Brickmakers in the U.S.-Mexico border region use low quality fuels such as agricultural waste, sawdust, scrap wood, waste plastics, and scrap tires to fire their bricks. Pollutants produced by the firing process include particulate, CO, hydrocarbons, and NOx. In an effort to reduce these emissions, Mexican and U.S. officials have been advocating the use of high quality fuels like natural gas and propane. The high costs of propane and natural gas make it desirable to supplement these clean-burning fuels with low-cost waste fuels. In this way, the environmental benefits of propane or natural gas can be combined with the economic benefits of a waste fuel. The University of Utah constructed a pilot-scale brick kiln to test different co-firing strategies and to quantify the gaseous and particulate emissions from three waste fuels: sawdust, pecan shells, and scrap wood. The inside dimensions of the facility were 3-by-5-by-12 feet and the firing rate was 20,000 Btu/h•ft2.
We tried three approaches to reduce emissions. The first involved the use of a grate to hold the solid fuel above the floor of the kiln. The hypothesis was that the grate would improve the circulation of air through the burning bed and reduce the rates of formation of soot, CO, and unburned hydrocarbons. The second approach involved firing natural gas above the burning bed. The hope was that the natural gas flames would help destroy soot, CO, and unburned hydrocarbons that were produced in the bed. The third approach involved simply using natural gas to preheat the kiln, so that when the gas was shut off and the solid fuels were fed, they would burn more cleanly and efficiently in a preheated kiln.
The use of a grate was not effective with sawdust because the dust simply fell through the grate and burned on the floor of the kiln. With the larger pecan shells the grate reduced emissions of particulate and CO by 15 and 50 percent. Firing a small amount of natural gas (25 percent of the total heat input) above a burning bed of sawdust caused a 20 percent reduction in particulate emissions but had little effect on emissions of CO.
Preheating the kiln with natural gas proved to be the most effective way of reducing emissions of gaseous and particulate emissions. Preheating the kiln to 1200°F reduced emissions of particulate and CO by 80 percent relative to emissions obtained on starting up a kiln which was initially at room temperature.
When the kiln's temperature was above 1000 °F (538 °C), further increases in temperature had a large effect on the size distribution and composition of the particles formed. At 1000 °F, sawdust produced particles that were mostly soot (54.5 percent carbon). At this temperature, only 32 percent of the particles (by mass) had an aerodynamic diameter less than 0.4 µm. At 1400 °F (760 °C), the particles were mostly ash (0.07 percent carbon), and 80 percent of the particles had an aerodynamic diameter less than 0.4 µm.
At all of the temperatures examined, more that 50 percent (by mass) of the particulate that is emitted is less that 2.5 µm in diameter. Increasing the kiln's temperature to 1400°F increases the amount of fine particulate to about 80 percent but it greatly reduces the total amount of particulate. Emissions of total suspended particulate while burning sawdust were 350 mg/SCM at startup conditions and 100 mg/SCM at 1400°F.
Based on these experimental findings, we conclude that the most cost effective way of using natural gas or propane to reduce gaseous and particulate emissions is to preheat the kilns to the greatest extent possible with gas before switching to more economical fuels. In addition, for waste fuels with larger sizes, the use of a grate to increase fuel-air contacting is also effective.
Future work should attempt to demonstrate these results on field -scale kilns. In particular, future work should establish a clear understanding of the trade-offs between reduced emissions and the cost of the natural gas or propane, based on field-scale data. In the absence of field-scale data, or to supplement field-scale data, the brick firing process could be mathematically modeled. The model could then be used to establish heating times and the economics of the preheating process.
This project has provided, for the first time, detailed information on particulate emissions from the firing of waste fuels under realistic kiln conditions. Approaches for reducing particulate and gaseous emissions that involve modification of the combustion process through the use of grates and preheating of the kilns have been identified and demonstrated on a pilot-scale facility. The data provide a basis for establishing emission factors for CO, particulate, and NOx over a range of operating conditions with three common waste fuels.
Introduction
Environmental pollution of the U.S.-Mexico border region is a growing problem. The border region, defined as the area extending 100 km north and south of the international border, has long been a free-trade zone attracting many U.S.-based manufacturing plants. As Mexico's economy has worsened, the population of the region has exploded as Mexican citizens have moved to the border cities to find work. The increased population of the area has overwhelmed the existing infrastructure and compounded the existing environmental problems.
The air pollution problem is most severe in the major metropolitan areas of the border region, such as El Paso-Ciudad Juarez, where frequent temperature inversions create meteorological conditions that fail to clear the atmosphere of pollutants (Affairs, p. 3319). For several years, El Paso has violated the EPA standards for the concentrations of ozone, carbon monoxide, and total suspended particulates (TSP) (Warner, 1991). According to a study by Applegate et al., two major sources of particulate pollution in El Paso-Juarez are unpaved roads and open burning (Appelgate, 1989).
A major source of carbon monoxide and particulate emissions are the area's 300 brick kilns. The brick makers use centuries-old technology and cheap waste fuels such as sawdust, pecan shells, or scrap tires to fire their bricks. In developing countries such as Mexico, the brick-making process has changed little for thousands of years. The process begins by mixing clay with an additive such as sand, sawdust, or manure. Workers often use their feet to mix a batch of materials until it has reached the proper consistency and is free of lumps. The mixture is then covered with plastic and allowed to sit overnight before it is molded into bricks. The bricks are molded by hand using open molds. The wet bricks are then placed on the ground and the excess water is removed by the sun. They are rotated occasionally so that they can dry evenly on all sides.
Once the excess water has been removed, the bricks are loaded into the kiln to be cured. In the small brick-making operations, the kilns are usually made from adobe brick and have a capacity of 10,000 - 50,000 bricks. The bricks are fired for a period ranging from 16 hours to 7 or 8 days, depending on the type of brick being made.
At the beginning of the firing process, the top of the kiln is left uncovered to allow the water still in the bricks to escape. After all the water has evaporated, the top of the kiln is covered with a layer of wet bricks or other covering to permit the kiln to heat up to the proper curing temperature. At the end of the firing cycle, the kiln is allowed to cool and the bricks are removed.
A local research institute, the Instituto de Investigaciones Ecotechnologicas (ECOTEC) is seeking to improve the air quality in the border area by promoting the use of clean-burning fuels, such as natural gas and propane. However, the high (and rising) cost of these fuels is making their use unattractive to the brick makers.
The high costs of propane and natural gas make it desirable to supplement these fuels with low-cost waste fuels such as sawdust, pecan shells, or scrap wood. In this way, the environmental benefits of a gas fuel can be combined with the economic benefits of a waste fuel. A pilot-scale brick kiln was constructed at the University of Utah to quantify the emissions (especially the particulate emissions) from burning the three waste fuels listed above. Several different firing strategies combining waste fuels and natural gas were then tested to determine their effect on reducing the emissions. The results of these tests are discussed in this paper.
The Pilot Kiln
A pilot-scale brick kiln was constructed at the University Combustion Research Center, located at 870 S. 500 W. in Salt Lake City, Utah. The kiln is about the same height as a field kiln, about 12 feet (3.6 m), though it has only about one fourth the volume. The internal dimension of the floor is 3 ft. X 5 ft. (0.9 X 1.5 m). The kiln was designed for a firing rate of up to 200,000 Btu/hr-ft2 (630,000 W/m2), but during the experiments described below, it was operated at 20,000 Btu/hr-ft2 (63,000 W/m2). This section contains a description of the pilot-scale kiln, divided into four subsections: the kiln shell, kiln insulation, burners and flame monitoring system, and the sampling duct.
The Shell
The shell of the kiln was fabricated from 3/8" carbon steel by Continental Steel of Magna, Utah. A diagram of the kiln, illustrating the dimensions and locations of all its ports, is shown in Figure 1. The "front" view in the figure represents the east and west walls of the kiln, while the "right" view shows the north and south walls.
There are two large, rectangular 30 3/4" X 36" (78 X 91 cm) ports located near the base of the kiln in the east and west walls. The large west-wall port was closed with refractory brick, insulating board, and a port cover during the kiln operations. The size of the east-side opening was reduced to about 15" X 12" (38 X 30 cm) by stacking Greenlite 30 refractory bricks in the port. It was through this opening that the solid waste fuels were fed by hand into the kiln.
In the north and south walls of the kiln are six 3" (7.6 cm) circular ports. These ports are situated in two horizontal rows of three, directly across from one another, 18" (46 cm) above the kiln floor and spaced 12" (30 cm) apart. The kiln's two burners were mounted in two of these ports, one in each wall. In the east and west walls of the kiln, there are ten more of these 3" (7.6 cm) circular ports. These are situated vertically, in two rows of five, beginning about four feet above the floor and spaced 15" (38 cm) apart. All of the circular ports were closed with refractory brick and port covers during kiln operation, except for the two in which burners were mounted.
In addition to the two large ports near the base of the east and west walls described above, the kiln has fourteen other rectangular ports, seven each in the north and south walls. Each of these rectangular ports is identical to the one directly across from it. The first port, moving from the base of the kiln toward the top, is a 12" X 48" (30 X 130 cm) port located 7 1/2" (19 cm) above the kiln floor and one foot (0.3 m) below the horizontal row of circular ports. The next one is a 3" X 48" (7.6 X 122 cm) port located 27" (69 cm) above the floor, 7 1/2" (19 cm) above the horizontal row of circular ports. Starting two feet above the second port is a row of five 3" X 36" (7.6 X 91 cm) rectangular ports, spaced 15 - 18" (38 - 46 cm) apart. All of the rectangular ports in the north and south walls were closed with refractory brick and port covers during kiln operation.
During runs in which start up conditions were simulated, two cooling coils were installed in the north and south walls of the kiln, in the second pair of 3" X 36" (7.6 X 91 cm) ports. The coils were located about 15" (38 cm) above the refractory planks. One coil was placed in each of the north and south walls. The coils were made from 3/8" stainless steel tubing and each extended to the center of the kiln. Water was run through the coils during startup runs to simulate the quenching effect of a charge of wet bricks. For runs in which the kiln was preheated to a temperature of 1000 °F (540 °C) or more, the coils were removed.
Three bare 1/8", K-type thermocouples were used to estimate the temperature inside the kiln. The first was located in the south wall of the kiln, 18" (46 cm) above the floor, in the left circular port (facing the kiln). The second was located in the west wall of the kiln, in the circular port located about four feet (1.2 m) above the floor. The third thermocouple was also located in the west wall, 30" (76 cm) above the second. Each of these thermocouples were 20" (51 cm) long and extended about 6" (15 cm) inside the kiln. The average of the three temperatures from these thermocouples was used to estimate the temperature inside the kiln.
The Kiln Insulation
The outer shell of the brick kiln is fabricated from 3/8" carbon steel. This outer wall is protected by two layers of insulating board and two layers of castable refractory. Figure 2 illustrates a cross-section of the kiln and shows the location and thickness of each of these four layers. All of the refractory and insulating board were purchased from A.P. Green of Mexico, Missouri.
Adjacent to the steel wall is a 3" (7.6 cm) layer of Insblock 1900. This insulation has a conductivity of 0.71 Btu-in/ft2-hr-°F (10.2 W-cm/m2-°C) and a temperature rating of 1900 °F (1040 °C). Next is a 2" (5 cm) layer of Insboard 2600. This insulation has a conductivity of approximately 0.80 Btu-in/ft2-hr-°F (11.5 W-cm/m2-°C) and a temperature rating of 2600 °F (1430°C). Adjacent to the Insboard is a 4" (10 cm) layer of Kast-o-Lite 30. This refractory has a conductivity of 3.8 Btu-in/ft2-hr-°F (54.8 W-cm/m2-°C) and a temperature rating of 3000 °F (1650°C). The inner layer is 4" (10 cm) of Ultra-Express 70. This refractory has a conductivity of 11.9 Btu-in/ft2-hr °F (171.5 W-cm/m2-°C) and a temperature rating of 3100 °F (1700°C). Even though it offers less resistance to heat transfer than Kast-o-Lite, Ultra-Express was chosen for the inner layer because of its high resistance to abrasion.
There are two natural gas burners mounted in the north and south walls of the pilot kiln, about 18" (46 cm) above the kiln floor. Both are venturi-type, inspirated burners manufactured by Buzzer (model number VNB150-HP), and purchased from Charles A. Hones, Inc., of North Amityville, New York. One of the burners is mounted in the 3" (7.6 cm) circular port located in the center of the south wall of the kiln. This burner was used both for preheating the kiln and for firing natural gas above the bed during the co-firing runs. The second burner was mounted on the north wall of the kiln in the circular port to the left (facing the kiln). This burner was used only for preheating the kiln or for maintaining the kiln temperature between runs. The gas flow to each burner was controlled by controlling the gas pressure at the burner inlet. This was done manually with a needle valve and pressure gauge located just upstream of each burner.
In the two circular ports directly opposite each burner, two UV sensors (Honeywell model #C-7027) were mounted to monitor the burner flames. When a burner's flame is extinguished, the UV sensor sends a signal to a solenoid valve that shuts off the gas supply to the burner.
Sampling Duct
At the top of the kiln, the 18" (46 cm) exhaust flange is reduced to 6" (15 cm) black stove pipe. The stove pipe extends for twenty inches (51 cm) above the kiln, then bends north and runs for ten feet (3 m) before tying in with the main building exhaust duct. Figure 3 shows the top and side views of the sampling duct. A "tee" near the tie-in with the main exhaust system allows dilutant air from the room to mix with the exhaust gasses, cooling them before they enter the main system. The draft through the kiln is controlled by a damper located 6" (15 cm) upstream of the tee.
The sampling port is located 51" (130 cm), or 8½ pipe diameters, downstream from the elbow and 68" (173 cm), or 11½ diameters, upstream from the damper. An air-cooled probe inserted into this port was used to sample both particulate and combustion gasses.
The pitot tube port is located 48" (122 cm), or 8 diameters, downstream from the sampling port. A pitot tube inserted into the center of the gas stream was used to measure the difference between the static and dynamic pressures of the flowing exhaust gas. This value was used, along with the temperature measurement at the same location, to calculate Vmax, the gas velocity at the midpoint of the pipe. Vmax was then used to estimate the average gas velocity (and flow rate) in the duct, using the empirical correlation in Figure 5-7 of McCabe and Smith (1993). This correlation shows the ratio of Vavg to Vmax as a function of the Reynolds number. A value of the Reynolds number is assumed. Then the correlation and the measured value of Vmax are used to calculate Vavg. This value is used to calculate a new Reynolds number and the process is repeated until the value of Vavg converges. For the pilot brick kiln, operating at a firing rate of 20,000 Btu/hr-ft2 (63,000 W/m2), the calculated Reynolds number in the sampling duct was 1.12 X 104, which lies in the turbulent flow region.
Two different methods for sampling particulates were used, depending upon the type of data desired. The first method described below was used to determine the total suspended particulate (TSP) in the kiln exhaust, reported as mg per dry standard cubic meter of gas. Standard conditions in this paper refer to conditions at one atmosphere and 68 °F (20 °C). The TSP sampling method was also used to analyze the combustion gas for O2, CO2, CO, NOx, and total hydrocarbons (THC). In the second sampling method described below, a cascade impactor was used to determine the aerodynamic particle size distribution of the particulate material.
Sampling for Total Suspended Particulate and Combustion Gasses
EPA document 40 CFR, Pt. 60, App. A, Methods 5 and 1A describe the procedure for TSP sampling in a small duct. According to this document, the sampling location should be at least eight diameters downstream and ten diameters upstream from the nearest flow disturbances. The velocity measurement site should be at least eight diameters downstream from the sampling site. The preceding discussion of the pilot kiln's sampling duct shows that it complies with these requirements.
Figure 4 is a diagram of the sampling train used to measure TSP emissions. The emissions were sampled from the from the center of the 6" (15 cm) sampling duct using a 3/8" stainless-steel, air-cooled probe. The sample gas was cooled inside the probe to about 300 °F (150 °C) before being passed through a 1-µm glass-fiber filter to remove the particulates. The water in the gas was then removed by passing the sample through a refrigerated knock-out pot. This was the only modification of the EPA method, which suggests using a series of impingers in an ice-cooled bath and a silica gel to dry the gas. A dry gas meter was used to measure the volume of the sample gas before it was analyzed for O2, CO2, CO, NOx, and total hydrocarbons (THC), using continuous, on-line gas analyzers. The particulate mass collected on the filter was divided by the measured sample volume corrected to standard conditions (1 atm., 20 °C) to determine the total suspended particulates (TSP) in milligrams per dry standard cubic meter.
To make sure the samples were taken isokinetically, a pitot tube was used to measure the gas velocity at the center of the duct. The sampling probe was also inserted to the center of the duct, 8 diameters upstream from the pitot tube. The measured gas velocity was multiplied by the cross-sectional area of the entrance to the probe to get the volumetric flow rate at the probe entrance. This flow rate was then corrected to the temperature and pressure at the rotameter to determine the desired flow rate at the rotameter. The needle valve on the rotameter was used to maintain the flow at this value for the duration of the run. The target sampling rate during the collection of TSP and gas composition data was between 12 and 16 standard cubic feet (0.40 - 0.45 SCM) per hour.
Sampling for Particle Size Distribution
The sampling train used to obtain aerodynamic particle size distribution data is shown in Figure 5. This train is similar to the one used to obtain TSP data, with a few modifications. The 1-µm particulate filter was replaced with an Anderson cascade impactor. The impactor was enclosed in a heated, insulated box to maintain a nearly constant temperature inside the impactor of about 250 °F (120 °C). This prevented water vapor in the gas stream from condensing on the impactor plates. It also reduced variations in temperature and volumetric flow from stage to stage within the impactor.
The impactor was calibrated for a flow rate of one actual cubic foot per minute. To maintain this flow rate, dilutant nitrogen was added to the sample gas just upstream from the impactor. During experimental runs, dry gas meter readings were taken every five minutes and the measured total volumetric flow rate was corrected to the temperature and pressure at the impactor to verify a flow of 1 acfm.
The aerodynamic particle size distribution data were reported assuming ideal impactor behavior (collection efficiency of 50%). This is equivalent to assuming that, at each stage, the mass of particles larger than the cutoff size that get past the stage is equal to the mass of particles smaller than the cutoff size that are collected (Hinds, 1982).
No gas analyses were performed during the impactor runs. After the gas passed through the sampling train, it was sent to the main building exhaust system.
In Mexico, brick makers often use waste fuels such as tires, plastics, scrap wood, sawdust, or pecan shells to fire their bricks. The firing process consists of two stages, a drying stage and a curing stage. During the drying stage, the top of the kiln is left uncovered to allow the moisture in the bricks to escape. After most of the water has evaporated, the top of the kiln is covered with a layer of wet bricks or other covering to permit the kiln to heat up to the proper curing temperature. After the bricks have cured for an appropriate length of time, the kiln is allowed to cool and the bricks are removed.
The first purpose of the project was to characterize the emissions from the brick kilns as they are currently operated. The emissions were measured during the two stages of the firing cycle. The first condition studied during these tests was the curing stage of the firing cycle when the kiln is at a high temperature. The kiln was preheated with natural gas to a temperature of about 1200 °F (650 °C). The gas was then turned off and solid fuels were added to maintain the kiln temperature. The second condition examined was the drying, or start up stage of the firing cycle. For these tests, a wood fire was started in a cold kiln and sawdust was fed on top of the burning wood. The cooling coils were installed and water was run through the coils at a rate of 1.2 gallons per minute. This provided a heat removal rate of about 5700 Btu/hr (1670 W), simulating the cooling effect of a load of wet bricks in the kiln.
During most of the experiments described in this and the following section, the kiln was operated at a constant firing rate of 20,000 Btu/hr-ft2(63,000 W/m2). This rate was estimated from fuel usage data from a field kiln in Juarez. The draft through the pilot-scale kiln was controlled using a damper in the 6"-exhaust duct and was maintained at about 140 standard cubic meters (SCM) per hour. At this draft rate, the pressure drop measured with a pitot tube at the center of the duct was 0.05 inches of water. With the draft at this level, and with the kiln operating at the above firing rate, the concentration of oxygen in the exhaust gas was 12-15 volume percent. This value agrees with the oxygen concentration data from a field kiln in Juarez. The conditions in the pilot kiln were similar to those in a field kiln.
Emissions from a Hot Kiln
For these tests, the kiln was preheated to a temperature of about 1200 °F (650 °C) with natural gas. The gas was then turned off and three waste fuels commonly used in the Mexican kilns were burned. These fuels included clean sawdust, pecan shells, and small pieces of wood. The ultimate analyses and heating values of the fuels, on an as received basis, are show below in Table 1.
|
|
|
|
| Moisture |
|
|
| Carbon |
|
|
| Hydrogen |
|
|
| Nitrogen |
|
|
| Sulfur |
|
|
| Ash |
|
|
| Oxygen |
|
|
| Btu/lb (kJ/kg) |
|
|
The pecan shells were obtained from Navarro Pecan in San Saba, New Mexico. The shells ranged in size from particles the size of course sand to pieces approximately " X ½" (2.2 X 1.3 cm).
The solid fuels were fed into the kiln by hand in a manner similar to that observed in Juarez. A 20-oz tin can on the end of a short handle was used to feed the sawdust and pecan shells. One can of sawdust was thrown onto the kiln floor every 15 seconds. This resulted in a mass feed rate of about 250-300 grams per minute, depending on the density of the sawdust. Pecan shells were fed at a rate of 1½ cans per minute. This corresponds to a mass feed rate of about 300 grams per minute. The wood pieces were fed into the kiln by hand at the rate of 1 kg every five minutes.
Due to their smaller size, the sawdust and pecan shells burned at a rate equal to or greater than the rate at which they were fed. The wood pieces, however, burned at a lower rate than their feed rate, resulting in the accumulation of coals on the kiln floor. The firing rate during the runs in which wood pieces were burned, therefore, was somewhat less than the target rate of 20,000 Btu/hr-ft2.
The sampling of total suspended particulate, CO, and NOx in the kiln exhaust was performed using the method for TSP and combustion gas sampling described above. The results from the three different fuels are compared below in Figures 6 and 7.
At the baseline temperature of 1200 °F (650 °C), the pecan shells produced somewhat higher particulate emissions than either the sawdust or the wood pieces. The wood pieces appear to produce significantly less particulate than either of the other two fuels. This is partly due to their lower firing rate in the kiln and to their larger size.
The NOx emissions from the pecan shells were nearly double those from the other fuels, due to their higher nitrogen content (see Table 1). The sawdust produced significantly more CO than the wood pieces and pecan shells. This may be due to the fine size of the sawdust, which led to quick bursts of combustion, and possible locally fuel-rich conditions. Once formed, the CO is difficult to burn at the relatively low temperatures inside the kiln because of unfavorable kinetics. The pecan shells and wood pieces, which formed much less CO, produced steady flames. At 1200 °F (650 °C), the hydrocarbon emissions from all of the fuels were negligible (less than 20 ppm).
Particle size distribution data at 1200 °F were also taken while burning sawdust and pecan shells. Figure 8 below compares the aerodynamic particle size distribution from burning sawdust to that from burning pecan shells.
The size distributions are very similar, indicating that the size of the emitted particles is not a strong function of fuel type. The sawdust emitted a slightly higher percentage of particulate in the larger size ranges (> 9 µm) and a slightly lower percentage of particulate in the smaller size ranges (< 0.6 µm). This was the case at the other kiln temperatures studied as well. This may be due to the difference in the particle sizes of the fuel. Small particles of the sawdust may be caught in the draft air before they have a chance to completely burn. This could result in a number of relatively large, incompletely-burned sawdust particles reaching the sampling duct. The larger, heavier pecan shell pieces are not as likely to become entrained, and are likely to remain on the kiln floor until they are completely burned.
Emissions During Kiln Start Up
To measure the kiln emissions during start up, wood and sawdust were burned with the kiln initially at room temperature. A fire was started in the cold kiln with about 2.5 kg of wood before data collection began. As emissions data were being taken, sawdust was fed into the kiln on top of the burning wood at a rate of 250 - 300 grams per minute. Every ten minutes, about 2.5 kg of wood were thrown into the kiln to build up the fire. Since neither the wood nor the sawdust burned as fast as it was fed into the kiln, it was not possible to know the firing rate during these start up runs; however, the conditions of low kiln temperature coupled with an excess of fuel accurately represent conditions in a field kiln during start up.
To simulate the cooling effect of a charge of wet bricks during start up, the two cooling coils were installed 15" above the radiation barrier. Water was run through each coil at a rate of 0.6 gallons per minute. The measured temperature increase in each coil was about 9.5 °F (5.3 °C). Together, the coils resulted in a total heat removal rate of 5700 Btu/hr (1670 W).
Figure 9 below shows the concentrations of CO, THC, and O2 as a function of time after kiln start up. The concentrations of CO for the first eight minutes exceeded the upper limit of the analyzer (2000 ppm).
During the first few minutes after kiln start up, when excess fuel is present, the oxygen concentration reaches a minimum of 8%. At the same time, the hydrocarbon (THC) and CO concentrations reach maximums of 800 and 2000+ ppm, respectively. As the temperature inside the kiln increases and the excess fuel is burned, the oxygen concentration gradually climbs to 14%, while the THC and CO concentrations decrease to their steady values of 250 and 1500 ppm. These emissions continue to slowly decrease as the kiln temperature increases, as shown in below in Figure 15. At 1000 °F, the measured THC concentration is about 20 ppm, and the CO concentration decreases to about 1100 ppm.
The results of these runs confirm that the highest particulate, THC, and CO emissions occur during start up, when the temperature inside the kiln is less than about 600 °F (315 °C). During this stage of operation, the measured TSP concentration was about 350 mg/SCM, compared with 60 mg/SCM at the baseline temperature of 1200 °F (650 °C). The CO concentrations averaged about 1800 ppm during start up (about 4 times higher than those at the baseline temperature), due to the low kiln temperature and an excess of fuel. THC concentrations ranged between 200 and 800 ppm, compared with less than 20 ppm at 1200 °F.
In this section, the effects of three alternative, low-cost firing strategies on kiln emissions are examined. The first strategy was to use a grate to hold the solid fuel above the kiln floor to improve the combustion by increasing the fuel-air mixing. The second strategy was to fire a small amount of natural gas above the bed of burning waste fuel. It was hoped that the natural gas flame would help complete the combustion of the particulates and CO given off by the solid fuel. The third strategy examined was to preheat the kiln to several different temperatures using natural gas before introducing the waste fuels into the kiln to determine the effect of kiln temperature on the emissions. The results from these three firing strategies are discussed below.
Fuel Grate
For these tests, a grate was used to hold the solid fuel about 5" (13 cm) above the kiln floor. The grate consisted of a stainless steel sheet, approximately 30" X 50", with 5/64"-diameter, staggered holes spaced 7/64" apart. The grate was about 1/8" thick and was supported on the kiln floor with refractory bricks. It was hoped that air drafting through the bed from below would increase the air-fuel mixing and improve the combustion. The grate was much more effective when used with the pecan shells than with sawdust. This was due to the larger size of the shells. When sawdust was fed onto the grate, much of the fuel fell through the holes to the kiln floor.
The kiln was preheated with natural gas to about 1200 °F (650 °C)
before pecan shells were spread on the grate at a rate of about 300 grams
per minute. The shells seemed to ignite much more quickly when they were
burned on top of the grate, igniting after about five seconds. When burned
on the kiln floor, they would smolder for 10-15 seconds before igniting.
Figures 10 and 11 below compare the emissions of particulate and CO from
burning pecans shells with and without the grate.
The emissions data indicate that the grate improved the combustion of the pecan shells. Particulate emissions were reduced by about 15% and CO emissions were reduced by one half, from 240 ppm to 110 ppm.
Co-Firing Waste Fuels and Natural Gas
This strategy consisted of firing a small amount of natural gas above a bed of burning sawdust to complete the combustion of the particulates and CO given off by the solid fuel. The total firing rate was held constant at 20,000 Btu/hr-ft2 (63,000 W/m2) while the percentage of the total heat input from natural gas was varied from 0% (all of the heat from burning sawdust) to 25% (75% of the heat from burning sawdust). The natural gas was fired from a single venturi burner mounted in the center of the south wall of the kiln, 18" above the fuel bed. As in the previous baseline tests, the kiln was preheated to a temperature of 1200 °F (650 °C) before the emissions data were taken.
To determine if the reduced particulate and CO emissions were attributable to oxidation of the pollutants in the gas flame or simply the effect of a reduced sawdust feed rate, the measured concentrations were normalized by dividing them by the total mass of sawdust fed during the run. The results are shown below in Figures 12 and 13.
The results from the co-fire runs are mixed. The reduced TSP indicates there may be some combustion of soot particles taking place in the gas flame. However, there is no significant reduction in the CO emissions, beyond the 25% reduction which can be explained by the lower sawdust feed rate. The relatively small reductions of particulate and CO emissions are probably not worth the high cost of implementing this strategy.
Preheating the Kiln with Natural Gas
During the previous experiments, it was discovered that kiln temperature had a greater effect on the emissions than any other factor. This suggested a third strategy for reducing emissions: preheating the combustion chamber with natural gas before feeding waste fuels into the kiln. To determine the effect of kiln temperature on the emissions, the pilot kiln was preheated with natural gas to temperatures from 800 to 1400 °F (427 to 760 °C). The gas was then turned off, and waste fuels (sawdust, pecan shells) were fed into the kiln. The results of the sawdust runs are shown in Figures 14 and 15 below. The startup emissions data are shown at T=600 °F. Similar results were obtained for pecan shells.
Between 800 and 1200 °F (427 and 650 °C), the particulate emissions (measured as TSP) were reduced from 330 mg/SCM to 70 mg/SCM, a reduction of almost 80%. Between 1200 and 1400 °F (650 and 760 °C), the TSP increased slightly. A similar trend was observed when burning pecan shells (see pecan shell data in the Appendix). As the preheat temperature of the kiln was increased from 800 to 1400 °F, the concentration of CO in the exhaust was reduced from almost 1400 ppm to less than 200 ppm, a reduction of 85 %.
A color change in the particulate filter cake was also observed between 1200 and 1400 °F (427 and 650 °C). At 1200° and below, the cake appeared black, but at 1400°, the color changed to a light gray. From this it was concluded that at lower kiln temperatures, the particulates were mostly soot, while at 1400°, they were mostly ash. Carbon analyses of the filter cakes support this conclusion. The results are shown below in Table 2. The values shown in the table are the average of two analyses.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Because of the increase in measured TSP at 1400 °F, as shown in Figure 14, and the very low carbon content of the particulate collected at this temperature (see Table 2), it was hypothesized that the decrease in the amount of soot particles between 600 and 1200 °F (315 and 650 °C) was offset by an increase in the amount of fine ash particles entrained in the kiln exhaust.
To test this hypothesis, aerodynamic particle size distribution data
were obtained at several kiln temperatures. First, size distribution data
were collected under start up conditions, while burning sawdust and wood
with the kiln initially at room temperature. Then the kiln was preheated
to 1000, 1200 and 1400 °F. Both sawdust and pecan shells were burned
at these temperatures and the particle size distribution data were collected.
Figure 16 compares the particle size distributions obtained during kiln
start up with those obtained while burning sawdust at 1000 °F (538
°C).
This figure shows that below about 1000 °F (538 °C), the particle size distribution is not greatly affected by the kiln temperature. Despite a temperature difference of over 400 °F (222 °C), the two size distributions are similar, except in the 2-6 µm range, where burning sawdust at 1000 °F emitted a higher mass fraction of particles. Under start up conditions, the kiln released a lower mass fraction of particles in the 2-6 µm range and a somewhat higher fraction in the smallest size ranges (<0.6 µm). However, these differences are probably due more to the difference in the fuels burned (sawdust and wood during start up vs. sawdust alone at 1000 °F) than to the difference in temperature.
Figure 17 shows the influence of kiln temperature on the particle size
distribution when burning sawdust at temperatures above 1000 °F. Similar
results, obtained from burning pecan shells, are found in the Appendix.
Conclusions and Recommendations
Of the three strategies tested for reducing harmful emissions from Mexican brick kilns, preheating the combustion chamber of the kiln to a temperature of about 1200 °F (650 °C) using a clean-burning fuel such as natural gas, resulted in the highest emission reductions. At 1200 °F, emissions of total suspended particulate (TSP) were reduced by 80% and CO emissions were reduced by 77% when compared to emissions during start up in a cold kiln. THC concentrations were reduced from an average of 420 ppm during kiln start up to negligible levels (less than 20 ppm) at 1000 °F (538 °C).
Burning the waste fuels on top of a grate resulted in much smaller, though significant, reductions when burning a larger waste fuel like pecan shells. The improved fuel-air mixing created by the grate reduced TSP emissions by about 15% and CO emissions by 50% while burning pecan shells at 1200 °F. Use of this strategy does not require a gaseous fuel, which makes it the most economical of the three strategies to implement. However, much greater emissions reductions can be realized if a fuel grate is used in conjunction with kiln preheating.
Firing natural gas above a bed of burning waste fuels did not burn significant amounts of particulate or CO. The 25% reduction in these emissions due to the reduced amount of waste fuel burned is probably not worth the high cost of the gas fuel used.
Above 1000 °F (538 °C), the kiln temperature had a large effect on the size distribution and the composition of the particles formed. At 1000 °F, sawdust produced particles that were mostly soot (54.5% carbon). At this temperature, only 32% of the particles (by mass) had an aerodynamic diameter less than 0.4 µm. At 1400 °F (760 °C), the particles were mostly ash (0.07% carbon), and 80% had an aerodynamic diameter less than 0.4 µm.
References
Find out more about SCERP's other projects involving the Cd. Juárez brickmakers.
Return to SCERP's home page • Contact the SCERP webmaster • Site map
Last updated 6/10/99