Field Testing of Brownian Diffusion Fiberbed

Mist Eliminators for Sulphuric Acid Plant Service

By Jim Dougherty of IMC New Wales, Florida &

Steve Ziebold, Monsanto Enviro-Chem Systems Inc., St. Louis, Missouri

Abstract

Brownian diffusion fiberbed mist eliminators have been used in sulphuric acid plants for over 40
years. Often there is a need for some simple way to check elements in the field to
ensure acceptable pressure drop and mist removal performance. A portable field test apparatus
was developed to measure bed resistance on-site for evaluation and comparison to expected
manufacturing levels. Often during maintenance periods, especially when elements have been
washed to remove plugging agents, there is a need to check elements to predict if they will
operate acceptably. This paper describes the portable field test apparatus, set-up and operation.
Field testing of different fiber bed packing types are compared along with an extended surface
area concentric fiberbed mist eliminator. In addition, the value of process gas pressure drop in
sulphuric acid plants is described.

Background

IMC Phosphates Inc. operates (5) 2900 STPD Monsanto sulphuric acid plants at their New Wales
Fertilizer Complex in Central Florida, USA. Six months following a routine turnaround on the
05 Plant, pressure drop in excess of 20" H20 was experienced in the mist eliminators of the inter-
pass absorption tower. An emergency outage was taken to replace the damaged mist eliminators,
which appeared to be fouled with carbon soot from a faulty oil fire.

A spare set of mist eliminators

was not available, so elements that Collapsed elements in a sulphuric acid plant
had been recovered from a (View under tubesheet)
mothballed plant at another IMC
location, washed, and stored were
installed into the 05 Plant inter-
pass tower.

Initially, pressure drop was a

normal 10", but eventually the
pressure drop built up to over 20"
and a second emergency outage
was required less than 18-months
later to install an entire new set of
mist eliminator elements.

1
 As a result of these events, IMC began discussions with Monsanto regarding methods to better
evaluate the condition of mist eliminators for re-use or re-wrap. The culmination of these
discussions was the field testing of mist eliminators elements during the February 2004 Plant
turnaround using Monsanto's portable element pressure drop testing equipment.

Description of Portable Field Element Pressure Drop Test Apparatus

The pressure drop test unit is designed to measure the flow resistance of individual mist
eliminator elements at customer's sites after the mist eliminator elements have been removed
from the vessel and cleaned. The unit is mounted on a skid that is 43 inches wide by 154 inches
long and weighs approximately 600 lbs. The unit consists of a 10 HP carbon steel blower, a
starter, a 12 foot length of 8 inch diameter CPVC pipe, a 10’ length of flexible duct, pressure
gauges, and adapter flanges. Mounted in the CPVC pipe is a calibrated orifice. Figure 1 shows a
diagram of the equipment arrangement.

The unit's flexible duct and the adapter flanges enable it to be connected to various diffusion style
candle elements. The elements are tested in the horizontal position for convenience and safety.

There are three pressure gauges mounted on a stand. One pressure gauge is provided to measure
pressure drop across the element. The second gauge measures the pressure drop across the
calibrated orifice and the third gauge measures the static pressure upstream of the flow orifice.

Checking the pressure drop of a fiber bed mist eliminator at a known flow rate is an excellent
troubleshooting tool. Comparing the actual pressure drop to the expected pressure drop allows
customers to identify elements that are plugged, (high pressure drop), or that have lost a
substantial amount of fiber due to corrosion, (low pressure drop). Customers who occasionally
clean their elements can evaluate the effectiveness of a washing procedure by checking flow
resistance after washing and drying elements and comparing these values to manufacturing data.

Figure 1 - Portable Element Pressure Drop Test Equipment Arrangement

BLOWER
STATIC
GAUGE
1 ELEMENT DPI
DPI GAUGE
GAUGE(S)

H L

H L
INLET FLOW
CONTROL
DAMPER FLANGE
ORIFICE ADAPTOR

[Caution – Never test elements that have not been properly cleaned and neutralized and
always review element handling procedures with regard to safety and personal protective
equipment required.]

2
Operation of Portable Field Element Pressure Drop Apparatus

In order to operate the portable field element pressure drop apparatus, proper connections of
impulse lines to the pressure gauges are first verified. The pressure gauges are then calibrated
and set to zero "0" with the blower off. Next the impulse lines are checked to be clean and free of
any liquid.

Portable Field Pressure Drop Test Skid with Operator

After preliminary checks and adjustments are made, the blower is started with the inlet damper
closed. The damper is then slowly opened and adjusted as required. The flow orifice pressure
drop and element pressure are then carefully monitored. These readings are allowed to come to
equilibrium (readings must stay constant over time). If the element is very wet, it will take some
time to get constant readings. If necessary, the blower damper is adjusted to obtain the desired
flow orifice setting. When the three pressure gauges are steady and reading constant, the data is
recorded in the log sheet.

The flow rate is first calculated as dry standard cubic feet per minute (DSCFM) from the equation
on the log sheet. Alternately the calculations can be converted to measure normal volume rate
(Nm3/h) for equivalent metric units. Next the standard flow rate is converted to actual flow rate
using a standard orifice equation as shown on the log sheet. (appendix A example).

The superficial log mean element flow area for cylindrical fiberbeds is calculated as:

Log Mean Element Flow Area = Hπ(OD-ID)/ln(OD/ID)

Where,
π = 3. 14159
OD = outer element diameter
ID= inner element diameter
H= effective element height

3
Bed velocity is then calculated by dividing the actual volumetric flow rate by the superficial log
mean element flow area.

Brownian diffusion fiberbed mist eliminators are essentially laminar flow devices in the regime
where they normally operate. Thus for a given type of bed construction, a dry bed pressure drop
model is used to calculate bed resistance. In the streamline region, based on D’Arcy’s law, the
Kozeny-Carman relationship is often used:

Pd = kµg L(1-ȏ)2/(df 2ȏ3)V

Where, k = constant
g = gas viscosity
L = Bed thickness
ȏ = void fraction 
df = fiber diameter
V = Bed velocity

Based on the specific filter construction, the above model can be modified to best fit the data
based on intrinsic fiber bed filter properties. At ambient temperature with air viscosity assumed
constant, the above model simplifies to:


Pd = CoV, where Co is a constant for a given diffusion bed type.

Co is sometimes called the dry element “pressure drop coefficient”. From electrical analogy, Co
can be thought of as resistance to flow or “bed resistance”. Pressure drop is analogous to voltage
and bed velocity (flow per unit area) is analogous to current.

Field pressure drop tester in operation (element in foreground)

4
Thus bed resistance or “pressure drop coefficient” (Co) is calculated by dividing the element
pressure drop by the bed velocity:

Co = ∆Pd/V

The pressure drop coefficient is then compared to the expected value to determine whether the
measurement is unusually high or low. If the measurement is out of line with expectations, then
further evaluation may be required.

Field Test Results at IMC New Wales, Florida

Field tests using the portable pressure drop tester were carried out at IMC New Wales, Florida in
early 2004. Three types of diffusion fiber beds manufactured with glass roving fiber were
evaluated: angle wrapped, parallel wrapped and angle wrapped concentric. A more detailed
description of Brownian diffusion mist eliminators is described later in this paper.

The three elements were two feet in diameter by ten feet long. One of these test elements is
shown in Figure 4. Test gas flow for all beds were normalized to 2500 cfm to simulate
approximate design flow at dry conditions. The results are shown below in Table 1:

Table 1: Field Pressure Drop Test Results on Diffusion Roving Beds

Angle Parallel Angle Optimized
Wrapped Wrapped Wrapped Angle Wrapped
Roving Roving Roving Roving
Element Element Concentric Concentric
Element Element
Dry Pressure 5.3 7.1 4.1 3.9
Drop (in. w.c.)
~Estimated Wet
Pressure Drop 10.6 14.2 8.2 7.4
(in. w.c.)

Estimated wet saturated operating pressure drop at process conditions is also shown in the table.
This value is about twice the dry pressure drop.

As result of the field testing, the angle wrapped roving concentric element design was optimized.
Thus at process operating conditions this element is expected to operate at about 48 percent less
pressure drop compared to the parallel wrapped roving bed. This is equivalent to a reduction of
about 6.8”wc operating pressure drop in a 2500 STPD sulphuric acid plant. The optimized angle
wrapped roving concentric element provides the extra pressure drop savings due to increased
surface area and the angle roving fiber arrangement.

5
Value of Gas Pressure Drop in a Sulphuric Acid Plant

Plant Blower Rate Limited

Assuming plant production rate is only blower limited, it is desirable to determine how much
extra production capacity can be achieved by reducing pressure drop of an individual equipment
item. In order to do this the new intersection of the blower curve with the total equipment gas
resistance curve can to be determined and compared to original design rate. Note in this example
production rate and gas flow rate are assumed to be proportional.

Figure 2 shows a typical blower curve. Superimposed are typical total equipment gas resistance
curves. Design blower static pressure and flow rate are designated as Po and Ro respectively.
This point is at the intersection of the blower curve and the original total equipment gas resistance
curve (solid curves). When the pressure drop of an equipment item is reduced, the equipment
gas resistance curve shifts downwards as shown (dashed curve). A new equilibrium point is
achieved where this curve intersects the blower curve e.g., at point designated by P and R.

Figure 2: Example Blower & Equipment Gas Resistance Curves

Blower Static Pressure

Po
P dP

Plant Rate Ro R

The blower curve over a wide range can be modeled by using a polynomial expression such as:

Pressure = A + BR+CR2+DR3

However, for a small change in operating pressure, the blower curve may be assumed linear in the
region of the design blower operating point:

P = Po + K(R-Ro)

Where K is the slope of the blower curve at the design point.

6
The total equipment gas resistance curve follows the approximate relationship:

P = AR1.7

Where A = constant.

Therefore, it can be shown the new total gas resistance curve can be estimated by:

P=(Po-dP)(R/Ro)1.7

Where dP = the reduction in pressure drop at original design rate by substituting a new individual
equipment item (designed for a lower operating pressure drop).

Equating this relationship with the above linear blower curve equation results in the following
expression:

Po + K(R-Ro)=(Po-dP)(R/Ro)1.7

Thus by knowing the slope of the blower curve at the original design operating point, the above
equation can be used to solve for the new rate R by iteration.

As a rule-of-thumb, on a basis of 180 in. w.c. blower discharge pressure, a 5 in. w.c. pressure
drop reduction equates to about a 1 percent increase in plant rate. Thus a 2500 STPD plant could
increase sulphuric acid production by 25 STPD. The higher pressure steam would go up
proportionately, less some increased steam to power the blower turbine drive.

Thus, comparing field pressure drop measurements described earlier, using the tested angle
roving elements reduces operating pressure drop by about 3.6 in. w.c. compared to using the
tested parallel roving beds. If a 2500 stpd plant is blower limited, the reduced pressure drop can
be utilized to increase sulphuric acid production rate by about 18 STPD. Currently this represents
an increase in value of over $100,000 USD per year since additional acid production is valued at
the incremental cost of purchased acid less sulphur cost plus the opportunity cost of co-
generation.

Using the new extended area optimized angle roving concentric elements reduces operating
pressure drop an additional 3.2 in. w.c. compared to using conventional angle roving elements.
Again if a 2500 stpd plant is blower limited, the reduced pressure drop can be utilized to increase
sulphuric acid production an additional 16 STPD. This represents an opportunity for an additional
increase in value of over $100,000 USD per year.

Plant Non-Rate Limited

If the plant has excess capacity, then the value of reducing operating pressure drop is related to
reducing the cost of operating the plant blower. In the case of an electric drive the calculation is
straightforward based on the value of reducing electrical usage. If the blower is steam driven
however, a reduction in operating pressure drop relates to the value of the high pressure steam
savings that would normally be used to add incremental pressure to the blower. The value of the
high pressure steam depends on whether or not there is a turbogenerator on-site with excess
capacity to produce incremental electric power.

7
Annual energy cost for any equipment item is calculated as if it were the last increment of
pressure being handled by the main blower. The energy cost attributed to adding incremental
pressure (Pb) to a blower is given by the following equation:

Eb = 4.311((STPD)FsTscd/(ysNckf))((Pd/Ps)^kf-((Pd-Pb)/Ps)^kf)

Where,

Eb = Equipment annual energy cost for adding incremental pressure ($/yr)

STPD = short tones per day
Fs = Fractional on-stream time
Ts = Compressor suction temperature (Deg-Rankine)
cd = Driver power cost ($/kwh)
ys = Volume fraction equivalent SO2 to acid plant
Nc = Compressor fractional efficiency (isentropic)
kf = power factor = (k-1)/k
k = Specific heat ratio (1.4 for air)
Pd = Absolute pressure at compressor discharge (in. w.c.)
Ps = Absolute pressure at compressor suction (in. w.c.)
Pb = Incremental equipment pressure drop (in. w.c.)

When the main blower handles gas containing sulfur dioxide then substituting the k value and
compressor pressures can solve the above equation.

When the blower handles air (k = 1.4), the above equation reduces to:

Eb = (STPD) FsTscd FpPb Ps -0.286/( ysNc)

Where the pressure factor (Fp) could be set up on a graph or can be conveniently calculated as
follows:

Fp = yPdz

And,

y = -9.8186E-08Pb3+9.1992E-06Pb2+3.9122E-05Pb+0.14992

z = 1.0554E-07Pb3-9.4143E-06Pb2+1.2372E-05Pb-0.22956

for the range 160<Pd<240.

For an electric motor, the driver power cost is given by:

cd = ce/Nd

Where,

ce = Electric power value ($/kwh)

Nd = Driver fractional overall efficiency

8
For a steam turbine blower drive:

cd = cs(TSR)/Nd

Where,

cs = Change in value of steam passing through turbine ($/lb)

TSR = Turbine theoretical steam rate (lb/kwh)

Assuming the following typical sulfur burning plant conditions:

Fs = Fractional on-stream time = 0.98

Ts = Compressor suction temperature, 550 Deg-Rankine
ys = Volume fraction equivalent SO2 to acid plant = 0.115
Nc = Compressor fractional efficiency (isentropic) = 0.76
Ps = Absolute pressure at compressor suction = 402 in. w.c.

The energy equation can be simplified to:

Eb = 1077(STPD)cdFpPb


  
    

Brownian Diffusion Fiberbeds

Dr. Joe Brink developed the first generation

Brownian diffusion fiberbed mist eliminator
in the 1950’s for chemical plant service.
This style element is basically a construction
of concentric wire mesh screens filled in-
between with hand packed bulk fine fiber
media. The element is able to achieve high
efficiencies ranging from 90 to 99.9+%
depending upon the choice of pack and
application in which the elements are
installed. In the case of strong sulphuric
acid applications, chemically resistant glass
fiber is used as the collecting media.

One drawback with bulk packed fiber beds is packing uniformity. The packing may have
significant variance by inexperienced packers or if there is variance with the raw fiber supply.
Note that the field pressure test described earlier does not detect anomalies within the packing.
The field pressure test only measures average bed resistance.

9
Figure 3 shows an example thermograph profile
for an unacceptable bulk packed mist eliminator
 
     
   
element. As warm air is forced into the filter, an
infrared thermograph provides a qualitative view *>91.7°F
of packing uniformity as it affects gas flow
distribution. As can be seen, the example test
element has a high gas flow variance especially 90.0
at one end. 88.0
86.0
Another issue with bulk fiber elements is there
are limits with operating velocity and mist 84.0
loading as related to potential entrainment 82.0
generation. Also when these elements need to 80.0
be repacked, the fiber packing is difficult to
remove unless a special procedure is used 78.0
separating inner and outer cages.
*<76.4°F

:;<    89



:
 ;
< 

!"#$%&
" '
# ()*
+%-#3-)1%2+
% &#'# , #%
45-+. $/),
5 2 01
"-#($..
+%+&4 + #(4+5#65!2+.(#.1'+
. -#2
--.+%$
,$/ +
. "+ "'()*
/)%. #% & 71% )(, #( # 1/6 0()2
The angle roving Brownian diffusion fiber bed mist eliminator is the second generation
developed for chemical plant service. The element is a wound roving fiber structure resulting in a
more uniform packing distribution. This greater uniformity can be qualitatively demonstrated in
the example thermograph in Figure 4. Computer controlled roving fiber feed and pressure
monitoring during manufacture assures exactly matched elements in order to optimize
performance in multi-element installations.

10
Figure 5 shows a close-up thermograph of a parallel roving wrapped diffusion fiber bed. This
style of packing is susceptible to high velocity areas in between glass rovings where parallel
alignment creates low resistance paths for gas flow as can be seen. Note the portable field
pressure drop tester cannot detect similar localized high velocity areas in a fiberbed. Thus it is
important to select a fiberbed design that assures consistent uniform gas distribution throughout
the packing.

?@=B>C
9>
 ;<D  E
Thermograph analysis is an excellent
qualtitative tool to quickly locate “weak”
spots in large fiber beds as shown.
A   
However, a more effective tool used to *>89.1°F
quantitatively measure packing variances
is a velometer probe (shown in Figure 5). 88.0

86.0
The velometer probe is used to measure
84.0
local gas velocities along the length of the
element. A velocity profile was measured 82.0

for the parallel roving element in Figure 5. 80.0

This velocity profile is shown in Figure 6 78.0
and verifies high velocity areas where
76.0
parallel rovings happen to align.
74.0

72.0

*<71.7°F

Figure 6 - Parallel Roving Element Gas Flow Variance

Local Surface Velocity (FPM)

600
3990 FPM
500

400

300

200

100

0
0 10 20 30 40 50 60 70 80 90 100 110 120
Distance Along Element (inches)

11
For many applications, angle roving diffusion fiberbeds typically provide up to 20 percent more
gas throughput compared to bulk packed fiberbeds for the same operating pressure drop and
collection efficiency (or equivalent reduction in pressure drop for the same number of elements).
This is due to more uniform packing distribution along with the unique wetting properties of the
roving fiber. The collecting fiber media is an angle wrapped interlocking roving and is more
stable and uniform compared to parallel wrapped roving, bulk packed or donut style packing.
This results in more predictable results when using the portable field tester described earlier.

In conventional diffusion fiber beds that only utilize a single collecting layer, part of the collected
liquid draining on downstream discharge gas surfaces often reentrains in the exiting clean gas.
This reentrainment increases with higher gas flows and inlet mist loadings and can result in
significant damage to downstream equipment or contribute to overall emissions. A standard
feature of the angle wrapped diffusion fiberbed is the bi-component fiber bed design to eliminate
reentrainment. The bi-component fiberbed design includes a proprietary drainage layer of coarse
fibers oriented downstream of the fine fiber collecting layer. Liquid films that would normally
form and burst at the gas discharge surface of the fine fiber collecting layer are drawn into the
drainage layer and drained by gravity. This allows the element to operate at higher bed velocities
and mist loadings compared to fiberbeds that only utilize a single collecting layer.

Compact Element Designs

One of the elements described earlier in the field measurements was a concentric element. As
plants continually find ways to increase capacity or “de-bottleneck”, one of these methods is to
replace conventional mist eliminators with what is an element inside an element or the
“concentric” element” design. A concentric element has an element in the normally void volume
in the center of the mist eliminator. The internal element is in opposite orientation as the main
mist eliminator (See Figure 7). Thus the gas splits between inner and outer element as it is
filtered.

Figure 7 - Concentric Element Examples

Hanging Style Standing Style

There are distinct advantages to this approach because use of concentric beds increases bed area
while using the same conventional bolting making retrofitting possible. The use of concentric
elements can lead to lower pressure drop by as much as 30 to 40 percent depending upon element
style and application. Reducing element pressure drop can sometimes allow for increased gas
flow and/or plant capacity.

12
As with anything else, there are tradeoffs with the concentric design. The elements are heavier,
so in some cases the heavier element weight may require additional tubesheet reinforcement.
Also this additional weight makes it somewhat harder for maintenance crews to handle and install
the elements so extra rigging may be required to assure proper attachment. In the case of
standing style, a seal leg is required for the internal forward flow element that may need to be re-
routed around exiting equipment or tower internals below the tubesheet.

Conclusions

1) A portable field test device has been developed and used at several plant locations as a
troubleshooting tool to measure flow resistances of Brownian diffusion fiber beds. This
test is especially useful to evaluate the flow resistance of used elements that have been
removed from service, washed and dried.

2) Because the portable field tester does not detect local high velocities in the fiber bed due
to packing anomalies, it is important to select an element that has proven uniform
packing distribution such as the angle wrapped diffusion roving fiberbed.

3) Based on a blower rate limited 2500 STPD sulphur burning acid plant in southeastern
United States, the tested conventional angle wrapped roving fiberbeds were estimated to
provide an incremental value of over $100,000 USD per year compared to using the
tested parallel wrapped roving fiberbeds.

4) Based on a blower rate limited 2500 STPD sulphur burning acid plant in southeastern
United States, optimized angle wrapped roving concentric fiberbeds were estimated to
provide an additional incremental value of over $100,000 USD per year compared to
using conventional angle wrapped roving fiberbeds.

5) Likewise for the same plant in 3) & 4) above, optimized angle wrapped roving concentric
fiberbeds were estimated to provide an additional incremental value of over $200,000
USD per year compared to using the tested parallel wrapped roving fiberbeds.

James Dougherty: Employed in several positions in plant operations and process engineering in Chlorine, Sulphuric
Acid, and Phosphate Fertilizer plants for Olin Corporation, Agrico Chemical Co., General Chemical Co., and IMC
Fertilizer over the past 16 years. Earned a BS degree in Chemical Engineering from Clarkson University and a MBA
degree from University at Buffalo. Currently working as a Process Engineer for IMC Phosphates at IMC’s New Wales
Fertilizer plant in Central Florida.

Steve Ziebold: Manager Technical Services for Brink® Mist Eliminators at Monsanto Enviro-Chem, has over 31
years experience in research, design, installation and start-up of fiber bed mist eliminator systems. He co-invented the
bi-component fiber bed mist eliminator and has several patents in mist eliminator and chemical plant designs.

13
Appendix A

PRESSURE DROP TESTER LOG SHEET EXAMPLE

Date
Element Identification Note
Orifice Pressure Drop (dp orifice), “WC
Static Pressure (SP), “WCg
Element Pressure Drop (dp element), “WC
Temperature (T) Deg F
Barometric Pressure (BP), “Hg
Flow Rate, DSCFM
Flow Rate, ACFM
Element Surface Area (Ab), FT2
Pressure Coefficient, Co

DSCFM = K dPorifice
460 + T 29 406.91
ACFM = DSCFM * * *
530 MW ( BP * 13.6 + SP)
dPElement * Ab
Co =
ACFM

Where:
DSCFM= Flow rate in Dry Standard Cubic Feet per Minute
K= Orifice Coefficient
ACFM= Flow rate in Actual Cubic Feet per Minute
T= Temperature, Degrees F
MW = Molecular Weight of Gas (in units of lb/lb-mole, if not available use 29)
BP= Barometric Pressure (in inches of mercury, "Hg)
SP= Static pressure (in inches of water- "w.c.)
Co= Bed Resistance of Fiberbed (units of " w.c./ft/min)
Ab= Log Mean Bed Area
dPelement= Pressure Drop of element at a given flow rate

14