Prototype Firing Range Air Cleaning System
GLISSMEYER et al
Proceedings of the 18th DOE Nuclear Airborne Waste
Management and Air Cleaning Conference.
Held in Baltimore, Maryland 12-16
August, 1984.
Sponsors: U.S. Department of Energy and The Harvard Air Cleaning
Laboratory.
Editor, Melvin W. First.
Published March, 1985
with an INDEX to the
17th and 18th CONFERENCES.
CON F-840806 Volume 2
PROTOTYPE FIRING RANGE AIR CLEANING SYSTEM
J. A. Glissmeyer, J. Mishima, and J. A. Bamberger
Pacific Northwest Laboratory
Richland, Washington
Abstract
|
Also
see: GLISSMEYER et al |
The Ballistics Research Laboratory, a component of the U.S. Army Research and Development Command, contracted with Pacific Northwest Laboratory (PNL) to provide a prototype air cleaning system for a new large caliber firing range where depleted uranium munitions are test-fired. The existing system consists of two banks of prefilters and a bank of HEPA filters in series at a rated airflow of 24,000 acfm. Experience at similar ranges indicated that the existing filtration system would be too costly to operate because shock waves and rapid particle loading result in short filter life necessitating frequent replacement and disposal as low-level radioactive waste. The rapid particle loading also results in decreased airflow causing an excessive waiting period before personnel can reenter the target area. The project's objectives were to provide a prototype air cleaning system that would substantially reduce operating expense, shorten the delay for personnel reentry and enhance the particle removal performance.
PNL's study proceeded by examining the characteristics of the aerosol challenge to the filtration system and the operating experience at similar firing ranges. Candidate filtration systems were proposed; including baghouses, cartridge houses, electrostatic precipitators, cleanable high efficiency filters, rolling filters and cyclones-- each followed by one or more of the existing filter banks. Methodology was developed to estimate the operating costs of the candidate systems. Costs addressed included the frequency (based on fractional efficiency and loading data) and cost of media replacement, capital investment, maintenance, waste disposal and electrical power consumption. The recommended system will be installed during calendar year 1984.
Introduction
The U. S. Army Materiel Test Directorate (MTD) and the Ballistics Research Laboratory (BRL) both operate two firing ranges (Ranges A, B, and C, D respectively) for the testing of large caliber depleted uranium (DU) penetrators. The targets are housed in enclosures which contain the DU aerosols and fragments produced by the test firings. One of the drawbacks of using a target enclosure is that the airborne DU must be removed by ventilation and air cleaning before personnel can enter the enclosure without respiratory protection.
BRL recently completed construction of Range D, shown in Figure 1. The Army's experience with air cleaning systems at such ranges has shown that the operation of the systems is very costly because of the short life of the filters. Our objective was to recommend an air cleaning system that would meet the requirements for effluent discharge and have substantially lower operating costs.
Figure 1. Range D
This paper describes the air cleaning challenge with a brief review of the performance of the air cleaning systems at the earlier ranges and the characteristics of the airborne DU aerosol. The candidates for the prototype air cleaning system are then outlined and the economic evaluation of them is detailed. Finally, the recommended systems and configurations are discussed.
Challenge To The Air Cleaning System
Prior to selecting candidate air cleaning systems to replace or augment the existing system, we investigated the performance of the existing cleaning systems at the ranges described below. In particular, we focused on the filter life, pressure pulses and measurements of aerosol characteristics.
System Descriptions
The existing Range D filtration unit, sized for an airflow of 24,000 cubic feet per minute (cfm), consists of three banks of disposable filters in series. There are twenty-four filters in each bank -- three filters high by eight filters wide. The filters are enclosed in a housing as shown in Figure 2. The double width-double inlet fan, with 60 hp motor, is rated at 24,000 cfm at 11 in. H2O and is also enclosed in the housing. Electrically driven opposed-blade dampers are located at the inlet and outlet of the housing to provide some measure of isolation when the system is shutdown. The dampers drive to their full-open position when the fan is started. The housing has a 12 gauge wall thickness and the roof has a 20 gauge outside liner.
The first filter bank consists of prefilters which have pleated cotton media with an ASHRAE 52-76 dust spot rating of 25%. The dimensions of a single filter are 24 in. x 24 in. x 2 in. The second filter bank contains filters which have pleated glass-fiber media reinforced with hardware cloth and cardboard separators. The filter rating is 95% ASHRAE. The frame enclosing the media is galvanized steel with dimensions of 24 in. x 24 in. x 12 in. The first two banks of filters are mounted in commercial side-loading housings. The third bank consists of HEPA filters with galvanized steel frames and aluminum separators. The HEPA filter mounting frame permits individual installation and clamping of the filters with filter gaskets on the downstream side of the filters to maintain the seal during a pressure pulse. The filter housing was not designed according to the typical standards for nuclear facilities.
Figure 2. Existing Range D Filter House
Ranges A, B, and C are functionally the same as Range D, shown in Figure 1, with a projectile entrance tunnel, target bay and air cleaning system; however, there are many differences in design and they are smaller overall. The target bays are box shaped and enclose 15,000 to 28,000 cu. ft as opposed to 54,000 cu. ft at Range D. The ventilation rates range from 6,000 to 18,000 cfm and the projectile entry tunnels are cylinders 6-7 ft in diameter and 30 to 150 ft long. The filter houses contain three banks of filters in series. The last two filter banks are of the same type as Range D (95% ASHRAE and HEPA's). At Range C the first prefilter is the same type as Range D (25% ASHRAE); however, 40-60% ASHRAE prefilters are used at Ranges A and B. At Ranges A and B, the filter house is located behind the target bay and connected to it by box shaped baffled shock attenuators. At Range C, the filter house is located on the roof of the target bay and connected to it with a duct containing fragment baffles but no shock attenuator. Ranges A and C use industrial grade
side-loading filter housings of about 16 gauge sheet metal. The housing at Range B is shop-built to nuclear facility standards using 0.5 in. steel, marine bulkhead access doors, bagout ports, and face loading filters.
Filter Life
Operating data from Ranges A and C were compiled and the filter life expectancies are shown in Table 1. Different filter change philosophies existed at the two ranges. At Range A, the two pre-filter banks and the HEPA bank were changed at pressure drops of approximately 1, 4 and 4 in. H2O respectively. At Range C the second prefilter and the HEPA were changed at about 4 and 3 in. H2O. The first prefilter bank did not have a pressure gauge and was changed when bank 2 was being changed.
Table 1. Filter Life
Number of Range of Filter Life, . Shots Filters Per Rounds/Change . Range Per Year Filter Bank 1st-, 2nd-Prefilters HEPAs . A 230 12 1-2 10-20 100-200 B 230 18 No data available, life estimated to be similar to Range A C 180 6 1-6 1-6 5-60 D 180 24 No data available, life estimated to be similar to Range C
Filter damage at Ranges A and B has been mostly limited to condensation and subsequent moisture damage to the filter media. At Range A, inadequate sheet metal seals in the filter housing permitted water entry. On one occasion, excessive plugging of the media caused the 1st prefilters to impact the 2nd prefilters and break their seals. At Range B, condensation was caused by temperature gradients in the system.
Temperature measurements at Range C showed that a heat front passes through the exhaust system approximately 1 second after the pressure shock. The heat front has caused ignition of the prefilter banks for some target/penetrator combinations, destroying the media and leaving only the wire support mesh and the cardboard frames. In one instance, sparks from the prefilters ignited the HEPA filters and melted the media. Since this incident, temperature sensors were installed in the exhaust duct and temperatures greater than 212°F have been recorded. It is unknown whether hot target or penetrator fragments had a role to play in these incidents. The absence of a shock attenuator at this range may be the reason that such incidents have only occurred at this range.
Pressure Pulses
Measurements of pressure pulses generated by static charges and the target/penetrator interactions were made at the ranges prior to the start of routine testing to verify the structural integrity of the target bay and the air cleaning system. During these tests at Ranges A and B no damage to the filters was observed by either visual inspection or standard aerosol penetration tests. At Range C, damage was caused by the temperature effects discussed above.
Range A
Blast overpressure measurements in the target bay and air cleaning system for five test firings are listed in Table 2. The pressures measured at the face of the prefilter banks are listed under Locations 9 and 10. Location 11 was downstream of the HEPA filters. In the filter house, the pressures pulses were of considerably lower magnitude and longer duration than in the target bay.
Range B
An extensive set of static and dynamic tests were performed at Range B. Table 3 gives the results of twelve static charge tests. The highest pressures observed at the two prefilter banks and the HEPA bank were 1.2, 0.6 and 0.7 psig respectively. Pulse durations were not available.
Table 4 presents the dynamic overpressure data for seventeen test shots with two penetrator and three target configurations. The filter bank pressures were significantly lower than in the target enclosure due to the action of the shock attenuator. In these tests the maximum pressures at the two prefilter banks and the HEPA bank were 0.89, 0.61 and 0.59 psig respectively. Maximum reflected pressures would be double these values.
Range C
Table 5 lists representative values of the pressure pulse and duration for the Range C static and dynamic tests. Peak overpressures of 4 psig were measured in the target enclosure. In front of the prefilters the average pressure dropped to 0.6 psig.
Range D
Range D has only been tested with static charges to date. The measured face-on pressures at the first prefilter have ranged from 1.05 to 2.35 psig. No damage was observed to the filters but the wall joints of the housing were severely damaged. No follow-up standard aerosol penetration testing has been performed yet.
Table 2. Range A Pressure Measurements in Exhaust Duct System Using a 3.4 kg DU Round
Pressure Transducer Location . Round Over Pressure-psi (Duration - ms) . Designation 1 2 5 6 8 9 10 11 14 1.0(5) 0.9(4) 0.6(1) - 0.3(499) 0.3(456) 0.2(479) 0.2(474) 15 1.8(4) 2.2(4) 8.1(2) - 0.6(535) 0.5(464) 0.4(477) 0.3(407) 16 1.9(4) 2.2(4) 1.4(9) 0.6(11) 0.4(521) 0.4(467) 0.3(479) 0.2(520) 17 1.8(2) 2.4(5) 5.1(9) 0.5(35) 0.4(518) 0.5(470) 0.4(480) - 18 8.9(3) - 4.0(5) 2.2(10) - - - - Pressure Transducer Location 1 Right wall, 4.5 ft from floor, 15.5 ft from rear wall, target enclosure 2 Right wall, 4.5 ft from floor, 17.5 ft from rear wall, target enclosure 5 Left wall, 5 ft from floor, 18 ft from rear wall, target enclosure 6 On air intake louvres, target side, 5 ft from floor 8 Behind baffles in exhaust plenum 9 1st filter, upstream side 10 2nd filter, upstream side 11 Fan compartment
Table 4. Range B Dynamic Tests: Peak Overpressure Data
Target 1, Target 2, Target 1, Penetrator A Penetrator A Penetrator B . Round Number 1 2 3 4 5 6 7 8 9 Sidewall, PR, psi 5.3 5.2 -- 7.4 17 23 4.7 4.9 -- Overhead, PR, psi 2.7 2.9 4.2 4.8 6.8 5.5 4.4 -- 11 Plenum, PS, psi 0.41 0.56 0.64 0.91 0.84 0.86 0.83 0.68 0.66 Filters, PS, psi First 0.16 0.24 -- 0.39 0.32 0.37 -- 0.36 0.32 Second 0.11 0.20 0.20 0.32 0.24 -- 0.18 0.30 0.26 HEPA 0.22 0.25 0.30 0.42 0.46 0.48 0.27 0.31 0.19 Fans, PS, psi 0.10 0.19 0.17 0.24 0.20 0.35 0.17 0.26 0.24 Target 2, Target 3, Target 2, Penetrator B Penetrator B Penetrator B . Round Number 10 11 13 14 15 16 17 18 Sidewall, PR, psi -- 21 4.0 3.1 4.7 11 -- 5.5 Overhead, PR, psi 3.7 -- 2.8 2.6 2.7 6.2 5.0 6.3 Plenum, PS, psi 1.3 1.2 0.21 0.22 0.28 1.1 1.3 1.1 Filters, PS, psi First 0.52 0.65 0.14 0.15 0.17 0.59 0.89 0.76 Second 0.42 0.54 0.10 0.12 0.13 0.50 0.61 0.53 HEPA 0.25 0.59 0.10 0.12 0.12 0.50 0.53 0.54 Fan, PS, psi 0.17 0.46 0.09 0.10 0.10 0.40 0.41 0.41 PR: Face-on, normally reflected pressure For the plenum, filter and fan data PR is double the PS value. PS: Side-on pressure
Table 5. Summary of Range C: Blast Overpressure Data
Pressure Gauge Location . Inside Ventilation Duct Center of Roof Between Baffles and Prefilters 21.7 ft Above Target Pressure . Pressure Duration Average Peak Duration Test (psi) (ms) (psi) (psi) (ms) 0.3.4 kg Pentolite 4.1 2.5 0.17 0.87 1300 (static) 2.1 kg Tungsten 1.7 3.0 0.07 0.29 1000 2.0 kg DU 4.1 3.0 0.29 0.58 2200 3.4 kg DU 4.4(a) -- 0.59 0.87 3000 (a) Maximum pressure recorded for a 3.4 kg round.
Aerosol Characteristics
Quantity and Composition of the Airborne Material
The impact of the DU penetrator upon the target results in aerosolization of a portion of both the penetrator and target. The fraction of the penetrator made airborne has been measured and ranged from 0.9% to greater than 70%.(2,3,4) The broad range is due to a variety of factors such as penetrator velocity and size, type of tar-get, method of measurement, etc. For the purposes of this study, 100% of the materials generated during test firings were considered to challenge the air cleaning system.
Unfiltered ambient air is drawn into the enclosure during test firings. The median background level for real-time, mass-airborne monitors currently in use onsite is approximately 30 micrograms per cubic meter. Assuming 8 hours per day of operation, as much as 10 g of fugitive dust could be drawn into the facility and challenge the air cleaning system. This does not include water or water vapor concentrations which vary greatly. The chemical composition of the fugitive dust is not known.
Small quantities of soot, unburned propellant grain and vapors (water, nitric acid, potassium hydroxide, sulfuric acid) from the burning of the propellant are found. The vapors could condense upon cold surfaces or be absorbed in water vapors or particles, thereby producing some potential for corrosion. The quantities generated per firing are estimated to be negligible.
Based upon analysis of the airborne particulate material from outdoor test firings, it was estimated that the composition of the airborne material will be 80% to 95% oxides of uranium (75% of which is U308), 4% to 19% iron oxide with minor amounts of aluminum oxide, titanium oxide, and silicon dioxide plus trace quantities of the pro-ducts from the combustion of the propellant. If the target material is non-ferrous, most of the iron oxide would be replaced by the new target material. It was further assumed that the total mass of air-borne material is equivalent to the mass of the penetrator.
Size Distribution of Airborne Particles
Particle size distributions of the airborne material under precisely the Range D operating conditions have not been found. Size distributions of depleted uranium oxide particles under similar conditions have been reported . The distribution with the largest fractions of small particles, Table 6, presents the most difficult challenge to the air cleaning system and was assumed for the purposes of this study.
Table 6. Approximate Aerodynamic Equivalent Particle Size Distribution
Particle Aerodynamic Mass Equivalent Diameter, Percent in Micrometers Size Range <0.18 31 0.18 - 0.56 14 0.56 - 1.8 15 1.8 - 5.6 13 5.6 - 18.0 11 18 - 56 7 >56 9
Airborne Mass Concentrations as a Function of Time
Measurements of the aerosol concentrations made 7.5 seconds after the shot during the unenclosed test firings ranged from 1.5 to 2.6 g/cubic meter(3). Measurements of the aerosol concentration made 5 minutes after the shot during the enclosed test firings ranged from 0.05 to 0.15 g/cubic meter.(4.6) The critical airborne concentration for this study is that which will challenge the air cleaning system. If the entire penetrator mass were airborne in the target enclosure, the airborne mass concentration would be 4.2 g DU/cubic meter. For this study a maximum airborne concentration of 0.5 g DU/cubic meter with the particle size distribution shown in Table 6 was assumed. This aerosol has a half life of 4.1 minutes. in the target enclosure.
Candidate Systems
The basic assumptions for selecting equipment were: 1) particle collection should not produce a liquid effluent requiring treatment, and 2) the final control device should be a bank of HEPA filters to act as a polishing and safety stage. The fifteen candidate systems are listed in Table 7.
Baghouses, precipitators, common rolling filters and cyclones are familiar devices in the air cleaning field. The lesser known devices will be briefly described in the following subsections; however, because the pulse-jet baghouse will be shown to be one of the recommended devices, a typical cutaway is shown in Figure 3.
Table 7. Candidate Air Cleaning Systems
1 Existing Disposable Filters
2 Existing Disposable Filters With Extended Surface Filter In Place of Bank 2
3 Rolling Prefilter + Disposable Filters
4 Peeled Roll Filter + Disposable Filters
5,6 Pulse-jet Baghouse + Disposable Filters
standard felt media
coated media
7,8 Shaker Baghouse + Disposable Filters
standard fabric media
coated felt media
9 Electrostatically Augmented Baghouse + Disposable Filters
10,11 Cartridge House + Disposable Filters
standard media
coated media
12 Cyclones + Disposable Filters
13 Electrostatic Precipitator + Disposable Filters
14 Cleanable High Efficiency Filter + Disposable Filters
15 Vaned Inertial Separator + Pulsed Panel Filter + Disposable Filters
Extended Surface Filter
The extended surface filter is a prefilter that fits into the same holding frame as the current second prefilter. The medium is a thick glass-fiber pad shaped into pockets or envelopes that are about 30 in. deep. The efficiency is about the same as the current second prefilter but has greater particle holding capacity.
Peeled Roll Filter
The peeled roll filter is similar to the common rolling filter; both use clean and used media reels. However, the airflow passes radially inward through the layers of media wound on the clean media reel and then out through the end of the reel's core instead of through the media stretched between the reels. Because the air passes through many layers of media the filtration efficiency is greater than for the common rolling filter although the media is essentially the same. When the outer layer becomes loaded with particles the layer is wound onto the other reel. Several media units are contained in a single housing. This filter was originally designed for viscous liquid aerosols.
Figure 3. Cutaway of Pulse-Jet Baghouse (after Mikropul(7))
Cartridge House
A cartridge house collects particles on a filter medium made of a treated cellulose paper which is pleated and formed into a cylindrical cartridge, very similar in appearance to a truck carburetor air-cleaner. The cartridge is about 12 in. diameter by 24 in. high. Figure 4 shows a typical cartridge house with several cartridges mounted below a tube-sheet in a configuration similar to a pulse-jet baghouse. Dirty air enters near the bottom of the hopper and passes upward through the cartridges to be exhausted at the top of the housing. Particles collected on the cartridges are removed by pulses of clean air directed through the cartridge in a direction opposite to the normal airflow. The removed particles that have agglomerated to a sufficiently large size settle into the hopper by gravity. Venturies between the compressed air nozzles and the cartridges amplify the cleaning airflow.
Figure 4. Typical Cartridge House (after Donaldson(8))
A preferable configuration has a walk-in plenum on top of the tube sheet. For filter maintenance a person enters the plenum, which is on the clean side of the cartridges, and the cartridges are removed through the tube sheet. Figure 5 shows another advantageous design where cartridges are mounted on a slight angle to the horizontal plane. The clean air side, tube sheet and pulse cleaning pipes are at the back of the cabinet. Pairs of cartridges are accessed through hatches on the front of the cabinet. This design has several more seals that must be checked for airtightness; however, with a modified hatch design, it lends itself to a type of bagout method for cartridge changes similar to that used in glove-boxes. The dirty airflow is from the top to the back of the cabinet which would enhance the movement of particles to the hopper during pulse cleaning. The cartridges are also compartmentalized in groups of twelve to reduce the tendency to blow particles to adjacent cartridges during cleaning.
Cleanable High Efficiency Filter
The cleanable high efficiency filter, Figure 6, contains one or two banks of high efficiency filters that look very similar to the HEPA filter used in the existing system. The filters differ from
Figure 5. Downflow Cartridge House (after Donaldson(8))
Figure 6. Single Stage Cleanable High Efficiency Filter (after Kermatrol )
HEPA's by a thin coating on the particle collection side so particles do not penetrate into the fibers of the media. The medium acts as a support for the porous coating which does the filtering. The filters are cleaned in-place by either a vacuum system or pulses of compressed air.
Another type of cleanable filter uses 80% ASHRAE dust spot efficiency cleanable filter panels housed in a steel box with hoppers beneath. The filter medium appears similar to that of the the medium efficiency prefilter in Bank 2 of the existing system except the medium is not pleated as deeply, and the individual filter dimensions are 2 ft x 4 ft x 3 in. For the design airflow the device would contain 24 panel filters. The filters are cleaned periodically with pulses of compressed air directed opposite to the airflow. This device will be referred to as the Pulsed Panel filter.
Vaned Inertial Separator
The vaned inertial separator is rated at 80% ASHRAE dust spot efficiency. Ninety percent of the airflow is sharply turned through vanes while 10% of the airflow proceeds in a straight line. Large particles cannot follow the rapid change in direction and follow the 10%. airflow in a straight line. The 10% airflow, now more heavily laden with particles, is passed through another dust collector (the manufacturer recommended the Pulsed Panel) which removes 80-90% of the particles. The airflows from the vaned inertial separator and the Pulsed Panel are recombined and passed through a second Pulsed Panel.
Economic Analysis Method
Because this air cleaning application is unusual and applicable data describing the performance of the filtration devices was lacking, costs attributed to the different system configurations were estimates based on several factors with considerable uncertainty. Factors considered in this study were the particle collection efficiency of the devices, individually or in series, and the frequency of mechanical repairs. The former proved to be the most significant because several important items hinge on the question of efficiency, including the life of media, frequency of media replacement, media disposal and the disposal of collected dust. We used the best efficiency data available to us from the literature or vendors; however, in most cases, extrapolations beyond the data were required. Estimates of maintenance and repairs were obtained largely from vendors. Estimating methods were applied uniformly to identify the relative ranking of the candidate systems even though the estimates were inexact. In most cases the candidate systems include all or a part of the existing filter banks as polishing filters. To explore which filter banks should be retained, cost estimates were generated for each practical combination of the existing filters with the candidate devices.
A complete estimate of procurement, operating and maintenance costs of each system was attempted. The cost elements addressed were
- Media Life -- how long the media lasts
- Media Replacement -- replacement and disposal of media
- Waste Disposal -- collected powder
- First Costs -- initial investment
- Maintenance -- operation, maintenance, repairs
- Electrical Power -- power for fan and auxiliaries
The general method of estimating these elements will be explained in the following subsections. Labor costs are assessed under the topics to which the labor is applied. The several assumptions, derived from current operating practice, that provided a starting point for cost estimating are enumerated in Table 8. Others will be introduced as they pertain to specific method elements.
Table 8. Assumptions Used for Cost Estimating
1 Labor rate for BRL personnel: $31/man-hour 2 Power cost: $0.04/kWhr 3 Waste disposal cost including container, shipping fees, labor, burial fees: $25/cu.ft 4 System equipment life: 20 years 5 Test firings during 20 yr life: 6000 6 Average weight of test round: 4.9 kg 7 Average aerosol mass per test: 4.9 kg 8 Bulk density of the collected dust: 1.5 g/cu.cm 9 Annual operating time for system 365 day x (5/7) x 8 hr/day = 2006 hr 10 HEPA filters will be retained as last device in system for a polishing and safety stage
Media life
The media life expectancy is a concern for all of the devices except for cyclones or ESP's. Mechanical media failure which results in particle leakage is the principal factor determining the life of cleanable media. The filtration efficiency of cleanable media affects only the subsequent disposable filters, because it determines the challenge aerosol. For disposable filters, the filtering efficiency and the particle holding capacity are equally important to media life. Therefore the question of media life was approached differently for the two media types.
Disposable Filters
The approach developed was applied to all the types of disposable media investigated: the rolling filters, the existing filters, and extended surface filters. To estimate media life, the filtration efficiency as a function of aerodynamic particle size and the dust holding capacity at the final operating pressure drop must be known; however, data covering the entire range of interest were generally not available and extrapolations were required.
The particle size range of interest was subdivided into seven size brackets equally spaced logarithmically and the mass of particles in each bracket was estimated as was shown in Table 6. All of the particles in a given bracket were assumed to behave as the particle whose size was the bracket midpoint. The collection efficiency of each device was obtained for each size bracket midpoint. The amount of particles collected by and penetrating a device was computed separately for each size bracket. The total amounts collected
and penetrated were computed by summing the bracket values. The size distribution of the penetrating particles was used as the input aerosol of the next stage of filtration. This approach was used to estimate the useful life of a bank of disposable filters in a series of banks.
Table 9 is an example performance calculation for the three existing filter banks in series. The initial mass distribution, fractional efficiency, mass collected and penetrated for each stage and size bracket are shown. Also shown are the overall and cumulative efficiencies for each stage.
Table 9 also illustrates how the cost of filter usage was calculated. The table shows the vendor's estimate for the maximum amount of dust that may be collected on a bank of 24 filters at the pressure drop which requires filter replacement. The mass collected by a filter stage was divided by the dust capacity to estimate the fraction of filter life, or life fraction (LF), used in the test. The inverse of the life fraction, 1/LF, is the number of tests the filter can be exposed to before replacement and is shown in the table as "Shot Life". While this method may be simple and have some theoretical shortcomings, the results shown in the table for "Shot Life" were within the range experienced at BRL and MTD, giving us some confidence in applying the method.
Table 10 shows how we calculated the estimated costs for replacing the banks of existing disposable filters as $267, $1949 and $4291 for stages 1,2 and 3 respectively. The calculations accounted for filter cost, labor cost and disposal cost as low level radioactive waste. The compacted volume fractions used (0.067, 0.088 and 0.25 for banks 1,2 and 3 respectively) were those achieved at BRL. The life fraction times the bank replacement cost was the estimated cost incurred for the test, which was multiplied by the number of tests expected in twenty years to estimate the cost during the facility life. Table 9 shows the results of these calculations for the existing filter system. Similar calculations were made for each candidate. system configuration. Table 10 is illustrative of the factors accounted for in computing the replacement costs of filters in the other types of devices.
Cleanable Filters
Estimates of media life for the cleanable media devices were obtained from the equipment manufacturers' experience. The control-ling factor for these devices is mechanical wear resulting in filter media failure. Most often these devices are used in harsh industrial ventilation environments where aerosol concentrations are higher and more continuous than expected for the firing range. Most vendors felt that the aerosol loading at the firing range is light compared to their typical applications. The life estimates for each type of device using cleanable media are given in Table 11.
To estimate the characteristics of the aerosol penetrating to the filters downstream and the quantities of dust to be transferred to disposal containers, a calculation method similar to that described above for the disposable filters was applied to most clean-able media devices. For pulse-jet baghouses an alternate method, based on the work of Leith and Ellenbecker(10), was used for estimating particle flux to downstream devices. This alternate method discards the notion of collection efficiency and relates particle penetration flux to the media area, dust cake on the media and the frequency of cleaning.
Table 9. Example Efficiency and Cost Calculation for Existing Filters
Size Distribution Bank One -- 25% ASHRAE Filters Midpoint Initial Fractional Grams Grams Dia. Microns Mass% Grams Efficiency Collected Penetrated 0.1 31 1519 0.00001 1.519E-02 1.519E+03 0.32 14 686 0.00001 6.860E-03 6.860E+02 1 15 735 0.007 5.145E+00 7.299E+02 3.2 13 637 0.82 5.223E+02 1.147E+02 10 11 539 0.983 5.298E+02 9.163E+00 32 7 343 0.995 3.413E+02 1.715E+00 100 9 441 0.9972 4.398E+02 1.235E+00 Total 4900 1.838E+03 3.062E+03 Cum. Total 1.838E+03 Stage Efficiency 3.752E-01 Cumulative Efficiency 3.752E-01 Max Load, g 6720 Shot life, shots 3.66 Bank Replacement Cost $267 20 yr Cost By Bank $438,260 System 20 yr Cost Bank Changes in 20 yr 1641 Size Bank Two -- 95% ASHRAE Filters Bank Three -- HEPA Filters Midpoint Fractional Grams Grams Fractional Grams Grams Dia. Microns Efficiency Collected Penetrated Efficiency Collected Penetrated 0.1 0.83 1.261E+03 2.582E+02 0.998 2.577E+02 5.165E-01 0.32 0.94 6.448E+02 4.116E+01 0.99968 4.115E+01 1.317E-02 1 0.987 7.204E+02 9.488E+00 0.9999 9.487E+00 9.488E-04 3.2 0.998 1.144E+02 2.293E-01 0.99999 2.293E-01 2.293E06 10 0.9994 9.158E+00 5.498E-03 1 5.498E-03 0.000E+00 32 0.9996 1.714E+00 6.860E-04 1 6.860E-04 0.000E+00 100 0.9997 1.234E+00 3.704E-04 1 3.704E-04 0.000E+00 Total 2.752E+03 3.091E+02 3.086E+02 5.306E-01 Cum. Total 4.591E+03 4.899E+03 Stage Efficiency 8.990E-01 9.983E-01 Cumulative Efficiency 9.369E-01 9.999E-01 Max Load, g 9600 9600 Shot life, shots 3.49 31.11 Bank Replacement Cost $ 1,949 $ 4,291 20 yr Cost By Bank $3,352,883 $ 827,574 Systems 20 yr Cost $4,618,717 Bank Changes in 20 yr 1720 193
Table 10. Cost of Replacing Banks of Disposable Filters
Bank 1 25% ASHRAE Prefilters New filters = 24 x $6.15 = 148 Changeout labor = 2 manhr x $31/manhr = 62 Compactor operator = 1 manhr x $31/manhr = 31 Disposal = 16 cu.ft x 0.067 x $24.26/cu.ft = 26 Total = $267 Bank 2 95% ASHRAE Filters New filters = 24 x $67.50 = 1620 Changeout labor = 2 manhr x $31/manhr = 62 Compactor operator = 2 manhr x $31/manhr = 62 Disposal = 96 cu.ft x 0.088 x $24.26/cu.ft = 205 Total = $1949 Bank 3 HEPA's New filters = 24 x $130 = 3120 Changeout labor = 16 manhr x $31/manhr = 496 Compactor operator = 3 manhr x $31/manhr = 93 Disposal = 96 cu.ft x 0.25 x $24.26/cu.ft = 582 Total = $4291
Maintenance
We accounted for waste disposal, equipment repairs and operating labor as maintenance costs.
Waste Disposal
The major waste disposal problem, other than disposing of filters (which was accounted for in previous sections), is disposal of dust collected in the hoppers of most of the candidate systems. The dust will be a mixture of about the same composition as the air-borne particles. The dust collected in the hoppers will be transfer-red to a container via an airlock valve. To estimate the disposal cost we made the following assumptions: 1) 55 gal drum, 7.35 cu.ft; 2) 2 hr labor involved in filling each drum; 3) powder bulk density the same as uranium oxide, 1.5 g/cu.cm, 4) $25/cu.ft disposal cost includes the container, shipping and burial costs. Allowing some head-space, each drum will contain about 300 kg of dust. Because the quantity of dust collected in the hopper varies with the equipment, individual dust disposal costs were estimated for each system.
Table 11. Media Life
Life Expectancy, Device Months Common Rolling Filter 1 Peeled Roll Filter 84 Pulse-jet Baghouse: standard felt 24 coated felt 27 Electrostatically Augmented Baghouse 24 Cartridge House 24 Cleanable High Efficiency Filter 80-120 Pulsed Panels 1
Repairs
The expected repairs included compressor maintenance,. pulse valve rebuilds, cleaning cycle control board replacement, blower repair, and a 16 man-hr annual allowance for miscellaneous repairs.
Operation
To estimate the labor cost of operating the systems we estimated the time involved for daily start-up and shutdown, annual inspection and annual housekeeping. For the daily operation we assumed 20 min was required to start and shutdown the existing system. For the candidate systems we added additional time for more complicated equipment. The approach for inspection and housekeeping was similar; one day for housekeeping and a half-day for inspection for the existing system per year. For the other systems additional time was allotted for more complicated equipment as appropriate.
First Costs
The first cost, or initial capital investment, included the costs of fabrication, flowrate controls, upgraded fans for various pressure drop requirements, air compressors, concrete pads, sheds, drum handling equipment, freight, ductwork of various types, installation labor and leak testing in the field as appropriate for each candidate system. The material of construction for most system components was mild steel.
Electrical Power
For BRL, power costs for the range were not part of its direct operating costs; however, power costs were estimated to provide a basis for discriminating between systems with otherwise comparable costs. Included in our estimates were the power for operating the fan, the control device and the auxiliary equipment such as compressors or blowers.
Total Costs
Following the method outlined, the twenty year cost of acquiring, maintaining and operating each candidate system was computed. Total costs were estimated for 294 different combinations of the 15 primary control devices and after-filters. HEPA filters were assumed to be used as the final stage of filtration to serve as a polishing stage and as a final barrier to particle release. It was further assumed that the HEPA stage would require a medium to high efficiency prefilter upstream that would collect the bulk of particles penetrating the first device. This ensures that the HEPA stage can remain functional even when the first filtration device fails catastrophically by explosive overpressurization of the system during a test firing. For the preferred after-filter combination of a particular candidate system there were several vendors which supplied estimated initial acquisition costs. The complete economic analysis was performed for each vendor's equipment and the results for maintenance, power, filter replacement, and etc. were averaged. The averages were then used as the cost estimates representative of the particular candidate system. Intercomparisons between the candidate systems were based on these averages.
Comparisons of Candidate Systems
Summarized cost estimates for candidate systems are shown in Table 12. The first two columns of Table 12 list the configuration recommended for each device. The initial investment values include equipment acquisition, installation, freight, ventilation fans, controls and desirable accessories for waste handling. The filtration penetration estimates are for the device plus the filter banks as a system. Given the uncertainties in filter efficiency data, the penetration estimates are probably accurate to 1-2 orders of magnitude. The totals in the last three columns exclude the cost of power as requested by BRL. The operating cost plus the initial procurement is the total number of constant dollars that would be spent on the system during the 20 yr. facility life. Amortizing the operating costs over 20 years at four percent interest and adding the initial cost yields the life cycle estimate in the last column.
The systems in Table 12 were listed in order of ascending life cycle cost. There were two distinct groupings with about an order of magnitude difference in total cost between the two. The expensive group was characterized by systems where most of the filtration duty is performed using disposable filters. This was true even of the cyclone system because the cyclone's low efficiency for the challenge aerosol puts the filtration burden on the disposable filters. The average life cycle cost for this group was about $6500K. The less expensive group was characterized by performing most of the filtration with a high efficiency device using cleanable filters or no filters at all as was the case for the ESP. The average life cycle cost for this group was $570K. The most important distinction between the two groups was the cost entailed in the useage and disposal of throw-away filters. Another significant difference, but of lesser impact, was the relative average costs of maintenance, $139K and $345K for the first eleven and the last five systems respectively. It should also be noted that none of the expensive group of systems met BRL's efficiency criterion of 99.998%, although they were close to meeting the criterion within the accuracy of the available data. The expensive group was then eliminated from further consideration.
Table 12. Cost Summary by Air Cleaning System
Calculated Maintenance .
Initial System Total Total
System Investment, Fractional Labor, Parts Total Power .
Device + Filter Bank $K Penetration Man hr $K Maintenance mw hr $K .
Cartridge, Coated Media 2,3 129 5.0E-08 3542 8 118 3475 139
Clnbl. Hi-Eff Filtr, 1 Stg 2,3 187 2.0E-07 3884 3 124 5750 230
Clnbl. Hi-Eff Filtr, 2 Stg 197 6.3E-07 4638 2 146 3850 154
Electrostatic Baghouse 2,3 167 6.3E-07 4946 3 157 4000 160
Baghouse, Pulse-jet, Std. Felt 2,3 157 2.0E-09 4748 8 155 4000 160
Baghouse, Shaker, Std. Sateen 2,3 134 1.0E-08 6192 1 192 2950 118
Baghouse, Pulse-jet, Coated Felt 2,3 157 3.8E-10 4434 8 145 4000 160
Cartridge, Standard 2,3 129 5.0E-06 3732 8 123 3475 139
Vaned Inertial Separator 2,3 119 7.0E-06 4158 2 131 4525 181
ESP 2,3 466 1.6E-06 3427 1 107 3100 124
Pulsed Panel 2,3 100 9.0E-06 4096 2 129 3275 131
Peeled Roll Filter 2,3 253 1.1E-D4 8047 1 250 4825 193
Cyclone 1,2,3 112 1.1E-04 8447 0 262 6575 263
Standard Rolling Filter 1,2,3 31 1.1E-04 10585 1 329 2950 118
Existing System 1,2,3 20 1.1E-04 12041 0 373 3200 128
Extended Surface @ Bank 2 1,2,3 10 2.5E-04 16470 0 511 2700 108
Waste Disposal Excluding Power .
Replacement Vol, Total Operating Operating + Life
System Media, $K Man hr cu.ft $K $K Initial, $K Cycle, $K
Cartridge, Coated Media 40 252 1043 34 191 320 413
Clnbl. Hi-Eff Filtr, 1 Stg 14 204 749 25 163 350 429
Clnbl. Hi-Eff Filtr, 2 Stg 15 205 776 26 186 383 475
Electrostatic Baghouse 20 214 1075 34 210 377 480
Baghouse, Pulse-jet, Std. Felt 60 259 1400 43 257 414 540
Baghouse, Shaker, Std. Sateen 49 221 1247 38 279 413 550
Baghouse, Pulse-jet, Coated Felt 94 208 1100 34 273 430 563
Cartridge, Standard 122 347 1528 49 294 423 567
Vaned Inertial Separator 216 440 1807 59 406 525 724
ESP 51 255 1004 33 191 657 751
Pulsed Panel 249 477 2008 65 443 543 759
Peeled Roll Filter 2649 3138 15345 481 3380 3633 5285
Cyclone 3031 3739 17607 556 3848 3960 5842
Standard Rolling Filter 3482 4814 19633 640 4451 4482 6658
Existing System 3631 5661 20920 698 4703 4723 7023
Extended Surface @ Bank 2 3640 4640 33064 970 5121 5131 7634
Several of the candidate systems in the lesser expensive group were eliminated for operational reasons. The vaned inertial separator and Pulsed Panel systems were eliminated because of the inconvenience and potential inhalation hazard inherent in requiring that the filter panels be changed by entering the contaminated hopper portion of the Pulsed Panels.
The ESP had the lowest estimated maintenance cost ($107K) of the remaining systems; however, the ESP had a significantly higher initial cost which outweighed the maintenance advantage. We also suspected that the ESP would be more susceptible to "puffing", briefly reduced efficiency for the instant of pressurization following a test shot, than would devices using filter media.
The shaker baghouse was eliminated for several reasons. First, it relies entirely on a dust cake for filtration, which would be disrupted during cleaning and probably during each test firing pressure pulse. Smith, Cushing and Carr (11) showed that shaker efficiency declines significantly following cleaning until such time as a dust cake is rebuilt, which would be a long time in the firing range case because the dust concentration is essentially atmospheric. Second, the pressure pulse could severely damage the unsupported filter bags. The filter bags are inflated by the ventilation airflow and the only supports to the fabric are rings sewn into the fabric at intervals. The brief pressurization following a test shot may over-stress an inflated bag. This argument would also apply to the electrostatically augmented baghouse due to its similar fabric support configuration. An additional drawback of the shaker baghouse was the very large number of bags (960) that would eventually require replacement.
The remaining devices, in order of initial cost, were
- Cartridge house ($129K)
- Pulse-Jet Baghouse ($157K)
- Cleanable High Efficiency Filter, 1 Stage ($187K)
- Cleanable High Efficiency Filter, 2 Stage ($197K)
Table 13 lists the pros and cons for these remaining candidate systems. Any of these three would be suitable air cleaning systems for this application depending on the preferences of the user for the positive aspects of each.
The cartridge house system with walk-in filter access was chosen for the firing range application. A single cartridge filter was adapted to the prefilter bank of Range C for some preliminary testing. After one test shot the filter appeared undamaged from the pressure pulse. Initial static tests in Range D resulted in failure of the original joints in the front and back filter house walls which have since been rebuilt. Because of that failure, the cartridge house will be fabricated of 0.25 in. steel and must pass a leak test at the design pressure of +3 psig. We anticipate that the cartridge house will attenuate the pressure pulse transmitted to the filter house.
Table 13. Pros and Cons For Remaining Candidate Systems
PRO CON Cartridge House 1. Filter removal 1. Standard high from clean side, pressure housing out of weather, unavailable 2. available Extensive user 3. experience Broad vendor support 4. Filters compact, easy 5. to change Lowest initial cost 6. Compact housing Pulse-Jet 1. Filter removal from 1. Walk-in filter access Baghouse clean side, out of requires high-bay weather, available 2. Filters awkward to 2. Extensive user change because of experience support cage 3. Broad vendor support 3. Highest number of 4. Standard high filter elements to pressure models change 5. available Fewest door seals Cleanable 1. Convenient filter 1. No existing systems High-Efficiency changeout of comparable 2. Fewest filter capacity elements 2. Many door seals that 3. Compact housing could leak 4. Bagout system for 3. Most expensive filters 4. Single vendor filter 5. Eliminate existing supply filters w/2 stage 5. No standard high housing pressure models 6. Two stages of filters 6. Single stage models with slightly larger had highest pressure housing drop
References
- ASHRAE 52-76. 1976. Method of Testing Air Cleaning Devices Used in General Ventilation for Removing Particulate Matter, American Society of Heating, Ventilating, and Air Conditioning Engineers, New York.
- Glissmeyer, J. A., and J. Mishima. 1979. Characterization of Airborne Uranium from Test Firings of XM774 Ammunition. PNL-2944. Pacific Northwest Laboratory. Richland, Washington.
- Gilchrist, R. L., and P. W. Nickola. 1979. Characterization of Airborne Depleted Uranium from April 1978 Test Firings of the 105 mm, APFSDS-T, M735E1 Cartridge. PNL-2881. Pacific Northwest Laboratory. Richland, Washington.
- Chambers, D. R., R. A. Markland, M. K. Clary and R. L. Bowman. 1982. Aerosolization Characterization of Hard Impact Testing of Depleted Uranium Penetrators, Technical Report ARBRL-TR-02435, U.S. Army Ballistics Research Laboratory, Aberdeen Proving Ground, Maryland.
- Trinks, H. 1981. "Gun Muzzle Blast Field Research: Multiphase Flow Aspects and Chemistry of Muzzle Flash Including Chemical Flash Suppression." Published in Proceedings from Sixth International Symposium on Ballistics, Joseph E. Backofen, ed. October 27-29, 1981. Orlando, Florida.
- Hanson, W. C., and J. C. Elder, H. J. Ettinger, L. W. Hantel, J. W. Owens. 1974. Particle Size Distribution of Fragments from Depleted Uranium Penetrators Fired Against Armor Plate Target. LA-5654. Los Alamos Scientific Laboratory. Los Alamos, New Mexico.
- Mikropul Corporation. Copyright 1982. "Mikro-Pulsaire Dust Collectors," Brochure PC-5, Summit, New Jersey.
- Donaldson Company, Inc. Copyright 1983. Torit Division, Brochure GB-83, Minneapolis, Minnesota.
- Kermatrol, Inc. 1980. Drawing No. 1000-1-21-02. Santa Monica, California.
- Leith, D., and M. J. Ellenbecker. 1983.. Dust Emissions from a Pulse-Jet Fabric Filter, Filtration and Separation, July-August 1983, pp. 311-314.
- Smith, W. B., K. M. Cushing and R. C. Carr. 1981. Measurement Procedures and Supporting Research for Fabric Filters. In Proceedings of First Conference on Fabric Filter Technology for Coal-Fired Power Plants. CS-2238, Electric Power Research Institute.
DISCUSSION
SCRIPSICK: What is the life extension of cleanable filters relative to the filters with no cleaning? What is the efficiency of the air cleaning system you described?
GLISSMEYER: Assuming effective cleaning action, the life of the cleanable filter is not directly related to particle loading in the same way as a filter with no cleaning. For a cleanable filter, the life is more often related to mechanical wear, which usually occurs during the cleaning cycle. One manufacturer of a cleanable coated filter recommends replacement after 2,000 cleaning cycles. A second manufacturer of a cleanable coated filter suggests a life of 80-120 months. There was no evidence from users of such filters that I contacted to contradict the manufacturers' claims as long as the cleaning was effective for the particles; however, their longest experience was only three years.
Our working criterion for the combined air cleaning system efficiency was 99.998%, to meet the sponsor's requirements. Using the estimating method outlined in the paper, all of the recommended systems, and some of the others, were expected to meet the criterion. The efficiency data, as a function of size that were used in the study will be included in a report to be published. Except for the rolling filters, cyclones, and vaned inertial separators, all of the cleanable prefilters should have an overall efficiency of around 99% or better for the aerosol distributions in the paper.
BERGMAN: I want to concur with and emphasize one of the findings of your study; namely, that the, use of cleanable prefilters may result in significant savings in total filter costs. We had installed an electric prefilter in a Pu process box in Rocky Flats and were able to extend the HEPA life by over 50 times and had 98% recovery of the Pu aerosols.
GLISSMEYER: Most cleanable prefilters have good efficiency and can result in a very significant cost savings over non-cleanable filters in applications with aerosol concentrations greater than ambient outdoor concentrations and may be of value for systems where operability during accidents must be assured.
CADWELL: The-components that are referred to as cleanable HEPA filters are not, by definition, HEPA filters because their initial resistance to air flow is substantially greater than the maximum allowed for HEPA filters.
GLISSMEYER: This is correct. These filters are also not DOP tested as required for HEPA filters. We, therefore, do not refer to these filter as HEPA filters but as cleanable high efficiency filters.
|
To
send us your comments, questions, and suggestions click
here |
