TOOL TO PREDICT AND SPECIFY MERCURY, SO₃,
CONDENSIBLES, METAL HAPS AND PARTICLE REMOVAL

 ICAC Conference

March 8, 2005

 Robert McIlvaine
McIlvaine Company

Problem: Utilities who are installing scrubbers for SO₂ removal cannot readily determine or obtain guarantees relative to the mercury, condensibles, discrete particles, and SO₃ which will be removed by the proposed scrubber.

Solution: A 100 cfm “Dust Difficulty Determinator” developed 30 years ago can initially determine what can be done and then can be used as the basis for supplier guarantees. (Figure l)

Analysis

Particulate

SO₂ scrubbers remove discrete dust particles.  If a precipitator is emitting more than 0.05 lbs/MMBtu of discrete particulate, there is likely to be substantial reduction in the downstream scrubber. But there is a big variation in dust removal performance between scrubber designs. Research has shown that the efficiency is directly correlated to the zone of maximum concentrated energy. 

A simple orifice scrubber (Dust Difficulty Determinator)  can then be operated at increasing energy input (velocity through the orifice) until the desired removal efficiency is obtained.

FIGURE 1. Dust Difficulty Determinator Graph as detailed below:

This concept was used to accurately predict the particulate removal performance of many SO₂ scrubbers including some as large as 1200 MW in 1968-72, when scrubbers replaced precipitators instead of following them. The more stringent particulate requirements of the Clean Air Act eliminated this practice.

This device was also used to disprove the claims of novice scrubber suppliers who thought their design would remove particles with much less energy consumption than the proven designs.  Even though particle size analyses seemed to support their contentions, when the Dust Difficulty Determinator was operated in parallel to the novel device it equaled the novel device performance at low energy input.

Not only did the Dust Difficulty Determinator disprove the claims of the novel scrubber supplier, it proved that particle size analyzers are very unreliable and cannot be used to predict scrubber performance.

Many utilities presently installing scrubbers are faced with reduction of discrete particulate including metal HAPS and particulate mercury. It is therefore critical that they know in advance the amount of these pollutants which will be removed.  The way to do this is to operate the Dust Difficulty Determinator on a 100 cfm slip stream from the stack and vary the pressure drop until the optimum removal is achieved.

Dust Difficulty Determinator Performance       

                                  

Then the utility should incorporate into its specifications a requirement that the SO₂ scrubber supplied by the system vendor equal the performance of the Dust Difficulty Determinator at X pressure drop across the orifice.

This requirement can be extended from just discrete particulate to also include condensibles, SO₃, metal HAPS, and particulate mercury. The removal efficiency of each of these pollutants is a function of the concentrated energy and therefore can be accurately predicted by the Dust Difficult Determinator.

Mercury

Mercury occurs in a number of forms in the environment – metallic (chemical symbol Hg[0]), oxidized forms (chemical symbol Hg[1] or Hg[2], depending on the degree of oxidation), and as part of organic compounds (CH3HgCl – monomethyl mercury, combined with chloride, is the most common organic form).

EPA has touted the co-benefits of SO₂ scrubbers as the most important way to remove mercury. However, a utility seeking a guarantee from a scrubber vendor is told that the guarantee will be qualified based on the speciation of the mercury. The greater the percentage of Hg(0) the lower the scrubber performance. Also there are a number of oxidized species. To complicate matters there is potential for additives such as bromine which would add oxidized species which do not exist in the normal flue gas.

With all these variables the mercury removal is directly analogous to dust removal. The Dust Difficulty Determinator should be operated on a slip stream and the pressure drop varied and total mercury removal measured. The optimum pressure drop should be determined.

SO₂ system vendors would then be required to provide scrubbers which are as efficient as the Dust Difficulty Determinator at X pressure drop. The 100 cfm DDD would then be operated in parallel to the commercial system to determine compliance guarantees.

Action Required

EPA, EPRI, or ICAC should take the lead and agree on a universal standard version of the Dust Difficulty Determinator. This device should then be available to all testing companies, suppliers and research organizations. Benefits are potentially huge including quick resolution of the question of whether the 34 tons of mercury is the correct co-benefit level.

Other Designs and Uses

The air pollution industry is equivalent to the motor industry without the concept of horsepower. The development of miniature reference devices with variable efficiency would rectify this situation.

Reference devices run in parallel with full scale designs and can be used for maintenance purposes as well as for prediction and guarantees. Take the example of a precipitator which was measured at 0.03 lbs/MMBtu emission previously but now is emitting 0.05 lbs. Is this deterioration in performance a function of changes in the coal and combustion conditions or is it a precipitator problem. If a reference 100 cfm precipitator were run in parallel to the full scale unit at both times, the question would be answered.

So in addition to the scrubber DDD the following devices should be standardized in miniature scale (e.g. 100 cfm).

1.      A fabric filter (e.g. Square footer) with a series of fabrics from low efficiency to high efficiency.

2.      A dry precipitator with variable SCA just by varying flow.

3.      A wet precipitator with variable SCA just by varying flow.

4.      A fabric filter with activated carbon injection variable from 0-10 lbs/MM acfm for mercury removal.

A further consideration is the use of these devices by regulatory agencies to define BACT AND MACT.  Instead of just a fabric filter without specificity being cited as BACT the regulation could site a reference device with a specified generic fabric of known efficiency.

There is a valid argument that regulations are written to favor one type of equipment by citing, for example, a generic fabric filter.  But if the regulation were more precise and specified a reference device with a specific fabric, the vendor of a novel precipitator could easily run the reference device in parallel and prove that his device was as efficient.