Double Loop Wet Limestone Scrubber Chemistry Control

This article is written to operators to provide information on the chemistry control of a double loop scrubber.  This article covers normal operation of the double loop scrubbers and does not address special circumstances such as aluminum fluoride blinding and particle size crashes.
I operated double loop scrubbers with a quencher capacity of approximately 60,000 gallons and an absorber feed tank of approximately 180,000 gallons.  The diameter of the tower was 36 feet.


Chemistry control variables that need to be carefully monitored and controlled are:

• pH
• Oxidation Air
• DBA concentration
• Percent Solids
• Gypsum purity
• Chloride concentration
• Anti-foam addition
• Mist eliminator wash

The goals of maintaining these variables are to obtain the following results:

• Maintain or exceed the removal efficiency of the tower
• Maintain gypsum purity according to specifications
• Maintain good crystal growth with low fines production

For instance, I operated two towers that served unit 3.  Unit 3 operated at a maximum load of 420 MW.  The removal efficiency required by the scrubbers based upon the Consent Decree was 95% combined on a thirty day rolling average which was measured by the continuous emissions monitoring systems (CEMS).  The gypsum purity of the slurry was maintained at 95% calcium sulfate as measured by thermo gravimetric analysis in the lab.  The crystal size of the gypsum was measured by particle size distribution in the lab and the average particle size that we tried to maintain was 50 microns with a fines distribution of 5% or less under 10 microns.
For unit 4, two towers serviced it.  Unit 4 maximum load was 485 MW.  The removal efficiency of the towers combined was 90% on a thirty day rolling average based upon the EPA Title V document for air pollution control.  The gypsum purity of the slurry was maintained at 95% calcium sulfate as measured in the lab by thermo gravimetric analysis.  The crystal size of the gypsum was measured in the lab by particle size distribution.  The average particle size that we tried to maintain was the same as the towers of unit 3 which was 50 microns and a fines of 5% or less under 10 microns.

PH

The most important operational parameter of the pH was monitoring the pH trend of the absorber feed tank and quencher using Ovation trends and Plant Information System Trends (PI).  We maintained a pH saw tooth pattern.  This was the most effective way of monitoring if the towers were operating properly.  It behaves like a heartbeat.  A loss of the pH saw tooth pattern would mean that something was wrong with the tower performance.  Therefore, it was very critical that the pH saw tooth pattern was maintained in both the absorber feed tank and the quencher.  Figure 1 shows what the pH saw tooth pattern should look line.  The pH pattern does not have to be a perfect, sharp saw tooth. 

Figure 1

The set point for the pH of both the absorber and quencher was set as high as possible as long as a good pH saw tooth pattern was maintained and the DBA concentration, absorber oxidation air flow, and quencher oxidation air flow was adequate to maintain the required removal efficiency.  The thermo gravimetric analysis of the absorber feed tank slurry solids and the quencher slurry solids were used to determine how high the pH could be set. For instance, if the calcium sulfate was 98% then there is room to increase the pH until the calcium sulfate was closer to 95% as long as the pH pattern is not lost.  If the pH pattern was lost or became too sluggish, it means that the absorber feed tank or quencher is too rich in calcium carbonate.  If the thermo gravimetric analysis showed that the absorber or quencher was too rich in calcium sulfate, it means that there is not enough calcium carbonate in the slurry, and the removal efficiency would swing with the pH pattern.  We did not want the removal efficiency to swing with the pH pattern.  When we monitored the removal efficiency trend, we wanted the removal efficiency to stay as flat as possible without much swinging. The thermo gravimetric analysis gives calcium carbonate purity.  A purity of 1.5% to 3 % carbonate was sufficient to maintain close to a stable removal efficiency trend.

Oxidation Air

The oxidation air flow should be adequate to maintain a good saw tooth pH pattern and good removal efficiency.  If the oxidation air flow is not adequate, the liquid phase sulfites in either the absorber feed tank or quencher will cause the pH pattern to become sluggish or lazy and the removal efficiency could drop below the removal efficiency limit.  The pH pattern could be restored by increasing the oxidation air flow by opening the oxidation air control valve some more.  

Table 1 shows the ranges of the oxidation air flow for both the absorber feed tank and quencher.  

Table 1

Absorber Feed Tank	6 000 scfm – 9 000 scfm
Quencher	3 666 scfm – 5 000 scfm

At particular intervals the liquid phase sulfite was measured in the lab to make sure that the oxidation air was adequate.  In addition, the thermo gravimetric analysis of the slurry solids would analyze for solid phase sulfites which would indicate if sulfites were precipitating out of solution.  If sulfites were precipitating out of solution, it could decrease the gypsum purity below 95% and make the dewatering of the slurry difficult on the filter drums.  The thermo gravimetric analysis was measured once a day to make sure that sulfites were not precipitating out of solution.   To avoid the loss of removal efficiency and low gypsum purity, the oxidation air headers and sparger for both absorber feed tank and quencher was internally cleaned on a regular basis.  In time the oxidation air headers and spargers become partially plugged which would unbalance the distribution of air in the absorber and quencher or lower the air flow.  Hence, cleaning the oxidation air headers and spargers once every other week worked well.  When an emergency situation occurred such as the loss of an air compressor, then the oxidation air headers and spargers were cleaned as soon as possible.

DBA

DBA which is called in the industry dibasic acid is a mixture of glutaric, succinic, and adipic acids. The addition of DBA helped increase the removal efficiency and maintain the pH saw tooth pattern. The DBA was added to the absorber feed tanks.  The DBA concentration was measured in the lab once per day and as needed as we requested it.  The DBA concentration range was 750 ppm to 1 000 ppm.  The lab measured the DBA using a buffer capacity test.  We tried to minimize the DBA addition (cost savings) to the absorber feed tanks by running the pH as high as possible without losing the pH saw tooth pattern and maintaining the gypsum purity near 95% calcium sulfate.

Percent Solids

The percent solids of the absorber feed tanks was maintained at 15%.  Since the slurry from the absorber feed tank overflowed into the quencher, the quencher percent solids was also maintained at 15%.  We called operating the towers at low percent solids “dilute operations.”  The lab measured the percent solids by collecting a sample of slurry and measuring the specific gravity of the slurry and drying a measured volume of the slurry solids and weighing the dried solids and then calculating the percent solids.  What we have observed with regards to maintaining a lower percent solids is that the gypsum fines production in the absorber feed tanks and quenchers were minimized.  Hence, by running more dilute in the absorber feed tanks, we were able to get lower gypsum moisture off the end product of the filter drums.  We were able to observe the effects of dilute operations by measuring the particle size in the lab using a particle size analyzer and by observing the crystals under a scanning electron microscope.  I also had an optical microscope in the lab that I used daily to look at the crystals since it did not require much preparation as the scanning electron microscope.  We had to send out samples to other labs to prepare pictures using a scanning electron microscope because our lab did not have a scanning electron microscope.  To maintain the percent solids at 15%, we used the mist eliminator wash water to add more water to the tower which then would flow into the absorber feed tanks through the bowl return line.  In addition, I started a project to add a density control feedback loop to add an independent source of water to the absorber feed tanks which should give even better control of the density.  Maintaining the percent solids was very difficult due to load swings of the unit; hence, adding a density feedback control loop should help.

The specific gravity of the absorber feed tank and quencher is monitored in the control room.  The specific gravity set point is set based upon the lab analysis to get the percent solids close to 15 percent.  In other words, we use the specific gravity measured in the lab with its associated percent solids to tune the specific gravity control of the absorber feed tank and quencher.  It is important to monitor the both the specific gravity of the absorber feed tank and quencher to avoid running too dilute which is below 8% solids because scaling could occur in the tower internals.  In addition, if the density of the quencher is too high at least 25% solids, we run the risk of plugging the blowdown line which we definitely want to avoid.  If the blowdown line gets plugged, then the quencher recycle pump drain valves were usually opened to keep the quencher level under control.  The slurry solids tend to settle in the drains and plug the drains and the slurry then overflows the drains and makes a mess.  On unit 4, to avoid putting slurry in the booster fans, the drains had to be opened immediately if the blowdown line was plugged.  Thus, to avoid shutting down the unit, the density of the slurry should be monitored very closely.

Gypsum Purity

I discussed in the section under pH, the gypsum purity of the slurry is measured in the lab using a thermo gravimetric analyzer.  The purpose of measuring the purity in the lab is to maintain the gypsum purity based upon contract specifications and to adjust the pH set point to meet the removal efficiency requirements of the scrubber.  Now what I am about to say may seem contradictory but it may make sense under special circumstances.  

In the absorber feed tank, the gypsum purity could be less than the desired 95% calcium sulfate.  The pH of the absorber could run as high as possible while maintaining a good saw-tooth pattern and the purity be 94.5% calcium sulfate.  The absorber feed tank overflows into the quencher via an overflow line.  As long as the pH in the absorber feed tank is not so high as to saturate the quencher with calcium carbonate such that the quencher pH saw-tooth pattern becomes lazy, you should be able to tolerate a lower calcium sulfate purity in the absorber feed tank.

In the quencher, the pH has to be set in such a manner that the purity should be 95% or better if the limestone grind is poor (70% or less passing 200 mesh).  

If the limestone grind is very fine such as 90% passing a 200 mesh or 300 mesh and the dewatering hydrocyclones are well maintained to give the sharpest cut possible, then the quencher pH set point could be higher such that the quencher slurry purity could be less than 95%.  The reason for this is the fine particles of limestone will be classified in the overflow of the dewatering hydrocyclones while maintaining 95% or better purity in the underflow stream of the dewatering hydrocyclones.   As a result, the end product gypsum purity from the underflow of the dewatering hydrocyclones should be 95% or better.  In the FGD system that I operated, the overflow of the dewatering hydrocyclones were used for density control and makeup water to the quenchers.  Hence, the finer calcium carbonate particles will be returned and dissolved in the quencher.  In my case, I had to operate the pH set point in the quencher to give 95% or better purity because the limestone grind was poor, 70% passing a 200 mesh.

 

Chloride Concentration

In the scrubber that I operated, we maintained the chloride concentration in the quencher at a maximum of 40 000 ppm and in the absorber feed tank at a maximum of 20 000 ppm.  Normally, at full load conditions (i.e. megawatts), the quencher chlorides was approximately   20 000 ppm and the absorber feed tank chloride concentration was approximately 10 000 ppm.  We tried to maintain as low a chloride concentration as possible.  To keep the chloride concentration under control, we discharged a slip stream from the dewatering hydrocyclones after treatment.  The normal discharge was approximately 200 gpm. The reason for controlling chlorides is to minimize corrosion of the scrubber internals.  The quencher was made of Inconel which can handle a high chloride concentration.  The reason that the quencher was made of Inconel was due to the quencher scrubbing out most of the chlorides from the flue gas; hence, the chloride concentration is about double that of the absorber feed tank.  In addition, the quencher protects the absorber feed tanks from very high chloride concentration.  As a result of the quencher decreasing the chloride concentration of the absorber feed tank, the absorber feed tank was constructed using carbon steel.  As of the creation of this article, the absorber feed tanks were replaced with fiber glass tanks.  However, we continued to maintain the 20 000 ppm chloride limit after replacement of the absorber feed tanks.

Anti-foam Addition

The use of an anti-foam liquid was added to the absorber feed tank to minimize foaming in the absorber feed tank and quencher.   When I began working in the FGD an oil based anti-foam was used.  The anti-foam was added by pumping the anti-foam with a diaphragm pump at the rate of 50 millimeters per minute.  Approximately, three years later, we switched to a water based anti-foam after tests showed the water based anti-foam was much more effective at minimizing foaming.  In the lab, slurry was collected from the absorber feed tank and the sample was shaken in the lab and anti-foam added with a syringe to the beaker to see how effective the anti-foamed worked.  In addition, I added Alka-Seltzer tablets to the sample to mimic the effects of oxidation air and added a few drops of the anti-foam using a syringe to observe the effectiveness of the anti-foam.  By trial and error, we can come up with a baseline dosing rate using these tests.  Since the absorber feed tank slurry overflows into the quencher, the quencher slurry got its anti-foam through the absorber feed tank.  

One important fact to remember is that some anti-foams can degrade if the temperature of the slurry is too high.  Hence, when replacing an anti-foam, keep that fact in mind.

 

Mist Eliminator Wash

Each mist eliminator valve was opened a minimum of 1.5 minutes per hour.  The minimum value time was recommended by the mist eliminator vendor as a baseline to start from.  Each mist eliminator valve had its own timer in the controls so that we could adjust the time the valves stayed opened.  We had differential pressure transmitters across the mist eliminator vanes to monitor the possibility of blockage of the vanes.  The mist eliminator flow rate, mist eliminator discharge pressure, and mist eliminator differential pressure was monitored in the control room instantaneously using trends.  If the mist eliminator becomes plugged and cannot be cleaned, the unit has to be taken offline to go inside the tower and clean the mist eliminator vanes manually with the use of water hoses.  Therefore, it is very critical to make sure that the mist eliminator wash is working at all times and all the valves are cycling properly and the flow meter, discharge wash pressure, and differential pressure transmitter are working properly.   Any malfunctions must be addressed immediately.

The consequences of plugged mist eliminator vanes is carry up of slurry into the environment and higher operating pressure in the tower which causes load restrictions.

In the four double loop towers that I operated, each tower had four mist eliminator valves.  Each tower had to share wash water which came from one source, the makeup water tank.  Hence, the mist eliminators was washed in a sequence starting at tower A, tower C, tower B, and tower D.  The logic was setup to open only one valve at a time.

In addition, we used the mist eliminator wash not only to keep the vanes clean, but also to help maintain the percent solids in the absorber feed tank.  Hence, each valve time was opened more than the 1.5 minutes per hour during full load (megawatt) conditions.  During low load conditions, the mist eliminator valve timers were decreased as necessary to maintain the percent solids.

Conclusion

I hope this helps to familiarize the tower operators on the chemistry controls of double loop scrubbers.  This is by no means a replacement for proper training.  It is important to have a good training program in place to train and re-train operators in the chemistry controls available to them.

 

Summary

• Maintain the pH saw-tooth pattern.
• Run the pH set point as high as possible while maintaining the pH saw-tooth pattern.
• Use the lab purity analyzer to determine how high to set the pH set point.
• Add adequate oxidation air to the absorber feed tank and quencher to maintain the pH saw-tooth pattern.
• Keep the oxidation air headers clean by pressure washing the headers and spargers online.
• Add adequate DBA to maintain the removal efficiency.
• Minimize DBA addition by running the pH set point as high as possible.
• Measure in the lab the percent solids and use the density and percent solids from the lab results to tune the density controls of the absorber feed tank and quencher slurry.
• Maintain the gypsum purity in the absorber feed tank and quencher based upon the contract specifications.
• The gypsum purity can be lower in the absorber feed tank and quencher as long as the limestone grind is high such as 90% passing a 200 mesh or 300 mesh and the dewatering hydrocyclones are operating properly.
• Maintain the chloride concentration of both the absorber feed tank and quencher by discharging adequate amounts of treated wastewater.
• Add an adequate amount of anti-foam to the absorber feed tank.
• Determine a baseline feed rate of anti-foam in the lab.
• Run the mist eliminator valve wash water timers above the minimum to keep the mist eliminator         vanes clean and to maintain the density of the absorber feed tank.
• Monitor the variables consistently using computer trends and lab analyses.