. Accordingly, tremendous amounts of rice husk and sugar cane bagasse, residues from processing of rice and sugar cane, are available as energy sources. Despite predominant portions of rice husk and bagasse are utilized by the Thai milling industries, significant amounts of these biomass fuels are being unused and eventually lost. Annual losses of rice husk and bagasse in this country are estimated to be and million tones, respectively. The corresponding aggregate power generation potential from the unused rice husk and bagasse is estimated to be MW e [1]. A significant amount of research work has been recently carried out on fluidized-bed combustion of rice husk, sawdust and other agricultural residues. Some literature references are focused on effects of operating conditions (e.g. excess air for conventional combustion, or secondary-to-total air ratio and fuel mass fractions for fuel co-firing) on temperature patterns, combustion efficiency and emission characteristics of fluidized bed systems fuelled with various agricultural residues [2 6]. As shown by different authors, the fluidized-bed combustion technology represents the most effective and environmental friendly technology for conversion of rice husk into energy. However, there are some problems related to the fluidized bed combustion of rice husk with high, up to 2%, ash contents in relatively short combustors. Under such conditions, the combustion efficiency of a reactor is rather low, of 81 86%, for different options of combustion air supply. These facts are mainly explained by significant values of unburned carbon in fly ash discharged from the reactor [2,3]. However, the combustion of medium-ash rice husk in advanced systems (e.g. in a circulating fluidized bed), is characterized by higher values of the combustion efficiency for wide ranges of operating conditions [5]. Meanwhile, the combustion of rice husk in fluidized-bed systems is accompanied by significant environmental impacts [2 6]. Because of elevated fuel-n, NO x emissions from conventional combustion systems are found to be in the range of about 1 to 18 ppm when burning this biomass fuel at the excess air values of some 2 to 1%, respectively. For low excess air (less than 4%), CO emission from the rice husk-fuelled combustor is very high, greater than 5 ppm, and strongly dependent on excess air. On the contrary, at excess air above 6%, the CO emission from the combustor is reduced to 6 11 ppm and almost independent of this operating variable [2]. Attempts to burn as-received sugar cane bagasse in a fluidized bed system are reported to be unsuccessful because of the high moisture content in raw bagasse [2]. Moreover, the high fuel moisture and the relatively large size of the fuel particles (of about.4 21 mm on average) of this fiber residue cause significant operational problems associated with fuel feeding. However, as shown in Ref. [2], sugar cane bagasse can be effectively fired (at about 99% combustion efficiency) at relatively weak environmental impacts compared with those from the rice husk combustion. This paper deals with an experimental study on co-firing of as-received rice husk and bagasse in a fluidized-bed combustor with the aim of achieving an effective utilization of the selected fuels via mutual elimination (or compensations) of the above drawbacks. Effects of the mass/energy fraction of rice husk in fuel blends as well as of operating conditions on major thermal and emission characteristics of the combustor were the focus of this study. 2. MATERIALS AND METHOD 2.1 Fluidized-bed combustor and auxiliary equipment The experimental tests were carried out on a conical fluidized bed combustor (FBC) consisting of two parts: cylindrical section (of the.9 m inner diameter and 2 m height) and conical section (of the 4 o cone angle and 1 m height). The combustor body was insulated with the 5-mm ceramic-fiber material ensuring minimized heat losses across the walls. A schematic diagram of the conical FBC is shown in Fig. 1. Silica sand (SiO 2 9%) of about.3.5 mm particle size and 4-cm static bed height was used as the inert bed material to ensure its effective fluidization in this conical FBC [2,3]. Corresponding author: kasama@siit.tu.ac.th 232

2 Fig. 1 Schematic diagram of the conical fluidized-bed combustor. A LPG-firing pilot burner was used for preheating the bed material during combustor start-up modes. The pilot burner was fixed at a.6-m level above the air distributor. Upon attaining the solid-air bed temperature of about 55 C (which was sufficient for the sustaining of ignition and combustion of the biomass fuels), the pilot burner was switched off and removed from the combustor with the aim to protect the burner from overheating when the major (biomass) fuel was fed into the combustor. For the fuel feeding, the conical FBC was equipped with a screw type feeder supplying biomass fuel (rice husk or fuel blend) over the bed at a.65-m level above the air distributor. A 25-hp blower supplied combustion air (under ambient conditions) through the air distributor located at the bottom of the conical part. The air distributing plate with nine air-bubble caps was connected to the bottom flange of the conical part. Each bubble cap was designed like a standpipe of 25.4-mm outer diameter and 5-mm height for providing the combustion air injection below the layer of the bed material through 56 holes. These holes of 2-mm diameter were declined at 45 o with respect to the horizontal plane and evenly distributed over the standpipe cylindrical surface. In addition, there were four vertical slots of mm in size, located on the top of each standpipe and used for the air injection. The volume flow rate of combustion air was measured in the tests with the use of a U-tube manometer installed downstream from the blower. Simultaneously, a Testo-35 gas analyzer was used for quantifying excess oxygen in the stack flue gas, i.e. at the exit of the ash-collecting cyclone (see Fig. 1). The Testo-35 gas analyzer was also employed to monitor gas concentrations of major pollutants (CO and NO) in the combustor as well as in the stack flue gas. During this procedure, combustion products were sampled through the holes in the combustor walls fixed at different levels (with respect to the air distribution plate) along the combustor height as well as at the cyclone exit. The relative measurement errors were expected to be about 1% for O 2 and 5% for CO and NO. Seven chromel-alumel thermocouples (of type K) were fixed at different levels along the combustor height and at the exit of the ash-collecting cyclone for temperature measurements (in the relative measurement error of about 1%) in the flue gas. 2.2 Fuel properties Raw (i.e. as-received ) rice husk and bagasse were used in experimental tests for which the fuels were blended at certain mass/energy proportions. Table 1 shows the fuel properties and their lower heating values (LHV), the latter being estimated by Ref. [7] using as-received fuel analyses for both fuels. As seen in Table 1, the moisture content in as-received bagasse is rather high (48.8%). All attempts to burn the raw sugar cane bagasse as well as the fuel blends at relatively low values of the rice husk mass fraction failed in preliminary tests of the conical FBC, mainly because of cooling down of the bed material below the temperature of stable ignition during start-up modes. The blended fuels used in these experimental tests were, therefore, fired at the 75% and 45% rice husk mass fractions. In addition, the experimental tests for firing rice husk only (i.e. for the 1% rice husk mass fraction) were carried out in order to provide wider comparisons of data for different fuel options. Table 2 shows the fuel feed rate, FR, as well as effective fuel moisture and rice husk energy fraction (or rice husk contribution to the total heat input by the blended fuel), EF rh, for different fuel options (in effect, for different rice husk mass fractions in the fuel blend) applied in this work. The effective moisture contents and rice husk energy fractions provided in Table 2 for distinct fuel options were determined using the fuel properties given in Table 1 and rice husk mass fraction in the blended fuel. 2.3 Experimental tests planning Despite apparent effects of the fuel mass fractions of rice husk and sugar cane bagasse on the feed rate of the blended fuel (for particular rpm of the screw feeder), it was managed to 233

3 maintain the fuel feed rate at (near) the same value in all of the test runs (see Table 2). The rice husk energy fraction EF rh combining the effects of the fuel mass fractions and LHV of individual fuels was, therefore, selected to be one of the main independent variables (together with EA) in this work. Table 1 Proximate and ultimate analyses of rice husk and bagasse co-fired in the conical FBC Analysis Rice husk Bagasse Ultimate analysis (wt.%, dry and ash-free basis): Carbon Hydrogen Oxygen Nitrogen Sulfur Proximate analysis (wt.%): Moisture ( as-received basis) Ash ( dry basis) LHV (MJ/kg) Table 2 Fuel feed rate, moisture content and rice husk energy fraction for the fuels used in the experimental tests Rice husk mass percentage in the fuel blend Fuel feed rate (kg/h) Moisture content (%) Rice husk energy fraction Experimental tests were carried out with the aim to investigate effects of the fuel characteristics (via rice husk energy fraction, EF rh ) and operating conditions (via percentage excess air, EA), on axial temperature and gas concentration profiles in the conical FBC as well as on the emission performance of the combustor. The fuel characteristics for different test runs corresponded to those given in Table 2; meanwhile, for each fuel option, the biomass fuel was burned at four different values of EA: 4, 6, 8 and 1%. Temperature ( o C) EFrh =.85 EFrh = Bed temperature ( o C) RESULTS AND DISCUSSION 3.1 Major combustion characteristics Fig. 2 depicts thermal characteristics of the conical FBC, such as the axial temperature profiles (Fig. 2a) and the bed temperatures (Fig. 2b), for various fuel options, or rice husk energy fractions. The bed temperatures were conventionally associated with the location of 1 m above the air distributor. Analysis of the thermal characteristics for different fuel options and operating conditions showed that the temperatures at all the location points in the conical FBC were almost independent of EA for the particular fuel feed rate; meanwhile, they were strongly affected by EF rh. In addition, the effects of the EF rh on the temperature characteristics were weakened for the higher contributions of rice husk to the heat input in its co-firing with sugar cane bagasse. All the above effects can be apparently seen in Fig. 2b for the bed temperatures. As an illustration, Fig. 2a depicts the axial temperature profiles in the conical FBC firing the fuels at quasi-identical values of the fuel feed rate (FR = kg/h) and percentage excess air (EA 6%) for different values of EF rh. With the reduction in EF rh, the temperatures at all the combustor locations were lowered because of influence of the fuel moisture affecting the fuel quality (LHV). Though for EF rh =.6 the bed temperatures were at a relatively low level (of o C), the fuel ignition and combustion were quite stable in this combustor fired with the blended fuel at the 45% rice husk mass fraction. As seen in Fig. 2a, in the bed region of the combustor (up to 1-m level above the air distributor), the temperature profiles were almost uniform. However, slight positive temperature gradients were observed in this region, and the peak temperatures were found at the level of about 1 m above the air distributor for all the test runs. Meanwhile, in the freeboard region (of 1 3-m height), the temperature profiles were characterized by noticeable negative gradients because of the reduced heat release in this region and heat transfer (heat loss) across the combustor walls. An interesting point related to the above data was that for the particular location point in both bed and freeboard regions, the temperature gradients were identical for different values of EF rh. Effects of the rice husk energy fraction and excess air on the oxygen consumption rate along the conical FBC are shown in Fig. 3. In Fig. 3a, the axial O 2 profiles demonstrate the effects EA = 4% EA = 6% EA = 8% EA = 1% Rice husk energy fraction Fig. 2 Effects of the rice husk energy fraction on the axial temperature profiles for EA = 6% as well as on the bed temperatures for various EA in the conical FBC for the biomass (co)-firing at the fuel feed rate of kg/h. 234

4 O2 concentration (vol.%) EFrh =.85 EFrh =.6 O2 concentration (vol.%) EA = 4% EA = 6% EA = 8% EA = 1% Fig. 3 Effects of the rice husk energy fraction and excess air on the axial O 2 concentration profiles in the conical FBC for biomass (co)-firing at the fuel feed rate of kg/h: EA = 6% for and EF rh =.6 for. of EF rh on the current O 2 concentrations (along the combustor height) obtained for the same operating conditions as in Fig. 2a. Unlike temperatures, the O 2 concentrations at various location points in the conical FBC were found to be almost independent of the fuel analysis. Accordingly, the axial O 2 profiles for various blended options were similar to the profile found for the case of firing rice husk only. With the relative error of about 15%, all the dependencies in Fig. 3a could be approximated by a single curve. Meanwhile, for the fixed EF rh, the effects of EA on the axial O 2 concentration profiles were, as expected, quite significant. Figure 3b illustrates these effects for the case of co-firing rice husk and bagasse at EF rh =.6 for different values of EA. As seen in Fig. 3, the maximum rates of the oxygen consumption occurred in the bed region indicating the highest rates of the combustion process in the conical part of the combustor for various operating conditions and fuel options. Small gradients of the axial O 2 concentration profiles at the combustor top indicated the occurrence of the combustion process in this region. Moreover, reduced O 2 concentrations in the flue gas at the cyclone exit (compared with those at the reactor top) confirmed the fact of oxidation of carbonaceous compounds along the gas path combustor top cyclone exit. 3.2 CO formation and reduction in the conical FBC Fig. 4 shows the effects of EF rh and EA on the CO emission characteristics of the conical FBC. The axial CO concentration profiles represented in Fig. 4a were obtained for EA = 6% when (co-)firing the fuels at different EF rh, whereas the CO emissions from the conical FBC (or CO concentrations in the flue gas at the cyclone exit) depicted in Fig. 4b were obtained in the tests of co-firing the fuels at EF rh =.6 and different EA. For all the fuel options and EA, the CO concentration profiles were found to have a maximum, CO max, whose location (above the air distributor) divided conventionally the combustor volume into formation (lower) and reduction (upper) regions, as may be seen in Fig. 4a. Actually, the chemical reactions responsible for formation and decomposition of the most compounds, involved in the combustion process, proceed in both regions. However, in this work, the terms formation region and reduction region are applied to the conical FBC with the meaning of the net result in each of the above regions. In the test run with the maximum mass fraction of bagasse in the fuel blend (or when EF rh =.6), the CO max values were significantly greater than those for EF rh =.85 and when firing rice husk only. CO concentration (vol.%) EFrh =.85 EFrh =.6 CO emission (vol.%) EA = 4% EA = 6% EA = 8% EA = 1% Rice husk energy fraction Fig. 4 The CO emission characteristics of the conical FBC operated at the fuel feed rate of kg/h: axial CO concentration profiles for EA = 6% and various EF rh, and CO emissions from the combustor for EF rh =.6 and various EA. 235

5 Two factors were likely responsible for this effect: (1) lower temperatures (see Fig. 2a) leading to the increased CO/CO 2 ratio in carbon oxidation and (2) high concentrations of water vapor enhancing the contribution of carbon wet oxidation (basically, to CO) occurring on the surface of char particles, i.e. in heterogeneous reactions [2]. As seen in Fig. 4a, a significant reduction in the CO concentration took place in the freeboard region along the combustor height. In this region, CO was likely oxidized in homogeneous reactions with OH radicals and oxygen, both being predominant in the freeboard region [9]. Because of elevated concentrations of water vapor (and, accordingly, OH radicals) in the flue gas, the highest rate of CO reduction was achieved in the test runs at EF rh =.6. Analysis of the data on CO concentrations for different fuel options and operating conditions indicated the significant reduction in the CO concentration along the flue gas path from the combustor exit to the cyclone exit. Thus, for the case of EF rh =.6 and EA = 6%, the CO concentration was reduced from about.8% at the combustor top to.5% at the cyclone outlet, mainly owing to the above reduction reactions. A small contribution to the CO reduction seemed to be done by the homogeneous reaction of CO with NO [2]. As follows from data in Fig. 4b (demonstrating the combined effects of EF rh and EA on the CO emission from the conical FBC), the effluent of this pollutant from the combustor could be effectively controlled by air supply via maintaining excess air. For the particular fuel option, the CO emission was apparently reduced with the increase in EA (varied within a reasonable range); however, the effects of excess air on the CO emission characteristics weakened for greater values of EA. Meanwhile, as could be generally concluded, with the increase of the bagasse mass/energy fraction in the blended fuels, the CO emissions from the conical FBC were increased when excess air was maintained at the same value. 3.3 NO formation and reduction in the conical FBC In this work, NO x emissions were represented by NO only because no NO 2 was detected in the test runs. Fig. 5 shows the effects of EF rh and EA on the NO emission characteristics of the conical FBC: the axial NO concentration profiles (Fig. 4a) and NO emissions (Fig. 4b) for the same fuel options and operating conditions as in Fig. 4. Like for CO, all the axial NO concentration profiles possessed a maximum, NO max, whose location made it possible to distinguish conventionally the formation region and the reduction region for this pollutant. Because of the low temperature level in the conical FBC (as seen in Fig. 2), the NO was expected to form in the biomass fuel combustion predominantly owing to the fuel-no formation mechanism. Basically, fuel-no can be formed in the combustion process through oxidation of nitrogen species, such as HCN and NH 3, released from the fuel particles with the volatile matter (mostly, in the bottom region) as well as through oxidation of fuel nitrogen retained in the char. In the freeboard region, NO reduction may likely occur in reactions with NH 3 and also with carbon and CO on the char surface [1]. As seen in Fig. 5a, for the particular EA, the axial NO concentration profiles were significantly influenced by EF rh. At all location points in the conical FBC, the NO concentrations were found to be greater for higher EF rh because of elevated fuel-nitrogen and higher temperatures. In the bottom region of the combustor, the rates of NO formation reactions (HCN and NH 3 oxidation) were obviously greater than those of the NO decomposition which resulted in the NO concentration increase along the bed height up to the NO max value. However, in the freeboard region, the axial NO concentration profiles were found to decline because of the prevailing of the above NO reduction reactions. Analysis of the dependencies in Fig. 5b showed with the increase of the bagasse mass/energy fraction in the blended fuels, the NO emissions from the conical FBC were mitigated when excess air was maintained at the same value. Meanwhile, for the particular fuel option, with the increase in EA, the NO emissions from the conical FBC were increased. Together with effects of the temperature, these facts confirmed the fuel-no formation mechanism occurred in this conical FBC during combustion of rice husk as well as of the fuel blends. 3.4 Effects of co-firing on combustion efficiency Though the determining of the combustion efficiency was not among the objectives of this work, some remarks could be made on influence of the rice husk and sugar cane co-firing on the major heat losses and combustion efficiency. As shown in Fig. 4a, the co-firing of rice husk with sugar cane bagasse led to the increased CO emissions in comparison with those for the rice husk combustion. Taking into account reduced LHV for blended fuels, it could be concluded that for the cases of the co-firing, the heat losses owing to incomplete combustion were always greater than those for the rice husk combustion. 2 3 NO concentration (ppm) 2 1 EFrh =.85 EFrh = NO emission (ppm) EA = 4% EA = 6% EA = 8% EA = 1% Rice husk energy fraction Fig. 5 The NO emission characteristics of the conical FBC operated at the fuel feed rate of kg/h: axial NO concentration profiles for EA = 6% and various EF rh, and NO emissions from the combustor for EF rh =.6 and various EA. 236

6 Assuming the heat loss owing to incomplete combustion to be.5% for the rice husk combustion [2], this heat loss could be roughly estimated as high as 3 4% for the case of co-firing rice husk and bagasse at EA of about 6% and EF rh of about.6. Because of the lower ash content in the blended fuels, the heat losses owing to unburned carbon for the cases of the co-firing could be predicted at lower values compared with that for the case of firing rice husk only. Depending on the design features, fuel properties and operating conditions, the reduction in this heat loss could be expected in a very wide range of 1 6%. The higher effects could be achieved when using high-ash rice husk as a fuel in relatively short fluidized-bed combustors. Thus, the combustion efficiency for the case of the rice husk and bagasse co-firing would be improved or deteriorated (with respect to the rice husk combustion) depending on the changes in the above heat losses. 4. CONCLUSIONS A conical fluidized-bed combustor (FBC) was successfully tested when co-firing kg/h of blended as-received rice husk and sugar cane bagasse for different values of the mass/energy fraction of rice husk in the blended fuel and excess air. However, attempts to burn raw sugar cane bagasse as well as the fuel blends at relatively low values of the rice husk mass fraction failed in preliminary tests of this fluidized bed system. Through co-firing with sugar cane bagasse, the combustion efficiency and emission performance of the combustor could be improved compared with those for the case of firing high-ash rice husk in relatively short reactors. Meanwhile, through the co-combustion with rice husk, an effective use of as-received sugar cane bagasse becomes feasible for energy production in fluidized-bed combustion systems. Some particular conclusions could be made on the thermal and emission characteristics of the conical FBC co-fired with rice husk and bagasse: the axial temperature profiles as well as the bed temperatures in the conical FBC were strongly affected by the rice husk energy fraction, EF rh ; however, the bed temperatures were almost independent of excess air, EA; the axial O 2 concentration profiles were almost independent of EF rh (or blended fuel analysis); with the increasing of EF rh in the blended fuel, the CO emissions from the conical FBC were reduced for quasi-identical values of EA; with the increasing of the bagasse mass/energy fraction in the blended fuel, the NO emissions from the conical FBC were mitigated when EA was maintained at the same value. Applied Energy, 16, pp [4] Armesto, L., Bahillo, A., Veijonen, K., Cabanillas, A. and Otero, J. (22) Combustion behaviour of rice husk in bubbling fluidised bed, Biomass & Bioenergy, 23, pp [5] Fang, M., Yang, L., Chen, G., Shi, Z., Luo, Z. and Cen, K. (24) Experimental study on rice husk combustion in a circulating fluidized bed, Fuel Processing Technology, 85, pp [6] Natarajan, E., Nordin, A. and Rao, A.N. (1998) Overview of combustion and gasification of rice husk in fluidized bed reactors, Biomass and Bioenergy, 14, pp [7] Basu, P., Cen, K.F. and Jestin, L. (2) Boilers and Burners, Springer, New York. [8] Sami, M., Annamalai, K. and Wooldridge, M. (21) Co-firing of coal and biomass fuel blends, Progress in Energy and Combustion Science, 27, pp [9] Tillman, D.A. (2) Biomass co-firing: the technology, the experience, the combustion consequences, Biomass & Bioenergy, 19, pp [1] Winter, F., Wartha, C. and Hofbeuer, H. (1999) NO and N 2 O formation during the combustion of wood, straw, malt waste and peat, Bioresource Technology, 7, pp REFERENCES [1] NEPO (2) Thailand biomass-based power generation and cogeneration within small rural industries, Final Report on Research Conducted by Black & Veatch, National Energy Policy Office, Thailand. [2] Permchart, W. and Kouprianov, V.I. (24) Emission performance and combustion efficiency of a conical fluidized-bed combustor firing various biomass fuels, Bioresource Technology, 92, pp [3] Bhattacharya, S.C., Narendra, S. and Alikhani, Z. (1984) Some aspects of fluidized bed combustion of paddy husk, 237