EPA-650/2-74-050
June 1974
                          Environmental  Protection Technology  Series

-------
                                 EPA-650/2-74-050
   MEASUREMENT  OF ENTRAINED
LIQUID  LEVELS  IN  EFFLUENT  GASES
    FROM  SCRUBBER  DEMISTERS
                     by

           L. D. Johnson and R. M. Statnick
             Control Systems Laboratory
             Process Measurements Section
        Particulate and Chemical Processes Branch
         Research Triangle Park, North Carolina
                ROAPNo. 21ACY-44
             Program Element No. LAB013
                  Prepared for

       NATIONAL ENVIRONMENTAL RESEARCH CENTER
        OFFICE OF RESEARCH AND DEVELOPMENT
        U.S. ENVIRONMENTAL PROTECTION AGENCY
     RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711

                   June 1974

-------
                          ii
This report has been reviewed by the Environmental Protection Agency
and approved for publication.  Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.

-------
                             iii
                          CONTENTS
Acknowledgements
Introduction
Theory
Experimental Work
Calculations
Results
Discussion
Conclusion
References
Fig.
1
  Title
                     LEST OF FIGURES
Effect of Stack Flow Rate on Mist
 Carry-over
                       LIST OF TABLES
Table
I

n
m
  Title
iv
 1
 1
 2
 3
 6
10
13
13

Page

 8
Input and Output Variables for Calculation
 of Entrainment                          4
Equations for Calculation of Entrainment    5
Experimental Mist Entrainment Results      6

-------
                                  iv
ACKNOWLEDGMENTS
     The assistance of F. E. Briden and R. S. Steiber of EPA  and
Gary Thorn of Zurn Industries is gratefully acknowledged.

-------
INTRODUCTION
     Wet scrubbers have been widely used for a number of years for
removal of pollutants from stack effluents.   A large number of dif-
ferent designs are possible, and many are in practical use today.
Most, if not all, of these devices produce entrained scrubber liquor
droplets in the air stream, and many require demister devices for
removal of this potential chemical pollution prior to discharge.
Although the quantitative determination of entrained liquid levels
in gases leaving scrubber demisters has long been recognized as an
important goal, no satisfactory and convenient method has been
available.  This report describes the theory and testing of a method
based on the use of an existing soluble ion  as a tracer agent.

THEORY
     The theory behind the present determination is fairly simple,
and is typical of most "tracer" type experiments.  In this case,
the assumption is made that the only major source of sodium ion in
the system is the scrubber feed liquor.  If  this is true, then only
entrainment of scrubber liquor can cause sodium ion above the demi-
ster.  If one measures the amount of sodium  ion in a given volume
of stack gas, it is then possible to calculate the total amount of
sodium ion passing through the stack.  The concentration of sodium
ion in the feed liquor to the scrubbers is easily determined, and
is usually constant for a given set of process controls.  Since
this concentration specifies the amount of liquor volume associated

-------
                                 2
with a given amount of sodium ion, it may be readily used along with
the sodium ion level in the stack to calculate entrained liquor
carry-over.  Although a constant sodium level  in the liquor is pre-
ferable, an average value may be used when this concentration varies.
Any major source of sodium other than the liquor would invalidate
the use of sodium as the tracer, but not the concept of the test.
In such a case, it might be necessary to add another element or com-
pound to act as the tracer.  The tracer method, of course, is quite
general, and therefore is not limited to use of sodium ion.  Any
convenient soluble ion may be used, and may be analyzed by any analy-
tical method available.

EXPERIMENTAL WORK
     The procedure described below was field tested at two locations.
The first was a large coal-fired power generating plant, which uti-
lized a turbulent contact absorber unit.  This unit was designed to
employ a limestone slurry as the scrubbing medium, but was temporarily
charged with sodium carbonate solution.  The stack was equipped with
a chevron type demister device, and an oil-fired flue gas reheater.
     The second installation was an oil-fired municipal power gene-
rating plant equipped with an inspiration scrubber and gull wing
demister.  The unit was being tested with sea water as the scrubber
solution, but was slated to eventually use a coral marl slurry.

-------
                                 3
     Sampling was accomplished by a  dry filtration technique similar
to that described in -he Federal  Register."'  Basically, the equip-
ment consists of probe,  filter,  impingers, pump, dry test meter,
pi tot tube, and inclined manometers.  Stack sampling procedures also
conformed to Federal Register requirements.  During the sample run,
the probe, cyclone,  and  filter holder were all maintained at 250°F.*
Two Millipore   type HAWP 14350  filters were used sequentially in
the same holder.   All  pertinent  instrumental parameters were recorded
during the test, and all  essential process data were obtained as soon
as practical.  For a listing  of  required  information, see Table I.
     After collection  of the  sample  and coolinq of the equipment, the
probe and cyclone were rinsed with distilled water, and the filters
were soaked overnight  in distilled water.  Volumes were determined by
graduated cylinders  for  each  of  the  resultant liquid samples, and they
were analyzed for sodium content by  atomic absorption spectroscopy.(2)
Three samples were produced in all:  liquid from impingers; wash water
from probe, cyclone  and  filters;  and blank distilled water.  Water soluble
sodium  blanks were  also determined  on unused filters.
CALCULATIONS
     Calculations were performed by  computer, but were based on equa-
tions given in the Federal Register.(1)   Input was flexible enough
 * Although EPA policy is to use metric units , certain non-metric units
 are used here for convenience.  The following factors  can be used to
 convert to metric units: m3 = 0.0283 x ft3; mm = 25. 4 x in.; m2 =
 0.0929 x ft2;  m3/min = 0.0283 x ft3/min; and °C = 5/9 (°F - 32).

-------
                               4
 Table I,    INPUT  AND  OUTPUT  VARIABLES  FOR'CALCULATION OF ENTRAPMENT

 Input
 VM - sample through dry  gas  meter, ft3
 PB - barometric pressure at  orifice, in.  Hg
 DH - pressure  drop across  orifice, in. water
 TMF - gas  meter temperature, °F
 T1F - flue gas temperature,  °F
 or
 BWO - volume fraction of water vapor in gas stream
 VIC - total  volume of liquid collected in impingers and silica gel, ml
 Ql - volumetric flow  rate  at stack conditions, ft3/min
 or
I OP - average pitot pressure  drop, in.  H20, and
IA - stack  area, ft2
 CI  - sodium concentration  in  impingers, mg/1
 VA  - volume of liquid  in  impingers after run, ml
 VB  - volume of liquid  in  impingers before run, ml
 CF  - sodium concentration  in  filter and probe wash, mg/1
 VF  - volume of filter  and  probe wash, ml
 CB  - sodium concentration  in  blank, mg/1
 CL  - sodium concentration  in  liquor to scrubbers, mg/1

 Output
 VMS - sample volume, STP  and  dry, ft3
 QS  - volumetric flow rate  in  stack, STP and dry, ft3/min
 NAI, NAF,  NAT - sodium in  impingers, filters and probe, and total, mg
 VL, VL6 -  entrained scrubber  liquid past demisters, in ml/min and
           gal/hr respectively

-------
to accept data in several  forms.  A  list of input and output variables,

along with their designations,  is given in Table II.  Equations used

are given in Table II,




          Table II.   EQUATIONS FOR  CALCULATION OF  ENTRAPMENT

                       DH
   VMS = 17.71VM (PB +13.6)
               TMF + 460


   BWO = 0.0474V1C
           VMS + 0.0474V1C
               *           ~
   either    VS= Kp Cp   DP  \    (PSXMS)
                = (85.48)   (.85)          /DP  (T1F + 460)
                     v/ZT             V     PB
                = 13.48    y/DP (T1F + 460)
                                  PB
   and
                                     /53Q	) /  PB   \
                QS = 60(1-BWO)(VS)(A)VT1F  +  460   )\ 2O2"/

                   = 1063(A)(VS)(PB)(1-BWO)  *  (T1F + 460)

   or           QS = 17.71Q1(PB)(1-BWO)  *  (T1F +  460)
      NAI = (CI(VA)  - CB(VB))  *  1000
      NAF = (CF(VF)  - CB(VF))  *  1000
      NAT = NAI + NAF
      VL = 1000 NAT  (QS)  *  (VMS(CL))
      VLG = 60VL t 3790
    *  VS,  Kp,  Cp,  TS,  PS, and MS are defined in Reference (1).

-------
                                 6
RESULTS
     Table in  gives  results in terms^ of both milliliters of entrained
liquid per minute and gallons of entrained liquid per hour.   These
entrainment values agreed within 10% of figures obtained by  plant
engineers from lengthy material balances and estimation.   The volu-
metric flow rate in the stack was also included since this important
value may be used directly to put the entrainment quantities into
perspective, or indirectly to calculate face velocities or other use-
ful operational variables.  Corrected sample volume and various  sodium
concentrations were also output from the equations used, but were not
general enough in interest to warrant their inclusion in the present
data table.  The list of output variables contains further information
on the result designations.

           Table III.  EXPERIMENTAL MIST ENTRAINMENT RESULTS
Test
SHI
KW1
KW2
KW3
KW4
KW5
KH6
VL,

1
4
2
3
26
ml/min
14
,200
,600
,000
,800
,000
17,000
VLG, gal/hr
0,22
19
73
31
60
400
260
qs,
21
37
57
26
35
57
56
ft /min
,000
,000
,000
,000
,000
,000
,000
PD, in.
-
8
8
8
12
12
12
H20








-------
                                  7
     The results from the second installation, all designated KW,
were further related to the pressure differential across the intake
tube for the scrubber, an important variable designated as PD.
When gal/hr entrained scrubber liquor was plotted against volumetric
flow rate in stacks, the curves shown in Figure 1 were obtained.
These are very similar in nature to curves obtained by the scrubber
manufacturer on pilot plant equipment.
     In contrast to the heavy entrainment levels found in the KW
series, the SH location showed only 14 ml of liquor loss per minute.
In view of the large flue gas and scrubber liquor volumes involved
in this operation, the low value indicated very efficient operation
of the demisters.  The figure is quite reasonable, however, since the
scrubber was temporarily without its "ping-pong ball" agitators, a
fairly low flow rate of flue gas prevailed, and operating conditions
during the test were very nearly ideal for effective operation of
the devices.
     An effect, which initially caused some concern, was observed at
the SH locations.  Even though the tracer method indicated that a total
of only 14 ml of mist per minute was carried up the stack (and therefore
that only about 0.1 ml should be collected by the sampling train during
the course of the test), 100 ml of water in addition to the initial
200 ml was recovered from the impingers.

-------
   500
 ra
 00

c/i"
OC.
400

-------
                                 9
     A series of calculations was made in order to account for this
difference, and the results are useful beyond this isolated sample.
They show the relative importance of various rcoisture contributions
to the flue gas, and may act as a guide when making material balances
or corrections such as those mentioned in the discussion section.
     The sources of water were: physical  entrainment of scrubber
liquor, trace amounts of water in the fuel oil to the reheater, com-
bustion products from the burning of the oil, evaporated scrubber
liquor, and water from the boiler gases.
     As mentioned previously, entrainment accounts for only 0.1 ml
of the total 100 ml; since typical water content of No. 2 fuel  oil  is
only a trace (0.1% is the maximum for ASTM requirements^) )t oniv
about 0.01 ml of the impinger catch would be likely to come directly
from the oil.  A typical  value for moisture formation from burning
No. 2 fuel oil  was obtained by means of tables and formulas from
"Combustion Engineering".^)  An estimate of the amount of water
contributed to the impingers by this mechanism was then easily cal-
culated.  Only 9 ml of the total 100 ml was computed to have come
from combustion of the fuel oil.
     Even though water is added separately by the boiler gases and
evaporation of scrubber liquor, the flue gas leaving the scrubber
will normally be saturated regardless of its moisture content  upon
leaving the boiler.  Therefore, it may often be convenient to  treat
the two sources as one.   In this case, for the sake of completeness,
the boiler gas moisture content was also calculated separately.
This moisture level is somewhat variable, but a typical value  is

-------
                               10
71g/m3 at a scrubber entrance temperature of 300°F.  Since it requires
only 60g/m3 of water to saturate the gas at its 110°F scrubber exit
temperature, the boiler gas typically provides slightly more than
enough for this purpose.  Boiler gas with moisture content  below
60g/m3 is  not especially unusual either, however, and in this case
the scrubber provides the extra water necessary for saturation.
     It. was reasonable to assume that the flue gas was saturated with
moisture at its exit temperature of 110°F.   Based on absolute humidity
figures from Lange,'^) it was calculated that it was possible to
condense 187 ml of water from the sample simply by cooling the satu-
rated gas from 110°F to 40°F.  It was readily apparent that this
boiler and scrubber humidification/condensation mechanism for con-
tribution of water to the impinqers was the dominant one, and far
outweighed the other three.
DISCUSSION
     Since the method described herein for determination of mist entrain-
ment is an indirect one, it might be beneficial to discuss more direct
methods.   The obvious approach to a problem, if workable at all, is
often the best.
     The most direct and obvious way to measure mist carry-over past
a demister device is to simply catch the liquid and measure it.   This,
however, requires a sample port Immediately after the demister,  and
that all  collection equipment be at the same temperature as the  stack
and demister.

-------
                                 11
     The location of the sample port would be especially critical,
in this case, since an attempt would be made to collect the entrained
mist droplets in a liquid state, and to avoid condensing and cap-
turing any of the humidity moisture at the same time.  Any heating
or cooling, even from passage through the stack, would cause evapo-
ration or condensation and would invalidate the test.  Evaporation
and condensation would not affect the tracer technique since the
quantity of sodium ion would be unchanged by these events.
     The direct procedure, like the tracer method and the semidirect
method discussed below, would normally require traversing the stack
and isokinetic collection to insure a representative sample.  In
the likely event that the flue gas temperatures at each of the tra-
verse points was not identical, it would be necessary to change the
temperature of all the sampling and collection equipment to be iso-
thermal with the gas at each point.  Worse yet, the liquid collected
at each point would have to be removed from the equipment before
the temperature change in order to avoid condensation or evaporation.
The result would be a number of liquid samples whose weight would
need to be combined to obtain the final result.
     Difficulties with this direct approach are readily apparent.
A sample port may not be available immediately after the demister,
and it may be physically or economically impossible to construct
one.  It would be especially difficult in most cases to match the
stack temperature throughout the entire sampling period.

-------
                                12
     Another possible route to entrainment determination is to
condense all water in the flue gas sample, either weigh or measure
volume, assume saturation, from the flue gas temperature calculate
the amount of water present due to saturation at that temperature,
subtract this amount from the total water condensed, and call  the
difference entrained mist.  A more refined calculation would need
to consider density differences between the saturation water and
the entrained liquor.  This procedure would be most effective if
sampling was immediately after the demister, but could be carried
out farther up the flue with decreased accuracy due to reheaters,
etc.
     Both of the above methods may probably be made to work in
certain specific cases.  In nearly all practical situations, however,
the tracer method will prove to be far easier, more economical, more
readily applicable, and probably more accurate.
     It should be mentioned that the filters used throughout the
present work were somewhat embrittled by heat and moisture combined.
This type of filter is composed of cellulose esters, which doubtless
undergo hydrolysis reactions under the conditions mentioned.  Filters
of the Millipore "Solvinert" series are said to be more stable, and
should be somewhat more satisfactory.  Glass fiber filter mats should
be satisfactory provided the sodium blank from these is consistent
enough.  Teflon filter mats are also available commercially, and might
prove useful for this application.

-------
                             13
CONCLUSION
     Based on the tests presented,  the method discussed  above
appears to be sound for the determination  of entrained scrubber
liquor carried past demisters.   All  results, both  experimental
and calculated, are reasonable  and  self-consistent.
REFERENCES

1.  EPA,  Standards of  Performance for New Stationary Sources, Federal
    Register 3£, No. 247,  Part  II, December 23, 1971.
2.  Perkin-Elmer Corporation, "Analytical Methods for Atomic Absorption
    Spectrophotometry," Norwalk, Connecticut, 1968.
3.  Fryling,  G. R., editor, "Combustion Engineering," Combustion
    Engineering, Inc., New York, 1966, pp.  14-14 and 21-13.
4.  Lange, N.  A.,  "Lange's Handbook of Chemistry," 10th Edition,
    McGraw Hill, New York, 1967, pp. 1424.

-------
                                    14
                                TECHNICAL REPORT DATA
                          (Plcate trail InMriictwiu on llw rrnrw lnjlrluixl
i  REPORT NO
 EPA-650/2-74-050
                                                       3 RECIPIENT'S ACCESSION-NO.
  TITLE AND SUBTITLE
 Measurement of Entrained Liquid Levels in Effluent
  Gases from Scrubber Demisters
            5. REPORT DATE
            June 1974
            6. PERFORMING ORGANIZATION CODE
7 AUTHORIS)
 L.D. Johnson and R. M.  Statnick
                                                       8. PERFORMING ORGANIZATION REPORT NO
9 PERFORMING ORGANIZATION NAME AND ADDRESS
 Process Measurements Section
 Particulate and Chemical Processes Branch
 Control Systems Laboratory
                                                       10. PROGRAM ELEMENT NO.
             1AB013; ROAP 21ACY-44
            11. CONTRACT/GRANT NO.
             NA
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 NERC-RTP,  Control Systems Laboratory
 Research Triangle Park, NC 27711
            13. TYPE OF REPORT AND PERIOD COVERED
            In-house; final	
            14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
1C. ABSTRACT
 The report gives results of the development and field-testing of a method for
 determining entrained liquid levels in gases leaving scrubber demisters, based
 on the use of a soluble ion as a tracer.  Trial measurements were made at two
 separate alkali scrubbing test facilities  located at power plant sites.  All results,
 both experimental and calculated, were  reasonable  and self-cons is tent. Agreement
 with independent estimates and practical experience was good.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                                           b.lDENTIFIERS/OPEN ENDED TERMS
                         c. COSATI I icId/Group
 Air Pollution
 Demisters
 Scrubbers
 Measurement
 Moisture  Content
 Entrainment
Air Pollution Control
Stationary Sources
13B
13A
07A
14B
07D
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (Tins Report)
Unclassified
                                                                    21 NO. OF PAGES
                                                                        18
 Unlimited
20 SECURITY CLASS (Thlipage)
Unclassified
                                                                    22 PRICE
EPA Form 2220-1 (9-73)

-------