![[logo] US EPA](https://webarchive.library.unt.edu/eot2008/20081107212525im_/http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Final Report: Development of a Membrane-Based Electrostatic Precipitator
EPA Grant Number: R828171Title: Development of a Membrane-Based Electrostatic Precipitator
Investigators: Pasic, Hajrudin , Alam, Khairul , Ingram, David
Institution: Ohio University - Main Campus
EPA Project Officer: Shapiro, Paul
Project Period: August 1, 2000 through July 31, 2002
Project Amount: $225,000
RFA: Exploratory Research - Environmental Engineering (1999)
Research Category: Engineering and Environmental Chemistry
Description:
Objective:This research project was conducted on membrane-based electrostatic precipitators (ESP) that are to be used instead of traditional plate-based ESPs. The two major objectives of this research project were to: (1) study the influence of flue gas-induced vibration of the membranes on the dust dislodgment; and (2) study the coatings of the fiber-based membranes to change/improve their properties, primarily their electroconductivity and abrasion resistance.
Summary/Accomplishments (Outputs/Outcomes):Experimental studies on combining flow-induced vibration and conventional rapping techniques for dislodgment of fly ash were first conducted at room temperature on a bench-scale ESP, with 30 x 30-cm thin carbon fiber membrane. To make comparisons, a series of similar experiments were conducted on a stiff steel plate of the same size. To measure fly ash removal efficiencies, we used: (a) a sonic horn in conjunction with both the metal plate and the membrane as a collection electrode; and (b) a pneumatic tension rapping to clean the membrane.
Velocity of the fly ash-laden gas ranged from 1 to 1.5 m/s, the distance between
the grounding and collection electrodes was 10 cm, and the operating voltage
was 30 kV supplied by the 100 kV, 0-10 mA Hipotronix transformer/rectifier (TR)
unit. For cleaning fly ash from collection electrodes, the Kockum Sonics horn
model MKT 75/440 has been used. The horn operated at 440 Hz and was run by compressed
air, producing the pressure level of 130 dB. In rapping experiments, a pneumatic
mechanism was used to stretch the membrane with a force of about 10 N/cm while
collecting the fly ash and then suddenly releasing it to dislodge the ash.
Some preliminary studies of flow-induced vibration of membranes in clean airflow
also were performed in a wind tunnel on a loose, 20 x 20-cm membrane. Vibrations
at flow speeds of 1-2 m/s were recorded/measured at room temperature.
Because the first results on both rapping and vibration indicated that scaling
effects are important and that there was a need for a larger experimental facility,
we decided to build a pilot ESP unit. For that, Ohio University (Ohio U) dedicated
a 2,500 sq ft ESP Laboratory area, part of a former Corrosion Center. The laboratory
space was renovated thanks to the funds provided by the Ohio Board of Regents.
Parallel to this project, the Ohio Coal Development Office (OCDO) funded some
ESP-related research as well, and the equipment in the ESP Laboratory was acquired
through these combined research efforts. These activities had not been planned
originally, because it was anticipated that studies conducted on the existing
bench ESP unit would suffice. This resulted in a prolongation of the project
by a few months, but also in building the state-of-the-art ESP Laboratory, which
will significantly enhance the future of Ohio U's ESP-related research capabilities.
The pilot ESP unit was then used to repeat the experiments obtained on the bench-scale
ESP-this time on 1.8 x 0.6-m carbon fiber membranes installed in the main ESP
section, which is 5.4 m long, 1.2 m wide, and 2.1 m tall. The pilot ESP also
is equipped with the same sonic horn and the pneumatic rapping mechanism as
in the bench-scale ESP. The flue gas is sucked by a 12,000-cfm fan installed
outside the laboratory. The power is supplied by the NWL Inc. AC/DC 70 kV, 0-400
mA switch mode power supply transformer unit. The ESP also is equipped with
a VibraScrew Inc. dust feeder, a differential mobility particle-size analyzer
(TSI 3071), a condensation nuclei counter (TSI 3022), etc. The membrane vibration
was monitored/measured with the Agilent dynamic signal analyzer 35670A and the
353B16 accelerometer.
Altogether, about 100 ash-removal efficiency tests were performed. Figure 1 summarizes the results of bench-scale tests performed on a 30 x 30-cm membrane/plate. Measured are the amount of dust (ash) removed by the gas flow, the amount of ash removed by the sonic horn/pneumatic rapping mechanism, and the amount of dust remaining on the substrate collection electrode. Combined flow-induced vibration of the membrane and the abrasion due to the gas flow removes about 34 percent of the dust even before cleaning takes place. Although the horn was operated at a very high frequency, sound-induced vibrations also are more intense with the membrane than with the plate, resulting in a higher cleaning efficiency. Finally, the amount of uncleaned dust (remaining on the membrane) is much smaller than on the plate or on the membrane cleaned by rapping, by pulling it.
Figure 1. Bench-ESP Experiment. (1) Dust removed by gas flow; (2) dust
removed by sonic/horn/rapping; (3) dust remaining on membrane/plate, when: 
membrane
is cleaned by sonic horn, 
plate is cleaned by sonic horn, 
membrane is cleaned
by stretching.
Figures 2, 3, and 4 present some of the results obtained in pilot-scale experiments in which the 180 x 60-cm carbon-fiber woven membrane was cleaned with the same sonic horn used in bench-scale experiments. The membrane was first kept straight but with minimal tension force, and the dust-removal efficiency was tested against gas flow speed (see Figure 2) and against dust-collection time (see Figure 3). Clearly, by increasing the gas flow speed, vibrations induced in the membrane are enhanced and more dust is removed during the collection time. At a gas speed of 1 m/s, that portion is about 9 percent, while at a speed of 1.5 m/s, it is 16 percent. The influence of the flow-induced vibration is even more pronounced as the dust layer gets thicker (see Figure 3).
Figure 2. Dust-Removal Efficiency Versus Gas Flow Speed in Pilot-Scale
ESP:
Dust Remaining on Membrane,
Dust Removed by Horn,
Dust Removed by Gas Flow
Figure 3. Dust-Removal Efficiency Versus Collection Time in Pilot-ESP
After: (1) 10 minutes; and (2) 20 minutes:
Dust Remaining on Membrane,
Dust Removed by Horn,
Dust Removed by Gas Flow
Figure 4 illustrates the importance of keeping the membrane loose to enhance both flow-induced and sound-induced vibration. Finally, Figure 5 compares dust-removal efficiencies from the large (pilot-scale) and small membrane, both kept at minimal tension. It is seen that sonic horn cleaning is more efficient when applied to large membranes, indicating that even better results could be expected if this technology is applied in industrial ESPs with large membranes.
Figure 4. Dust-Removal Efficiency Versus Tension Force in Pilot-Scale
ESP:
Dust Remaining on Membrane,
Dust Removed by Horn,
Dust Removed by Gas Flow. Dust collection time: 20 minutes.
Figure 5. Dust-Removal Efficiency for: (1) Large; and (2) Small Membrane:
Dust Remaining on Membrane,
Dust Removed by Horn,
Dust Removed by Gas Flow.
The results of the experiments using sonic horns to clean a fly ash in membrane-based dry ESPs clearly indicate that this technology can be more beneficial than when the same technology is applied on plate-based ESPs. Light membranes can more easily be excited by the propagating sound wave. In addition, vibration induced in the membrane by the oncoming gas flow also contributes to enhanced fly ash removal, because some of the attracted particles do not settle on the vibrating membrane, but rather, "flow" down into a hopper beneath. Comparisons of the results on a small-size membrane and a larger size membrane indicate that these effects should be even more pronounced in real dry ESPs with much larger membranes.
In the experiments presented here, a small-size sonic horn was used. It operated at too high a frequency-440 Hz-much higher than the optimal one that would be able to bring the dust layer to a resonance; about 120-150 Hz. It is, therefore, expected that utilization of a more appropriate horn would further increase dust dislodgment efficiency. Clearly, additional experimentation on membranes in real-size ESP and utilizing adequate horns is necessary.
For Objective 2, the ultra-high vacuum magnetron-sputtering system to deposit titanium and aluminum on fiber strands was used (see Figure 6). This research has been conducted in collaboration with the Ohio U Physics Department and their Accelerator Laboratory. Initially, the focus of studies was to deposit titanium dioxide on silica fabrics and yarn. The coatings were expected to be used not only to improve mechanical and chemical properties of fibers/membranes, but also for adsorption of mercury from exhaust gases, because the control of mercury vapor emissions has recently become a critical clean air issue. A rather sophisticated apparatus (see Figure 7) was developed and built to move the fiber strands within a relatively small space in the magnetron, and the coating was produced and tested for conductivity and quality.
Figure 6. The Magnetron Sputtering System
Figure 7. Coating Fixture
The results from the coating of silica yarns and fabrics (to make them conductive) demonstrated that the coating of nonconducting fibers or fabrics by a conducting metal (for high-temperature ESP applications) is too complex, and hence, prohibitively expensive. In addition, the coating is too brittle and tends to break apart. Therefore, it led us to consider the utilization of conductive fabrics for application to ESPs. The most common conducting fabrics that are resistant to corrosion/oxidation are the carbon fabrics. However, the difficulty with using a commercial carbon fabric is that very thin carbon fibers (few microns) that break often tend to act like a pointed electrode that, in a high-voltage ESP field, create an electrical short and are eventually removed by abrasion. This leads to malfunction of the ESP, which must be avoided. In the end, it was concluded that a promising method for eliminating all of those problems is to put a coating on the carbon fabric, not on the individual fibers/strands. A novel solution could be to take a carbon fabric and coat it with a thin Teflon layer, so that the conductivity is not affected significantly, but the body of the fabric is held together by the polymer so that frayed carbon filaments remain bound to the fabric. This can be further improved by adding short carbon fibers to the Teflon coating so that a conductive path is maintained between the carbon fabric and the surface of the Teflon-coated carbon fabric, as well as by impregnating the mixture between the membrane strands/fibers. Preliminary coating tests with pure Teflon and Teflon/carbon fiber mixtures have been successfully conducted (see Figure 8).
Figure 8. Carbon-Fiber Membrane (30 x 30-cm) Coated/Impregnated With Teflon
Next, various types of fiber-based membranes, with and without coating, have been tested for their strength and durability by installing them in Ohio U's heating plant's ESP inlet duct that connects the furnace and the ESP. The plant has four boilers and fires Ohio's bituminous coal with about 10 percent of ash (weight percent). The flue gas contains 12.5 percent carbon dioxide, 82.4 percent hydrogen, and 5.1 percent oxygen (volume percent). The membranes-about 30 x 30-cm in size-were mounted in metal frames and set parallel to the gas stream, 3.6 m from the furnace, and 9 m before the ESP, in a 1.5 x 1.5-m duct and were exposed to the 326°F temperature for 6 months. Two Teflon-coated silica fiber woven membranes and three uncoated carbon fiber woven membranes were tested. Tests included measurements of loss of membranes' weight and their tension strength-as a possible indicator of a damage induced primarily by abrasion by the particle laden gas flow and by high temperatures.
Contrary to all expectations, after 6 months, the membranes seemed to be undamaged and had no apparent flaws. Only the Teflon coating was partially abraded from the silica membrane, probably as a result of an inadequate coating quality. The other Teflon-coated silica membrane appeared untouched and had insignificant loss of weight. However, uncoated carbon fiber woven membranes gained up to 30 percent in weight; the increase being attributed to the fly ash self-impregnation between the membrane strands/fibers. This also resulted in increased smoothness of membranes with almost unnoticeable strands/fibers, and they appeared as if they were coated with tar.
After these tests, the membranes were tested for mechanical strength (i.e., the force required to rapture them was measured on 25 x 250-mm specimens), using a Tinius Olsen tension testing machine. This force then was compared with the force required to rupture the virgin membrane. The relative force loss required to rupture the membranes that had undergone harsh ESP conditions gives an indication of potential usefulness of the membrane.
Typical results for the carbon membrane are shown in Figures 9 and 10. After 6 months, the membrane tension strength dropped considerably. However, it is still large enough to keep the membrane stretched/straight, because the force required for that is less than 5-10 N (per inch)-much less than the value of 400 at rupture (see Figure 10).
Figure 9. Rupture Test of Virgin Carbon FDI-1300 Fabric.
Figure 10. Rupture Test of ESP-Treated Fabric.
Because the weight loss was insignificant, it was concluded that the drop in membrane strength was primarily due to its exposure to high ESP temperatures rather than abrasion (i.e., that the damage was induced at the very beginning of the 6-month membrane exposure to high temperature in the ESP inlet duct-perhaps within the few first hours or during the first day). To check for that, we kept the membrane fabrics in a laboratory furnace overnight at the same temperature as in the ESP inlet duct and tested its tension strength afterwards. We found that our hypothesis was confirmed. However, it is clear that the laboratory furnace environment differs from the one in the ESP duct primarily in the differing oxidizing mechanisms. Hence, these results are still to be regarded as very encouraging but inconclusive. Therefore, although fiber-based woven membranes can most likely sustain harsh ESP environments for time periods much longer than 6 months, the longest lasting period that they can stay in full service is still not clear. The best way to check for that is certainly a long-term exposure of membrane materials in real ESP environments and their periodical rupture testing to clarify damage progression and, based on it, make extrapolations.
In conclusion, our experiments have shown that using fiber-based woven membranes as collection electrodes in electrostatic precipitators is a very good alternative to using rigid plates, especially when used in conjunction with sonic horns as a means for cleaning/rapping precipitated fly ash. However, experiments on coating membranes or their strands/fibers to protect them or change their properties have turned out to be virtually unfeasible.
On the other hand, we have been pleasantly surprised by findings that the membranes, when exposed to the real ESP environment, endured those harsh conditions for a long time due a to kind of self-protection, as they have been coated and impregnated with all kinds of flue gas ingredients. Although the mechanical strength dropped, it is still far above the necessary minimal bearing capacity. However, it is not clear if that drop is a final one or whether it will further weaken in time. Therefore, more mechanical tests on membranes exposed to real ESP environments for longer periods-several years or more-are needed.
Possible Applications and Commercialization. The membrane-ESP technology has been developed by Ohio U (US Patent 6,231,643) and licensed to Southern Environmental, Inc. (SEI), Pensacola, FL. Thus far, SEI has built a pilot unit on a Lime Kiln at Georgia Pacific's Cedar Springs Georgia Plant. It operates as a wet ESP and the results are excellent.
SEI currently is building another pilot wet ESP unit at First Energy's Bruce Mansfield Power Plant to collect primarily SO3 mist (sulfuric acid mist). This is a U.S. Department of Energy (DOE)-funded project. It will start operation in June 2003.
Thus far, membranes have not been implemented in dry ESPs, primarily because it was initially thought that they would not sustain high temperatures and abrasion. Although still not conclusive, our research results presented above indicate that it is not the case and that carbon fiber-based woven membranes, if combined with sonic horns to clean the dust, could be a very good solution to replace plates in ESPs. It is therefore to be expected that the SEI could start production of dry membrane ESPs soon.
Journal Articles:No journal articles submitted with this report: View all 2 publications for this project
Supplemental Keywords:electrostatic precipitators, membranes, woven membranes, fly ash, cleaning, rapping, sonic horns, coating, fibers, precipitation, particulates, clean technologies, innovative technology, cleanup, engineering, ecology, measurement methods, chemicals, pollution prevention, modeling. , Air, Scientific Discipline, Engineering, Chemistry, & Physics, Environmental Chemistry, heavy metals, ambient emissions, particle chamber, chemical composition, membrane-based, particulates, gaseous organic compound, atmospheric particles, gas flow rates, heavy particles, air quality standards, air modeling
Progress and Final Reports:
2001 Progress Report
Original Abstract