November 1, 1999, through October 31,
2000
Project
Title:
Plasma-Assisted Removal of
SO2/NOx/Hg from High-Sulfur Coal Combustion
Gases
ICCI Project Number: 99-1/1.2B-2
Principal Investigator: Shirshak Dhali, Southern Illinois University
Other Investigators: T. Wiltowski, Southern Illinois University
ICCI Project Manager: Ronald Carty, ICCI
ABSTRACT
The objective of the research is to develop a novel process to replace reagents with plasma in a wet scrubber for SO2/NO/Hg removal. A multi-stage reactor has been fabricated and the system has been tested. The reactor is modular, and additional stages can be added in parallel to scale to larger gas flow. Results have been obtained on power consumption as a function of number of stages. The power consumption increases almost linearly with the number of stages. The SO2 removal scales linearly with the reactor length, applied voltage to the reactor and the power frequency. The removal also increases with increasing water flow through the reactor.
A slipstream has been designed from the scale-up parameters. The slipstream specifications are for a 10-15 cfm flue gas from a stoker unit at the Southern Illinois University Physical Plant.
The results of the tests on the multistage wet-plasma reactor are very promising. This is the first report of data from a wet plasma scrubber. The removal efficiency is far better compared to the dry plasma techniques being investigated throughout the world.
EXECUTIVE SUMMARY
The objective of the research is to develop a novel process to replace reagents
with plasma in a wet scrubber for SO2/NO/Hg
removal. The plasma-based reactor
has been developed and fabricated.
It consists of four modular reactors in
parallel. Additional stages
can be easily added to handle increased gas flow without additional pressure
drop. We have designed and tested
plasma reactors that can flow a maximum of 2
cfm. For efficient plasma
generation, the tube diameter is limited to about 2
cm. This limits the flow through
each unit. With 6-8 units in
parallel, the flow rates of 10-15 cfm can be achieved for the slipstream
tests.
We have calculated the pressure drop in our reactor for a flow rate: 5000 sccm = 83.3 sccs = 0.0029 ft3/s. With these Assumptions: gas temperature = 311K, Density of gas = 0.072 lb/ft3, and Viscosity = 1.90 x 10-5 Pa-s the Re > 2200 indicating that there is a turbulent flow. Therefore we may calculate the pressure drop using the orifice equation:
DP = C1 rg U2
Where: DP = pressure drop in H2O/ft packing, C1 = constant for a packing. Since we are using the glass spheres 5 mm in diameter, we assume C1 = 0.8, rg = gas density, lb/ft3 and U = superficial gas velocity in ft/s. The superficial velocity will be influenced by the fill factor of the packed beads. For hard spheres, assuming closed hexagonal packing, the filled space is approximately 74%. Therefore the superficial velocity will increase, and for annular region of 0.75 inch OD and 0.25 inch ID, the velocity = 8.5 ft/s.
Therefore for our system,
DP = 0.8 x 0.072 x (8.5)2 = 4.2 in H2O/ft packing or (7.5 mm of Hg/ft)
The actual pressure drop will be higher due to under estimation of superficial velocity. Different glass bead sizes were investigated for the packed bed dielectric. The three sizes are 3mm, 4mm, and 5mm. We measured the pressure drop across the reactor for these beads. Due to low values of pressure drop, the experimental measurements were not very accurate. We estimate the pressure drop in the range of 3-5 mm of Hg per 2 cm of the reactor height from the largest to the smallest bead. This was done with a tube of 1.6 cm ID. Typical reactor lengths to be expected are about 10 cm. This would result in a pressure loss of 30-50 mm of Hg.
We have studied the power requirement for different number of stages (up to 4). The power requirement is very much a linear function of the number of reactor. The load is highly capacitive typically with a leading power factor of 0.1. Our results indicate that the kVA (reactive power) can be reduced significantly by introducing an inductor in parallel with the reactor. The load capacitance can be calculated for coaxial reactor geometry. We estimate for four reactors in parallel the capacitance is approximately 1 nF. For an operating frequency of 5 kHz, the required inductance is about 1 Henry.
The removal of SO2 scales very well with voltage and length of the reactor. Clearly the reactor design lends itself to scaling to handle large flow rates and large SO2 concentrations. Our current power source is not capable of going up any further to determine the saturation limits.
Removal studies were done for no water flow and with different water flows through the reactor. Data from the CRC physical and chemical handbook shows the solubility of SO2 in cold water (0 oC) as 22.8 g/cc and in hot water (90oC) as 0.58 g/cc. In room temperature water we see a reduction of SO2 from 3500 ppm to 2330 ppm with water alone (without plasma). For 240 ml/minute of water flow through the reactor, this amounts to 5.5 mg of SO2/minute. The concentration of SO2 in water is 2.3 mg/100 cc of water. Therefore the nearly 30% reduction in SO2 concentration with water alone is well within the solubility range of SO2. With the plasma turned on, the removal efficiency falls from 84% to 40% when water flow is stopped. Clearly, the process under investigation with water flow in the plasma is superior to the dry systems being currently investigated by other investigators.
Preliminary results show that there is slight reduction in the removal of SO2 with a combined flow of 600 ppm NO. This result is to be expected because NO and SO2 compete for the same radicals (OH) for removal.
Shown below is the design specification for the slipstream:
1. Slipstream flow rate of 10-15 cfm of flue gas at 300-4000F.
2. Number of modular reactor required for the above flow rate is 4-6.
3. Water flow rate is 200 ml/minute per cfm of flue gas.
4. Power source of 100-300 V at 1.5 kVA.
5. Transformer 100:1 at rated voltage of 30 kV(RMS) and 1.5 kVA.
The SIU physical plant has several facilities for generating steam and power. It uses Illinois coal #6. Its main facility is based on fluidized bed combustion technology. The SOx/NOx concentrations from fluidized bed combustor are low and not suitable for our tests. In addition the physical plant has three stoker units, which are used for additional steam generation during peak season. After meeting with the physical plant managers it was decided that using the flue gas out of the stoker units would meet our specification.
The ductwork consists of 1-inch pipes capable of flows of 10-25 scfm. Since the pressure is negative a fan (Cole-Palmer) is used after the plasma reactor to maintain a flow. A gas chromatograph for the SO2 and NOx analysis is part of the on-line diagnostics. This instrument is Gow_Mac 550p equipped with TC (thermal conductivity detector) and Porapak Q columns. The gas at the exit of the rector is sampled with a six-port VALCO sampling valve.
The flue gas from the boiler (steam capacity of 80 lb/hour at 300oC) goes through an economizer and bag-house and from there to the stack. The slipstream is designed to be right after the boiler. The flue gas temperature is in the range of 400-600oF depending on the load. The expected SO2 concentration is in the range of 1500-2500 ppm. The NO2 concentration is in the range of 500-800 ppm. In addition the flue gas at this point has particulates.
As a conclusion, the testing of the modular reactors provided data for the scale-up to slipstream tests. The fractional removal of SO2 scales linearly with the length of the reactor, the voltage and frequency. Increased flow is handled by increasing the number of reactors without additional pressure drop. This is extremely encouraging for scale-up, since we have not reached saturation levels for the removal of SO2. As a result of this research, we were the first to show that large quantities of SO2 removal works better with an wet plasma reactor instead of the dry reactors, which are being investigated by other researchers. A complete set of design specifications were obtained for slipstream tests with a flow of 10-15 cfm.
Funding cycle of 2 years would be needed to perform thorough slipstream tests. A 1-year funding is not adequate to perform reliable slipstream test that can be carried one-step further to pilot plant studies.