Dust Collection Guide

(A comparison of different methods)

There are several approaches to the issue of extinguishing sparks in a gas stream.

Important Factors in Spark Arrestor Selection

(1) There is no such thing as an efficiency rating for spark arrestors. They either work or they don’t. Remember, it takes only one spark/ember getting through the device to cause a fire or explosion.

(2) Maximum turbulence is the key to effective spark arresting and in the selection of a spark arresting device. Some devices do not impart enough turbulence (and/or pressure drop) to be 100% effective. The recommended pressure drop for an in-line device (one that is installed in a section of the ductwork) is between 0.75 and 1.5 inches WC. Anything less is highly risky. This is a basic law of physics.

(3) Pressure drop across a QUENCHER style of spark arrestor is a function of the Reynolds number which is proportional to the density for air. This means that a unit can be sized smaller if operating at a higher temperature. For instance a spark arrestor operating at 440 degrees F is 2/3 the size of the typical unit applied at 70 degrees F and the pressure drop will be designed the same. This lowers the cost of the spark arrestor and ensures its effectiveness. The density is also affected by the water vapor in the gas stream. It has little effect at temperatures below 125oF but can be a major factor when operating at higher temperatures.

(4) If the gas stream has dust that might drop out in the duct at the velocities in the blender style or QUENCHER spark arrestor, a booster must be provided to periodically remove this accumulation. If this unit is not kept clean, it might pose a threat by putting an extra load on the ductwork. Without an automatic duct cleaner-booster system, the spark arrestor would require periodic manual cleaning.

(5) The duct cleaner – booster design is also temperature sensitive and must be altered to accommodate changing gas stream conditions.

(6) Most suppliers do not have the capability to modify the designs as referred to in item (3), (4) & (5) above.

Blender Type Air Mixers

A number of these air blender/mixers have been applied with varied success as in-line spark coolers, arrestors and suppressors. Over the last several years standard air mixers have been adapted and applied between the spark generating process and dust collector. They were applied in processes where fires in the dust collectors had previously occurred. One supplier hired a consultant to develop a market for these air blender/mixers as a spark arrestor/cooler. This air blending or mixer style design was an outgrowth of mixing two gas streams of different temperatures to insure a uniform temperature after the static mixer. It was deduced that the gas stream produced turbulent flow as it passed through the blades and this was the reason it could be adapted to spark cooling. However, these are air mixers first and spark arrestors second. They are marketed as having low pressure drop (maximum 0.5 inch WC) through them. There are performance limitations because not enough turbulence (and related pressure drop) is imparted to the spark/ember. To achieve spark suppression, we need to go from laminar to highly turbulent flow in the duct which strips away the hot air envelope around the spark/ember thereby cooling it and starving it of fuel (oxygen). For air blending this is not a requirement. Also, these devices have large gaps between the mixing blades, when looking through the inlet and downstream of the device. These gaps can allow a percentage of sparks/embers to slip through and cause a fire or even an explosion in the dust collector.

Improved In-Line Spark Arrestors

QAM developed the QUENCHER, which is a variation of the blender/mixer design. It is also an in-line spark arrestor. Employing a 60 year old spin vane mist eliminator technology developed by Sly Manufacturing in the early 1960’s, led QAM to vary the blade designs to have the most effective performance, inducing maximum turbulence to the gas stream, and lowering the cost. Maximum turbulence (and the pressure drop that results from it) is the key to spark arresting. After several tests it was found that the air blending/mixer design did not impart enough turbulence and some sparks got through, especially at low gas stream velocities. Eventually, there was a specific design which imparted the most effective swirling and turbulence thereby extinguishing the sparks quickly and most effectively. In fact, during testing of the QUENCHER, the arrestor cell would light up as a ball of fire, however, one inch past the cell nothing was left in the gas stream. These designs were incorporated into the QUENCHER. QAM has developed special application data in which the blade angles are adjusted to produce minimum effective pressure drop for different temperatures and gas densities. To our knowledge, no one else accounts for the gas density effects on spark arrestors. In truth, due to the advanced design, even applying the incorrect parameters to a QUENCHER may not result in a failure to put out sparks. Since the pressure drop across the device are a function of the velocity through it, the development of a pneumatically operated booster was introduced to prevent dust dropout accumulating in the static arresting cell. It also blows out accumulations on the blades.

Link to Quencher spark arrestor

Liquid Spray Systems

For many years these systems were the only method to prevent fires caused by sparks. The system consists of electronic detectors that detect sparks and react to their presence. When a spark is detected liquid sprays are actuated and water sprayed into the duct. The sprays actually cool the gas stream below the dew point. However, in dust collection systems, the water then wets the filter bags or cartridges. This prevents fires but the gas flow is interrupted and the bags must be either replaced or dried out before the process can resume. It takes a whole day or two to dry out the bags or even to prevent blinding and replacement. The detector sensitivity can be lowered to prevent excessive actuations, but, this reduces the reliability of the systems. The detector missing a spark is an ever present danger and a fire may occur. Bag or cartridge replacement is definitely required.

Cyclone Dust Collectors

Contrary to common belief cyclones are not effective spark arrestors. For a spark arrestor / cooler to work, there must be high turbulence in the air stream. If you have turbulence in a cyclone the pressure drop is very high. Cyclones are designed to avoid turbulence. Many bag house fires occur in systems with cyclone pre-cleaners. Amazingly the inlet baffles on the baghouse are more effective as spark arrestors, however they are not foolproof.

Static Blade Spark Suppressor (Tri Pass)

These were developed in Japan to replace multiple cyclones in coal fired boilers. They found that the multiple cyclones did not stop sparks from entering the dust collectors. The first ones were installed in the early 1970’s. They ran at 1.5 inches of pressure drop and were fabricated from structural angles to resist the wear of the abrasive ashes in the coal that they fired. There are several of these applications installed in the USA and Canada designed by one of our colleagues.

Static Baffle-Box Spark Arrestor

Many dust collector suppliers offer this type of device as a spark arrestor. It consists of air entering at one end of a baffle box running over a baffle plate which drops out the sparks and much of the dust collected. The air exits at the other end, and then travels to the dust collector. The big drawback is that a hopper and flexible or solid hose connection to a collection barrel is required. Also, these devices do not eliminate all of the sparks. There is not enough turbulence generated to ensure 100% spark arresting. Sparks may also ignite the contents of the collection bin under it.

Mesh Filters

This is a common stop-gap measure where the filter is placed at the exhaust duct of hoods or installed in the ductwork. When clean, the mesh filter will stop at best 80% of sparks. These filters do not produce enough pressure drop to be fully effective. It only takes one spark to ignite dust in the duct or set a dust collector on fire. The only thing these filters do is clog up and add to your maintenance.

We trust that the above information will enable you to evaluate and select the most suitable method and supplier for your application. Buying our QUENCHER/BOOSTER spark arrestor combination will give you a risk free unit, fine tuned for each application.

Equipment:
C5-2500, orifice scrubber style wet dust collector, rated for 2500 CFM, purchased to handle explosive aluminum dust particles.

Problem:
The dust was going right through the collector and packing into the fan / outlet compartment. Very little dust was collected in the dust collector sump.

Investigation and observations:
We requested pictures and system layout drawings (sketches were actually provided). From these we observed that the client did not describe the application accurately at the time of purchase.
1. The dust was produced from a spray coating operation. Therefore, it was fine powder type aluminum dust. Wet collectors are designed for metal dust 5 microns and larger, as generated from grinding and cutting operations.
2. The inlet was connected to a properly sized 8” duct but was over 20 feet long with three elbows. These units are designed for maximum 10 feet of duct directly to the collection point.
3. The client also decided to exhaust the discharge of the collector to the outdoors with another 20-25 feet of duct, and three more elbows. These collectors are designed for an open, unrestricted discharge on the top of the unit.
The result of this was a questionable capability of collecting the powder type dust. The biggest problem was that the collector performance was choked by far too much resistance to airflow in the installation. By doing this the air entered the collector with far too little volume to cause the necessary turbulent energy in the “omega” style baffles. The necessary wetting action of the dust particles was not taking place and filtering action was non-existent.

Solution:
1. We asked the client to place some of the collected dust in a closed jar with water. Then shake it and let it sit for a few minutes. If the dust settles, it can be collected. If it doesn’t, a different dust collection solution must be found.
2. A wet dust collector is very particular about the airflow through it. You need to be in a range of +/_ 10% of the rated CFM for that model dust collector. In this case, the minimum flow that could be tolerated is 2250 CFM. Conversely, with more than flow than the 10%, water gets drawn up too much and discharges out the unit. We recommended bringing the collector closer to the application and remove the duct on the outlet. Alternatively, add a booster fan to overcome the restriction.

Other Considerations:
A. If the excess resitriction is minor (within the 10% range) but dust is discharging at the top, You would add more water to the collector, in small increments, until the dust/water stops coming out the top. The added water compensates for the higher restriction. Then reset the float control to maintain this new water level.
B. In some cases the discharge is required to be exhausted by code. An example is beryllium. In such a case, do not attach a duct on the outlet. Instead build a capture hood over the top of the wet collector outlet, approximately 4-6 above the top, and duct that to the outside. Install a fan to give you 5-10% more air flow than what is running through the collector, to ensure no contaminents escape back into the room.
C. If you oversized the wet collector, Do not restrict the flow with a damper more than the 10% tolerance. Install a bleed-in on the inlet duct with an adjustable shut-off damper. Open the damper to the point where the collector performs properly. This is often a scenario when you size the collector for multiple collection points but don’t have them all installed until later.

This is a touchy topic and greatly misunderstood. Today, plant operators recognize that there is more danger from lawyers, on this issue, than from actual explosions in the dust collector. In fact, the accumulation of dust in the plant itself is a greater danger for conflagration than a properly engineered dust collection system. Go to our website at http://www.qamanage.com/ and view the videos on “Combustible dust in the workplace” by 60 Minutes on CBS and the U.S. Chemical Safety and Hazard Investigation Board.

Explosions in Dust Collectors

Explosions in pulse jet collectors invariably are when cleaning off-line and can be prevented by sound practice. We run into a risky process when we write the safe procedure because it is application dependent and relies on common sense of the operators. If your design is good engineering, there will be no explosion. All explosions in pulse jet dust collectors, we have investigated (about 100 or so), have been clear cut stupidity. Among the typical ones; horseplay, ignoring warnings posted on equipment and disgruntled employee sabotage.

The most risky application is a mechanical cleaning (shaker style) dust collector when cleaned off-line, and, you can only clean them off-line. A dangerous spark is generated by static charges produced during the shaker action. If I were there, I would tell someone else to turn off the collector or put a long delay on the shaker actuator while I go to the restroom. There is no particular reason to warrant us to observe an explosion first hand. I am cowardly since I heal from injuries slowly.

Advanced Technology vs Poorly Designed Dust Collectors

An overwhelming number of explosions have occurred on badly designed collectors with bottom inlets in which the fine dust has difficulty in making its way to the hopper until the fan is shut down. Advanced technology dust collectors (such as ULTRA-FLOW), with their high side inlet and high ratio cleaning system, have some marked advantages that further reduce the risks involved in explosions. One major advantage is the extremely high efficiency of these designs which prevents dust being returned to the plant, thereby reducing the hazard of accumulated dust referred to the introductory paragraph of this bulletin. The cleaning system more thoroughly cleans the bags and the inventory of dust on the bags is very low, and usually not sufficient to cause the dust concentration to go above the lower explosive limit in the event of an explosion front traveling into the collector from the inlet ducts. With these collectors, the fire and explosion generally occurs outside the collector, in the ductwork, and is drawn into the collector. In normal operation there is only a very small part of the collector that passes through the lower explosive limit. This consists of a narrow band about 1/8 to 1/4 inches thick that surrounds the bag when it is cleaned. Eggshell or singed finishes on the filter bag is recommended to further reduce dust inventory on the bags.

Placing the explosion vent below the filters (i.e. in the hopper) is a bad idea. It allows the pressure to build up in the housing before it can be released by an undersized vent in the hopper. This is especially true with conventional dust collector designs that have venturis restricting the neck of the filter cage at the tube-sheet. With advanced technology designs (having no venturi), we have 12-15 times more open area to the outlet which in itself is a natural explosion vent. We place the vent in the housing side where it offers the most protection by venting the explosion immediately where it occurs.

Woodworking

Because of the NFPA rules do not directly apply to dust collectors, there is much latitude in their interpretation. The solution is to apply sound engineering to assess the risk and to provide equipment suitable for a particular service. Venting woodworking applications is probably the largest number of installations in the dust collector industry. Explosions have occurred and the venting has been quite effective in controlling them.

The norm in the industry, for the last 30 or so years, has been to provide a 60:1 vent ratio. This has been sufficient for this service. ULTRA-FLOW uses a standard 20:1 vent ratio for its explosion vents, which further protects against the harmful effects of an explosion.

Vent Ratio

This was developed by UL labs. It is the ratio of the volume of the dirty air compartment of a dust collector to the area of the explosion vent. For example; a cylindrical bag dust collector with (24) 6 inch by 6 foot bags, dirty air housing size of 6ft x 4ft x 6ft, hopper which is 1/3 x ( 6 x 4 x6 ). The gross volume of the collector = 192 cu.ft. The volume of the bags is 24 x 1.2 cu.ft./ bag = 28.2. The volume of the collector = 192 – 28.2 = 164 cu.ft. Therefore, if we want a vent ratio of 20:1; 164/20 = 8 sq.ft. of explosion vent.

Ultra-Flow dust collectors use a vent ratio of 20:1. In general the insurance companies determine the specification that they want and we supply it accordingly. In the end, good engineering is the key.

Kst Ratings

This issue is very complex and not as easy as just meeting a “Kst” deflagration rating. It is an NFPA 68 test requirement for ideal lab conditions. “Kst” refers to the rate of pressure rise in an explosion. Unfortunately defining of the number is difficult since NFPA never really measure it except when they use a sealed globe enclosure, stir the dust in it and then try to ignite it with a sparkplug. This is not the real world of dust collectors.

A more accurate test was performed by AAF specifically on dust collectors. See the “Combustible Dusts” chart at the end of this bulletin. That chart shows the “Explosion Pressure” or burst pressure where theoretically a dust collector will blow apart in an explosion. If a dust collector is built of 12 gage steel to withstand +/- 20 SP (inWG). The burst pressure is usually a factor of 4 times that or 80 psi. As an example, for wood dust it was determined that a vent ratio of only 180:1 was safe in a dust collector. The chart says the burst pressure would be 35psi which is less than the 80psi allowed. Ultra-Flow uses a 20:1 vent ratio, therefore it is 9 times that value, so, you are as safe as you can get. No matter what you do, there will always be some risk. All we can do is make it inconsequential. As mentioned above, there is a far greater risk from dust in the plant than you will find in the dust collector itself. Look on the home page of our website for the news reports on “Combustible Dusts in the workplace”.

Use this link to view and print the “Combustible Dusts Chart” at the end of the original article.

Use this link to view the videos on “Combustible Dusts in the Workplace“

Some people have used Quenchers, and other style spark arrestors in plasma and laser cutting applications but still experienced fires in their dust collectors. Sparks are only one issue to deal with these applications. A good spark arrestor is definitely needed to stop sparks and embers, but, it is no guarantee against fires in the dust collector.

The problem:

1. The operator may have to reset the heat setting of the plasma head. It could be generating too much atomic static particles. This causes a “painting” effect on the cartridge media, eventually clogging it.

2. Large heavy particles of molten metal can be generated in the process. 

3. You should use spun bond wide pleat cartridges, to ensure proper clean out of the cartridges. That way the dust will spread over a large surface of media, instead of on the outer surface only.

 4. Current cartridges that are clogging over time (can vary from hours to weeks, depending on loading). When clogging occurs, the air flow drops and sparks can slip through any spark arrestor (not just the Quencher). This sets fire to the combustible dust accumulated on the surface of the cartridges.

Normally, plasma cutters have different characteristics depending on the settings of the cutter torch. The quantity of dust produced is relatively small. At some torch settings the dust is reactive by initiating an atomic bond between the dust and the surface of the cartridge, forming a hard durable impervious coating which totally or partially plugs the filter media. This mechanism is an inherent part of the plasma coating process to put wear resistant coatings on shafts, turbine blades etc. that allow the parts to receive very long lives. In the plasma coating machinery, the key to collecting the overspray in cartridge or fabric collectors is to allow the atomic bond to dissipate. This is accomplished by extending the time that particles travel from the torch to the filter media elements. In plasma coating systems at this time, depending on torch settings will vary from 0.5 to 0.8 seconds depending on the metals being sprayed. 

In plasma cutting applications often the dust being emitted from the torch does not require any special considerations. In fact, collectors can operate for many months quite well with moderate pressure drops. Then the torch settings are changed because of various factors such as the composition or thickness of the pieces that are cut. As the settings of the gun or the speed of the cut is changed, the dust can act as a plasma coating torch and the cartridges start plugging. Sparks are often produced. If the dust is combustible the sparks may ignite the coating on the cartridges. Normally the fuel on the cartridge surface is not very heavy so the fires do not damage the housing of the collector. The cartridges are then usually replaced. The QUENCHER spark arresters are sometimes applied to limit the risk of fires and extend cartridge life. In the tandem horizontal type collectors, the cartridges are usually tight spaced, so, as the pressure drop rises, the pleats are pinched in the valleys so the pressure drop goes up. Combustible dusts can put pounds of dust to be stored in the cartridges to fuel a fire in the collectors. However, the squeezing of the pleats also causes pressure drop to increase and slow the flow through the dust collector. This often allows dust to be released into the work area. 

Although spark arrestors will protect the system from sparks, pieces of molten metal go through the spark arrestor unaffected. These heavy, hot particles lodge on the surface of the cartridge and ignite the combustible dust coating. The heavy molten particles need to be dropped out of the system prior to the spark arrestor and collected safely, so as not to cause a fire in that collection device. Cyclones and drop out boxes are sometimes used for this. However, be aware that these devices have little effect on sparks / embers which are light buoyant particles and slip through to the dust collector. 

An excellent example of these effects was the experience of the Day division of Donaldson who supplies this design. In cutting the filter mounting plates for their design they plasma cut holes in a 1/4 inch thick plate. They found that the filters plugged quickly in the after filters. They added distance in the filters venting the operations. This experience occurred 20 years ago and we do not know how this operation is now performing. 

Our recommendation is to replace the current cartridges with a wide spaced stiffened spun bond media carried and precoat the cartridges with a 1/64 inch thick coating of inert pre-coat material. 

We suggest you send each job application data (layouts & pictures) to QAM technical support at garyb@qamanage.com and/or call him at 800-267-5585. We’ve dealt with plasma cutting applications for decades and feel that yours would be a common problem. If you contact us, we’ll be happy to work with you on this. 

For more information on; Quencher spark arrestor

For more information on; Dust collectors and dust collection solutions

Philosophy of design

1. Component and Fabrication costs
This line of collectors, Torit being the most popular, was developed for low initial cost, and to develop the replacement cartridge market. Donaldson has the lowest costs of production of any supplier. The replacement cartridges are at least 25% lower cost than the nearest supplier. They can sell cartridges at competitor’s costs and still have a good profit margin. With high volume customers, they use this ability to remove serious competition in a big account. For the other components such as valves and timers, they have the same advantage based on their purchase quantities. They buy in quantities of 25,000 to 50,000 and the competition buys in lots of 100-1,000.
On this line of collectors with the nominal 12 inch by 26 inch cartridge, they build the collectors in modular form so that there are two types of modules. There are two end modules and any amount of middle modules to come up with any size collector. The modules are bolted together so they match up and assembly is always the same. They can build modules in lots of 25 or more and assemble modules to make a finished collector. This lowers cost more than 35% compared to other suppliers with the same size and design collector. The modules are bolted together and they can ship collectors in two weeks with the first week devoted preparing the paper work. Since they use multiple inlets and outlets, it is hard to make a mistake in manufacturing that cannot be fixed by adding or removing modules.
 
2. Engineering
The holes in all of the flanges are gang punched with dies punched so they always match in spacing both for modules and hoppers. They promote their multiple hoppers to reduce engineering. The horizontal cartridge designs are designed for lowest possible headroom.

3.Shipping Costs
All assemblies are selected to be shipped with standard trucks and railroads without special permits.
 
Limitations of this design approach

This design requires special inlet and outlet manifolds. If cost of these manifolds are included, the cost of the system can be higher than competition. The configurations available are either right hand or left hand. Other competitive designs are more adaptable. Because of the bolted construction the tolerances between parts can produce misalignment when the cartridges are installed. One unit which we examined, which was a single cartridge instead of a tandem unit, we found that tolerances were such that the seals were not functioning. One side of the seal was bottomed and the other side was open so that dust continually leaked around the cartridge.

Capacity of cleaning system is limited by size of cleaning valve for the tandem set. This capacity is a function of the compressed air flow in the valve. As listed below the maximum filter flow rating for the tandem set is 810 CFM. The quantity of unplugged media whether fabric, cellulose or other media is also a function of valve size:

 Valve size                Filter flow    Media Cleaned  On line/off line
 1/2 inch valve       360 CFM         23 sq.ft./ 28.5 sq.ft.
 3/4 inch valve       810 CFM       49 sq.ft./ 61.2 sq.ft.
 1 inch valve          1440 CFM        87 sq. ft./108.8 sq. ft.
 1 1/2 inch             3240 CFM      193 sq ft. / 241.2 sq. ft.

Analysis of Two tandem cartridge design
This design can be analyzed as follow:
Valve: 0.75 inches to dean two cartridges
Maximum filter flow on line 810 CFM
Cartridge media area (two cartridges) 450 sq. ft.
From Table above      49 sq. ft. cleaned on line / 61.2 sq. ft. off tine
Cleaned area    49.5 sq. ft / 61 sq. ft.     Plugged 400 sq. ft. / 389 sq. ft,
Cartridge weight is 40 pounds

Approximately 550 grains per sq. ft. of 60 Ib. per cu. ft. dust are collected when media is plugged. 550 gr/ sq. ft x 400 sq. ft. = 220,000 grains.
220,000 / 7000 gr./lb. = 31.5 lbs per tandem set.
Total weight of cartridge set is 40 Ib. + 31.2 Lb. = 77 Ib.

Cleaning System Actuation (Recommended)
Another factor is that the cleaning action is generally initiated by a pressure switch. The recommended prevalent pressure switch setting is about 3 1/2 inches. For most applications the pressure should be about 3/4 to 1 1/2 inches w.c. above the initial pressure drop. Typically initial pressure drop through the cartridges is 0.3 -0.5 inches of water column. Therefore, at 3 1/2 inches w.c. pressure drop, less than 17 sq. ft. of the media is available because the dust bridges across the pleats rendering the rest of the media in a condition where it cannot be cleaned by the reverse jet flow. The cartridge must be cleaned three times more frequently than if the switch were set in the proper range. This also adds about three pounds to the operating weight.

Compressed air consumption
Compressed air usage is a function of the square of the operating pressure drop. The usage is also related to the dust loading and the fineness of the dust. With a typical dust from material handling operations such as sand or rock dust and 5 grains per cu. ft., the air consumption per thousand CFM of filtered air would be as follows:

Operating Pressure drop       SCFM at 85psig and 1000 CFM of filtered air
           1.5 inches                            0.2
           3.5 inches                            1.1
           4.5 inches                            1.8

Filter life
The filter life is related to the cleaning frequency (or compressed air consumption) so a collector running at 1.5 inches of water will last five times as long as one running at 3.5 inches all other factors remaining the same.

Efficiency
Over 95% of the dust penetration through the filter comes immediately after each cleaning cycle. The dust penetration is related to pressure drop. With the same conditions of loading described above, the efficiency and penetration at the three pressure drops follow:

Operating pressure drop     Penetration in grains per cu. ft.     Efficiency
                    1.5 in. w.c.                                    .00005                              99.9990%
                    3.5 in. w.c.                                    .00027                              99.9946%
                    4.5 in. w.c.                                    .00045                              99.9910%

Conclusion
This design has some serious limitations, some of which can be remedied by the QAM retrofit design which will reduce pressure drop, increase efficiency and extend cartridge life.

For more information:    Retrofitting dust collectors

Part 1 – Explosions in dust collectors
Dust explosions are possible whenever the process produces combustible dusts. Not all combustible dusts will produce explosions. For instance, even combustible dusts may not have the characteristics to produce an explosion. A coarse combustible dust such as coal may burn well but not explode depending on how fine the dust is. To produce a conflagration the dust must have a sufficient ratio of surface area to weight to sustain the rapid oxidation for creating and sustaining an explosion. When a dust can sustain an explosion, the dust concentration must be within the explosive limits.  These are often defined as:
L.E.L. (Lower Explosive Limit): Below this level of concentration, an explosion will not occur and propagate itself. There is not enough concentration of fuel to allow the flame front to grow. A typical range of values would be 20-30 grains/ cubic foot.
U.E.L. (Upper Explosive Limit): Above this limit the concentration of dust is so high that there is insufficient oxygen to oxidize the fuel and the unburned fuel stops the spread of the flame front.

Ignition of the dust depends on several factors
 (1) Chemical Composition
 (2) Shape and fineness, briefly described above.
(3) Dust distribution in the gas stream or atmosphere
(4) Concentration of oxygen in the gas stream.
 (5) Initial temperature and pressure of the gas.
 (6) Energy level available to detonate the explosion

Intensity of the explosion is dependent on the rate of pressure rise and maximum pressure developed. Factory Mutual ran lab tests to determine these values and are contained in their publications. It must be pointed out these tests and values are run with a spherical test chamber with power ignition source in the center of the sphere. These numbers are relatively high when referring to explosions in a dust collector housing, because the bags usually obstruct the expansion of the explosive flame front.

Limiting exposure to hazards
 A) Eliminate ignition sources. One source of ignition is sparks, often produced in the hoods venting processing machines. Sometimes the machinery can be modified to prevent spark generation. Another method is to install spark suppressors prior to the dust-laden gas entering a dust collector. For sparks to be carried along to the collector the flow must be laminar. Most dust collection ductwork is deliberately designed to operate with laminar (smooth) flow to reduce pressure drop but laminar flow produces a system that is an excellent vehicle to send sparks into a dust collector. Recently, there are offered some excellent designs of spark coolers that turn laminar flow into turbulent flow for very short distances then revert back to laminar flow. These devices require pressure drops of less than one inch water column and are easily installed. An automatic self cleaning device for these spark suppressors is available.
 B) Isolate Operations. The collectors may be located outdoors or away from the main production areas.
 C) Introduce inerting systems. Inert dust can be introduced into the system so that the lower explosive limit of the dust mixture can be eliminated therefore the mixture will not explode. This is especially effective where the dust concentration is very low. As an example, if we have a fume dust with loading of 1 grain per thousand cubic feet of gas which is combustible, and the system has 25,000 CFM, adding an inert dust at double that load to prevent it from burning would require:
 25,000 CFM  x 2 gr./ 1000 CFM  x  60 min./hr x 8 hrs/shift =24,000 gr./shift
24,000 gr./shift divided by7000 grains/ lb. = 3.4 lb./shift.
The other possibility is to mix the gas with another gas stream that may have the oxygen already oxidized as with a combustion process.
D) Explosion Vents can be provided on ducts and on the collector housings. These vents must be directed either outdoors or to an area where the explosion can be safely dissipated.
E) Changes in Design of collectors The collectors can be braced to withstand higher pressures so the explosion venting is more effective. Cylindrical collectors are more resistant to damage then rectangular collectors but rectangular units can be braced to withstand higher explosive pressures. Grounded bags are often supplied to drain off the static electricity charge that is a possible source of ignition. Grounded bags can provide a false sense of security to the operator and designers. Most often the dust that holds a charge insulates the static charge from the media. When the filter element is cleaned, sparks can be generated.
F) Changes in operation of electrical controls; Referring to figure 1, which is a view if a cylindrical bag in a Pulse jet collector.

The bag is the cylinder between the two dark hollow sections of cylinders. When the bag is cleaned a small volume of dirty air is propelled from the bags and extends a fraction of an inch between the bags. This forms a hollow cylinder of dust laden agglomerated dust. It is this hollow cylinder which is possibly between the LEL and UEL. Even if a source of ignition occurs from static charges the volume of the collector housing between these limits is usually less than 3% of the housing volume so the explosion would dissipate itself and cause no further damage. In investigating explosions in dust collection systems, there have been no verifiable explosions where the explosion was started from inside the collector from sparks while there was on line cleaning in the collector.
G) Sources of detonations In pulse jet cleaning collectors which experience explosions are as follows:
a) An explosion front traveling through the duct and enters the collector and dislodges the dust from the bags and then a secondary explosion occurs as the dust concentration in the housing goes between the LEL and UEL. An explosion can occur in the collector. Some methods to reduce this danger are to use smooth finish bags where residual dust on the outside of the bags is reduced. Egg shell bag finishes are a good selection. Another approach is to use laminated PTFE bags.
b) Off- line cleaning (in which the fan is shut down) increases an explosion hazard. If the collector is to be cleaned off-line with hazardous dust, the cleaning should be operated very slowly, with perhaps 3-6  minutes between pulsing a single row of bags. This will diminish the chances of the dust cloud passing through the LEL and UEL.
c) Hopper fires can occur if the hopper is not cleaned out before de-energizing the fan. These can be difficult to deal with. If a hopper door is opened an explosion can occur. There have been cases where operating personnel have tried to put out a hopper fire with a hose. The water stream agitated dust and formed a cloud of dust that passed between the LEL and UEL and the fire in the hopper provided a detonation source and serious explosions have occurred in the hopper. The best approach is to inject inert gas into the collector and allow it to cool below ignition temperatures before doing anything.

Explosion hazards in other Dust Collectors

Mechanical shaker collectors are inherently more hazardous than cylindrical bag collectors.. These collectors are cleaned off-line with no gas flow. There is always a potion of the collector which will pass through the LEL and UEL. Often the whole collector will pass through the limits during cleaning. The main approach to reducing the risk is to be careful to limit the sources of detonation when cleaning. Grounded bags are always recommended when explosive dusts are collected. Another approach is to use many small collectors instead of a central system. As discussed before, placing collectors outdoors is an option.

Pulse Jet Cartridge (Pleated filter element) Collectors are more and more of a selection in dust collector systems. When pleated filter elements were first introduced, they made the incorrect assumption that the main criterion was the filter ratio. While it is a criterion it is only one factor. A more important factor was the fact that a certain cleaning jet can only clean a fixed area of media. The rest of the area is plugged and it holds a lot of dust. This was discussed in lesson 14 (History of Cartridge Collectors). When it comes to explosion and fire hazards this plugged media contributes a lot of dust to fuel a fire or explosion if one is initiated. Some newer designs with pleated elements run at higher filter ratios and diminish the hazard by a wide margin.

Part 2;  FIRES

Requirements
As we discussed above some combustible dusts, may not have a LEL in any concentration of dust in a process gas stream. However fires can occur in ducts and in a dust collector. Fires in ducts are usually a result of poor duct design so that dust drops out in ducts.
Fires can occur in exhaust ducts as well as inside dust collectors. Requirements of fires or any combustion process are: a) Fuel, in gas liquid or solid form. b)  Oxygen (Atmosphere consists of 20 per cent oxygen)  c) Fuel must be raised to the ignition temperature to start burning.

Sources of ignition include: Overheating of coils, motors, Friction, spontaneous combustion, static discharge, burning debris drawn into the vent system.
          Spontaneous combustion occurs when dust slowly oxidizes in a collector or in any accumulated pile. The fuel oxidizes very slowly but the fuel is insulated by the dust.  A “hot spot “develops. When the collector flow is resumed or the dust pile is agitated it often acts like a spark to ignite the dust (fuel).
          Static discharge- Generally speaking static built up in a collector is reduced or eliminated by the jet cleaning system. The jet cleaning action dissipates most charge build up on the surface of the bags.
Burning debris drawn into the exhaust system can be a source of ignition.

Transport of sparks through ducts. Referring to the sketch below, there is a glowing ember surrounded by some hot air which gives the sparks buoyancy. This spark and the hot gas associated with the spark can travel hundreds of feet in a duct. The ductwork is designed to give laminar (smooth ) flow. This is illustrated on the left. Spark suppressors are placed in the duct to change the flow to turbulent (coarse) flow, as shown on the right. This agitation or turbulence strips the air from around the ember removes the fuel (oxygen), therefore extinguishing and cooling the spark below ignition temperature.

Prevention depends on eliminating the causes of ignition. Spark traps can change laminar to turbulent flow and extinguish any sparks in a duct. Design for proper dust transport velocities. Install pneumatic actuated duct booster to flush dust into dust collector.  Use air jets to remove electrostatic charges on the duct surfaces.

Spontaneous combustion in Pulse Jet Collectors can be prevented by pulsing the collector when the system is idle. This cools off the hot spots. For instance, brass furnace fires can be prevented by pulsing the collector every hour when the fan is not running.

Putting out fires can be accomplished by one of the following approaches: Cooling below ignition temperature, cutting off fuel supply, cutting off oxygen supply.

Water Hose and nozzles. This is an attempt to cool the solid fuels below the ignition temperature and to cut off the flow of oxygen to the fuel. It also takes away heat by turning water into steam. To boil one pound of water consumes over a 1000 BTU raising temperature of water by 200 degrees takes away less than 250 calories per pound. The steam generated can cause serious injury or death. The steam also displaces the oxygen in the air making it lethal but will often act as an inert gas to prevent oxygen from reaching the combustibles. Many dust collectors are equipped with spray nozzles. The hoppers should have automatic drains to prevent the water from doing structural damage. A 10 ft by10 ft collector with ten foot long filter bags can accumulate 40 tons of water if the sprays are not shut down or drained after the fire is extinguished.

Inert gas systems such as carbon dioxide or nitrogen gas are sometimes provided. Usually fire dampers will be provided to contain the inert gases. This will cut off the supply of oxygen to the fuel (dust and media)

Fan Operation during a Fire Whether to shut down a fan on a dust collector because of fire can be a difficult decision especially, if the collector is vented outside. Often, collectors are ignited at night and the smoke is not detected. The next morning the dust collector is virtually intact except the bags and dust have been consumed. For example a 10,000 SCFM collector removes heat at the following rate with inlet temperature of 100 deg. F. and various outlet temperatures:
 OUTLET deg F.   BTU/hr
 150                         500,000
 250                     1,000,000
 300                    2,000,000
 400                    3,000,000
 500                    4,000,000
 550 `                 4,500,000
With 1000 sq. ft. of cloth, the cloth would weigh about 100 lb and the dust about 50lb. If we assume a heat of combustion of 5,200 btu per lb, the BTU generated in a fire for this collector would be 780,000 btu. If we assume an outlet temperature of 450 degrees, it would take 30 minutes to burn itself out and the collector would probably not have damage to the tube sheet or cages. If the fan were turned off, immediately the temperatures would easily reach 1250 degrees and smolder for hours and the tube sheet and cages would be destroyed.
If the gas stream was re-circulated the decision is of course to shut down the fan.
Other fire extinguishing systems 
Manual Fire extinguishers are usually either inert gas like carbon dioxide or inert powders. The gas extinguishers usually cool and put a layer of inert gas between the fuel and the flame. The manual gas extinguishers should not be operated through doors of the dust collectors as the displaced gases can be vented out through the access doors. The other manual extinguishers usually spray powder into the flames. They are not suitable to spray through access doors.
There a several other systems to fight fires in dust collectors which will not be covered in this chapter.

For information on:

In-line spark arrestors (traps); http://www.qamanage.com/products/quencher_spark_arrestor.htm

Duct cleaner – boosters; http://www.qamanage.com/products/Booster.htm

Equipment:   Cartridge Collector with cellulose Media, 8 pleats per inch.

Application:  System designed to coat outside surfaces of gloves with starch to keep them from sticking together. It was effective and was meant for surgical use. The coating readily fell off the gloves when they were worn and before they were put to use

Observations: The starch dust was very coarse in the 5 to 10 micron size. The dust would easily fall from the fingers when squeezed and moved together. When the system pulsed large puffs of dust were observed at the collector outlet. We examined the dust under the microscope and it was uniform in size in the 5 to 10 micron range. However, the surface of the dust particle spheres were very smooth, almost polished. The collector ran with less than one inch pressure drop across the media, the same pressure drop as a clean cartridge. The lack of the ability to interlock the dust to form a cake meant that, during pulsing, the dust would tumble and could not agglomerate and this allowed the dust to penetrate the media during and after a cleaning cycle. The collector exhaust was returned to the work space, dust and all.

Recommendations, action, remedy and conclusions.
We recommended adding a small amount of ordinary cornstarch into the collector on an intermittent basis, once every four hours. This ordinary powdered (non-spherodized) starch formed a stable filter cake and the collector ran at 2 inches of pressure drop across the media, solving the penetration problem. The contaminated starch collected from the hopper was then sold for animal consumption but it was not a significant cost for the process.
Another solution was to use a standard tubular shaker collector with laminated (Gortex) media which does not require interlocking dust to form a cake.

This approach is suitable for other dusts that do not form granulated powders that don’t interlock to form a cake. Another example of this kind of dust was on a process where names were engraved on plastic. There was no granulated dust. The laminated media on cylindrical bags allowed the exhaust to be re-circulated.

Dust Collection Solutions and Applications and Troubleshooting

Although cement dust is relatively coarse and has optimum agglomeration properties, cement plants have some of the most demanding applications in the dust collection market.

First, all cement dust is dense and very abrasive. Even in the relatively less demanding application such as bag loading and bag unloading, these properties are important.

For many years, on the relatively light loading applications, shaker collectors were the preferred dust collectors. In fact the envelope collectors were very effective. The release of cement dust from shaker collectors was generally not a problem. In cylindrical bag collectors, the dust collected on the inside of the bag. The fastest moving air with its dust load was at the opening in the bag. There was a potential for wear near the bag entrance. If a bag developed a hole, the dust coming out of the tear would quickly abrade the surrounding bags. It seldom was possible to detect leaks until several bags, or even all the bags were damaged.

To filter a continuous process, off-line cleaning was necessary. On larger units, the collectors, in parallel sections or modules, were isolated and cleaned while maintaining flow through the remaining sections. On cement mills and other high load applications the collectors were usually preceded by cyclone pre-cleaners and the collectors often had dropout boxes or settling chambers.  On these applications, shaker filters ran at filter ratios below 2:1.  Air horns were often added to shaker collectors to improve cleaning and increase dust holding capacity.

In the late 60′s the Fuller Corporation introduced their version of the pulse collector. This compartmented collector had large diaphragm valves that discharged into the clean air plenum. The burst of air agitated the filter bags, producing a thorough cleaning. More dust collected on a unit of filter area to allow handling of heavier dust loads. The name of this collector was “Plenum Pulse”. It was able to handle the heavier load for applications such as raw mills and finish mills. The first installations vented clinker cooler vent systems. A measure of the efficacy of this cleaning system was that the clean air plenum was built with heavier gauge metal because of weld failure with 12 gauge construction. While this collector was touted to be a pulse-jet collector, it was not. The compressed air did not create a reverse air jet. The collector operated more like a shaker than a reverse jet design. It produced more energy to clean the bags than existing shaker designs. This was especially necessary on the heavier loading processes.

Another effort to handle this high loading was the combination of shakers and reverse air cleaning systems, provided by “Norblo” a company that was based in Cleveland Ohio.

However in the late sixties, MiKroPulverizer (later renamed MikroPul) and their then licensee, FlexKleen, introduced the reverse pulsejet collectors. They were first applied to vent the raw and finish mills in the cement production process. The reverse jet collector was developed by MikroPul for their Pulverizers. These were similar to cement mills and were able to handle the heavier loads of the cement mills without pre-cleaners.

These collectors ran quite well, with filter lives exceeding three to five years. Then in 1969, when MikroPul’s patent was challenged and declared invalid in court, many competitors copied their design. To counter the imitators, they made a major change in their design. The bag length went from six to ten feet. The pulse pipe orifice area was increased by the same ratio. Unfortunately, the venturi diameter was not changed. The velocity of the cleaning jet increased. It went from 15, 000 fpm to 25,000 fpm. Nobody recognized that the venturi velocity is proportional to the velocity of the dust leaving the bag towards adjoining bags during cleaning. In a dust like cement, it causes partial blinding and abrasive wear on adjoining rows during the cleaning pulse. Average pressure drops increased from 3 to 5.5 ” wc. Bag lives were reduced by 50-60%. Compressed air usage went from about 0.5 SCFM per 1000 cfm of filtered air to 1.2 SCFM per 1000 cfm of filtered air. To counter these effects, the industry made some basic changes in selecting pulsejet collectors. The changes treated the symptom rather than the cause.

• Filter ratios were drastically reduced. Bag life was increased and pressure drop was decreased to the 5 inch w.c. range.   
• Pressure actuated cleaning systems were introduced. This kept the abrading and blinding of bags from cleaning pulses to a minimum. (The cleaning frequency and air consumption per 1000 cfm of filter air remained much higher than with the old six foot long bag designs.)

The development of the ULTRA-FLOW advanced technology design was able to remedy all the shortcomings of the 1O foot bag design. These designs will be discussed in future papers of this series. Find out about Ultra-Flow at Advanced Technology Baghouse Dust collectors

Generally, the operators have chosen very conservative air to cloth ratios.  Cartridge collectors are quite effective on less demanding applications.

Application Engineering Data Filter ratio      Cartridges
The belief was that more filter media (and associated filter ratio) made the selection more conservative. This idea is firmly entrenched in the cement Industry. This is generally true with pulse jet fabric collectors with high velocity cleaning jets in that it extends bag life. The fact is that the opposite is true with cartridges because of bridging across the pleats if the media is not cleaned. As dust accumulates in the valley of the pleat, it bridges. During the cleaning cycle the cleaning air looks for the easiest path from the inside to the outside of the filter cartridge. That path is above the bridge.

What contributes even more to this phenomenon is the fact that a certain volume of reverse air can only clean a certain amount of media. Because the cartridge may contain huge amounts of media that cannot be cleaned, the media not cleaned will plug. In most designs running at filter ratios of less than two, an operating cartridge may contain 10 to 40 pounds of dust. Table 1 illustrates the area of media that can be cleaned with various orifices and /or converging diverging supersonic nozzles.

TABLE 1

Orifice / Nozzle diameter = area of media cleaned

0.250 in = 5.5 / 8.8 sq.ft.

0.312 in = 8.6 / 13.75 sq.ft.

0.375 in = 12.3 / 19.8 sq.ft.

0.500 in = 22 / 35 sq.ft.

0.750 in = 49.5 / 79 sq.ft.

1.000 in = 88 / 132 sq.ft.

1.500 in = 198 / 376 sq.ft.

There are new advanced technology cartridge dust collectors available today which incorporate wide pleat spacing, vertical cartridges, and supersonic nozzles, and, can be found at Advanced Technology Cartridge Dust Collector

Dust Collector Selection

The best designs for a fabric media pulse jet collector on these applications are those offered by ULTRA-FLOW with low jet velocities and higher filter ratios.  The characteristics of these designs are listed below:
Average velocity at bag opening  = 10,000 feet per minute
Bag opening (no venturi) = 4″ diameter
Jet volume  = 740 CFM
Bag diameter and length  = 4 inches x 96 inches
Bag area  = 10 sq. ft.
Filter volume rating per bag  = 190 CFM
Nominal filter ratio = 20 FPM
Average pressure drop  = 2 1/2 inches water column
Average Air Consumption  = 1/2 SCFM/1000 CFM of flow
Average dust penetration at 5 gr. / cu. ft. load  = 0.0005 gr. / cu. ft.

Conclusions

The best dust collection choices are a fabric collector with low jet velocities and a high inlet. The filter ratio is dependent on what the customer will accept, but ratios of over 18:1 are easy and reliable. Some modifications must be made to the inlet baffle because of abrasion.

Welding

Two Stage Electrostatic Collectors; Venting welding fume operations poses some difficult application decisions. Years ago, the preferred method of collecting weld fume was with two stage electrostatic precipitator dust collectors. These had several advantages; they were relatively compact and were generally very effective on general ventilation applications. They could handle relatively large gas volumes through the collectors and generally were located near the roofs of buildings.

Efficiencies of general ventilation; The collection efficiency was variable depending on the velocity going through the collection plates. The lower the velocity through the collectors the higher was the collection efficiency. The same collector might have an 80% collection efficiency at 6,000 CFM and a 98.5 % efficiency at 1,000 CFM The same collector could be applied to different exhaust volumes that would vary as much as a ratio of 6:1. The higher the volume would produce the lowest efficiencies. But the same air would be re-circulated and an acceptable level could be maintained in a particular room or building.  The cleaning of the precipitators were accomplished by a detergent wash system.
Loading for general ventilation; The loading for general ventilation units were from 0.1 to 0.5 grain per thousand cubic feet of volume. The washing frequency was typically once or twice a week. The presence of condensed hydrocarbons along with the fume was not a problem. Generally these would be oxidized into solids by the time the filter was washed. These collectors were generally the same ones that were applied as air filters in HVAC systems. The washing systems were designed for 1000 cycle life. This would translate to over ten years of life under these low loading conditions.

Hooded Systems; The trend was to hood the welding operations. The venting of hoods had some pronounced effects on the application of these precipitators. The load would vary from 5 to 20 grains per1000 CFM.

Effects of hooded systems; Usually the washing requirements were to wash the filter every shift or twice per shift because the load was so much higher. On a two shift operation, and washing twice per shift, the washing system had a life expectancy of less than 52 weeks.

Plating; It was necessary to operate particular precipitators at lower volumes with their associated higher efficiencies, because of a phenomenon called “plating” Referring to figure 1 below:

The precipitator will ionize the gas and the particles. As the dust passes through the precipitator it forms a bubble type shape, containing charged particles, which were not collected on the collection plates. The gas quickly loses it’s charge. However the dust that was not collected keeps it’s charge a little while longer and loses its charge as it leaves the boundary of thebubble marked “A” in figure 1. If the precipitator has a low efficiency the bubble is much bigger as marked by “B”. This low efficiency bubble is 2 to 20X as bigger in volume than the high efficiency bubble.
Under certain atmospheric ambient conditions, this low efficiency bubble starts to grow rapidly until the whole room atmosphere is ionized and the room and all the contents become collection plates for the dust. The dust s attracted to the walls, windows, machines, eyeglasses and every object that is grounded. All the surfaces turn blackened within seconds. There have been cases where this happened after the walls were painted white. After the atmospheric conditions go back to normal the plating stops
Bad Inlet conditions. All precipitators either single or two stage need even velocity distribution across the plates. If we had a gas stream averaging 100 fpm that would be designed to operate at 95% collection efficiency and  the real velocities entering the plates varied from 50 to 150 fpm, the section at 50 fpm might have a collection efficiency of 98% and the section at 150 fpm would be running at 80%. This would mean that the overall efficiency might operate at close to 85%. This condition would cause phenomenon described above
Lower Efficiencies, caused by running at lower average velocities were common. In the example above the collector might be selected to run at 125 fpm and an average efficiency of 85%. Plating may be produced because of these lower velocities and lower collection efficiency. As a result, many collectors were purchased based on volume and the supplier’s guaranteed higher collection efficiencies. There were practically no way to specify the collectors except based on supplier claims. Velocity was not a good criterion. Some collectors at relatively higher velocity and longer sets of collection plates would achieve the same result as a collector with short plates and a slower velocity.
Electrical Controls; To further aggravate the problem improper electrical controls were offered. Some operators interpreted SIC controls to mean arcing was not allowed. To eliminate the arcing across the plates, they lowered the voltage controls to halt this sparking. Unfortunately the voltage was lowered so much that the particles were not charged. There were cases where the metal pre-filters were more efficient than the precipitators.
Insulator Coating; The main collection plates are at ground in a two stage electrostatic precipitators. The charged particles will be attracted to the lower voltage intermediate voltage level plates and to the grounded collection plates. The insulators were also at ground level and some of the dust (a very small percentage) stuck to the surface of the insulators. This was a very strong bond and the spray cleaning systems could not keep these insulators clean. Eventually they were coated badly enough that the power supply could not keep the charging electrodes to ionize the gas and the particles. To correct this problem required a major overhaul of the precipitator. The charging electrodes were made of very fine wires and would eventually break and require replacement. Most electrical maintenance men were not familiar with high voltage supplies and maintenance was neglected.
Innovations in Design; In the late 70’s, two stage precipitators with pressurized insulators and more rugged washing systems were introduced. The insulators were subject to a gas stream that entered the collection compartment at a higher velocity than the collection velocity across the plates. This protected the insulators from charged particles. The charging electrodes were made heavier to give much longer charging electrode life.
These innovations increased the collection costs with electrostatic collectors to the point where cartridge dust collectors introduced at the same time were more economical to purchase and operate.
The advantages of the electrostatic collectors were:

1. The pressure drop was constant and usually low.
2. They could collect liquid droplets
3. They had the potential of long periods of service without maintenance.

Cartridge Dust Collectors
With the development of cartridge collectors, another method of collecting fume dusts became available. The standard design pulse jet fabric collectors with cylindrical bags did not work because the cleaning systems propelled dust through the cake of adjoining rows of bags during the cleaning cycles. In the late 70’s and early 80’s thousands of cartridge collectors were applied to both hooded and non-hooded ventilation systems.
Problems developed in many systems after the mid eighties. High-pressure drops and short cartridge life developed in many systems. The causes were one or more of the following:
The presence of thin films of oil on the surface of the parts that were welded. When electrostatic powder coating finishing systems were widely applied to reduce or eliminate hydrocarbon generation from paint systems, the faces of steel parts required protection from oxides on the surfaces. During the welding process condensing hydrocarbons were liberated and swept up into the ventilating systems and their associated collectors. One of several results followed:
a) The solids to liquid ratio was so high that the dust blotted the liquids and the collection system was not affected by the liquid droplets.
b) The solids to liquids ratio was in range where the powders and hydrocarbon mixture formed a paint and the collection media was gradually plugged. This could take days, weeks or months, but the net effect was the  cartridges had to be replaced or laundered pre-maturely.
c) The solids to liquid ratio was so low that liquid wetted the cartridges and they were plugged as in (b) above. Even in cases where the coating was barely discernable, this could occur.

A case in point was in a plant making stainless steel mufflers. The metal was washed after forming and the load in solids was 0.02 grains per 1000 CFM, the pressure drop rose in a six month time period. The re-enforced cellulose media would be air-dried and the pressure drop would be reduced from six inches to 0.3 inches after the elements were installed. After 4 months the pressure drop went up to six inches. After washing the pressure drop went down to 0.8 inches.  The next washing cycle came two months later and the pressure drop returned to 1.2 inches. It was less expensive to replace the cartridges than to wash them in such a short interval.

Washing cellulose media cartridge elements: After each washing, the media is wetted, the permeability of the media diminishes, even if no dust remains at or below the surface of the media. The wetting causes the media to matt. If oil wets the media it is a good blotter and the fibers may grow. This causes the pressure drop and base permeability to decrease.
Other media are available that can be washed and are not wetted by oils. These are referred to as oleophobic media. This is a coating on the fibers that does not change the permeability. Otherwise they will be called washable. Often they can collect a mixture of fumes and hydrocarbons because the fibers do not swell.
Treated Spun bond medias are widely applied. Some of these are excellent choices but have limitations. For instance, with tight pleats, the top of the pleat may squeeze so the media in that portion of the pleat may make contact on the clean side when the pressure drop rises. On some applications, over 80% of the pleat of the media may not be effective.  The remedy is one of the following.
A) Provide pleats with wider spacing and make them shorter in depth. This will allow full use of the media    available in the filter element.
B) Provide a media that has stiffness and will not collapse on itself.
C) Provide a laminated media with the clean side backing very open so that if the pleat squeezes there will be flow through the media.

Fume Generating Processes Similar to Arc Welding and Gas cutting

Thermal Deposition Processes
Spray Coating The first type was a flame spray coating machine. These fed a material into a high heat gas torch. The temperature achieved was so high that feed material would produce a material in gaseous/liquid form that started to condense into molten droplets. Though the process is not understood, it is presumed that some of the adhesion was from a nuclear bonding, in addition to the cooling of the molten droplets on the piece to be coated. There were some materials that were too porous and there was limits to the thickness of the coating. The over-spray that did not adhere varied from about 5 – 20% of the material fed into the coating generating gun. The over spray was generally collected by medium pressure air washer scrubbers at a 99% collection efficiency.
Plasma Arc Spray To get smoother surfaces and better adhesion to the target surfaces, an electric arc was added to gas flame. This produced much higher temperatures in the gun at the point where the powder or wire feed entered. It generally produced more over spray (10%-40%). This over spray was much finer and would lose its ability to stick and adhere to surfaces. This over spray was too fine to be collected efficiently with air washer wet scrubbers. Fabric or pleated cartridge collectors were necessary.
One serious problem was encountered. This involved residence time of the dust between the gun and the media collection surfaces. In a system installed in 1975, on a plasma arc spraying machine for coating electrical capacitors. The process was coating plastic surfaces with metal. The cartridge collector filter elements, venting the over spray, plugged up in less than ten minutes. The six cartridges each with 50 square feet of filter, (300 sq. ft. total) received less than 250 grains of dust. The dust collector was connected within 20 inches of the gun. The over spray dust adhered to the media surface and blocked the pores.
Through experimentation and field experience it was determined that if the dust stayed in the gas stream for relatively long periods of time, it would lose its ability to coat the media. Depending on various factors such as the feed rates of gas, solids and the arc current, this time varied. It varied from 0.5 to 1.0 seconds. Referring to the figure, the residence time will be analyzed.

The part to be coated is placed in a hood with the gun at the front of the hood. The hood is 6 foot long and is rectangular with a 4 x 4 opening.  The face velocity of the hood is 350 feet per minute. The duct is sized at a 2500 feet per minute duct velocity and the duct is 15 feet long. We will assume the back of the hood has a transition 2 foot long, designed like an evasé to have uniform velocity distribution.
1) Time to travel through the hood 6 ft / 350 FPM = 0.017 seconds
2) Time to traverse duct to the collector 15 ft / 2500 FPM = 0.006 seconds
Residence time = 0.017 + 0.006 = 0.023 seconds.
The flow through the system is 350FPM x 16 sq. ft. =  5600 CFM

To re-design collection for longer residence time the length of travel in components are altered and the velocity can be modified. The hood is the first to be looked at.

3) The hood would be made 10 foot wide with the same 4 foot by 4 foot opening for the gun and part. The velocity in the wide part of the hood would be 5600 CFM / 100 sq ft = 56 FPM
The residence time in this portion of the hood would be 18 feet divided by 56 FPM = 0.32 seconds.
4) The duct could be extended to 200 ft by putting in ductwork in a “serpentine fashion” and enlarged to drop duct velocity to 1,000 FPM. The residence time in duct would be 200 feet/ 1000 FPM = .0.20 seconds

The residence time of the system would be 0.32 + 0.20 = 0.52 seconds.

High Temperature Cutting Processes
This high temperature flame coming from the gas gun proved to be an excellent improvement in flame cutting. Instead of jagged edges near the cut, it became much smoother and for most applications it did not require smoothing the edge or the operation was very quick. With digital cutting machines the precision rivaled other cutting processes.

Plasma cutting and laser-enhanced cutting are in common use. The type of dust produced runs the gamut from arc welding to that of metalizing operations. Most dust is more similar to venting systems for arc welding operations, but to get some cutting characteristics the temperature and flow in the gun are adjusted. This may produce a dust that is prone to coat surfaces and media.  When this happens the residence time requirements may be in the same range as the electro deposition processes. Laser cutters work well with 1 second residence time. Some flame cutters have been applied to non-metallic pieces such as wood and plastics. These dusts can contain tars, and oils from non-metallic parts and the collector media can get plugged easily, within a few seconds. With metallic parts, the oils can be an imperceptible film on the metal or originate from the compressed air compressor. In that case, a low-pressure scrubber may be a good choice. Roll filters with replaceable media or a self-feeding pre-coat material system have been employed.
It is crucial to have the correct airflow at initial start-up. Too much airflow will reduce residence time and cause the painting effect. Install a control damper in the main duct and use an approved method to accurately measure the exact airflow. Use the damper to choke the system if needed.
Recently, it has come to our attention that some plasma cutting processes are throwing out the 1 second residence time rule of thumb. Either the process temperature is being cranked up so high that the molten metal atoms still don’t have enough time to form molecules or the dust concentrations are so low that the atoms never get a chance to collide with one another in the laminar flow of the duct system. In these cases, finding the correct residence time
is almost a trial and error process. A new product has come on the market, called a Quencher, which is inserted in the ductwork as close to the source of dust as possible and no less than 10 duct diameters upstream from the collector. This device imparts a high energy multi-directional swirl to the airstream which cools the metallic atoms, accelerate their oxidization, and forces them to collide together and form molecules which can safely be collected without the painting effect.

De-agglomerating dust
Normally we would run a properly designed dust collector at 1 to 1.5 inch water column pressure drop.  Sometimes a system will only stabilize at a higher reading (E.G. 3 to 4 inches). One possibility is that it takes 3 to 4 inches to a agglomerate and fall to the hopper.  It may be de-agglomerating when you pulse at lower pressure drops.  In that case off-line cleaning should drop out the de-agglomerating dusts.  Some dusts are more susceptible to this phenomenon than others.  Often, they put an anti-rust wipe on the material being cut.  If it contains ceramics then we will have this problem.

To get more information:

Consulting services

Quality Air Management

May, 2009
Tire Recycling plant, Dunnville, Ontario
Equipment; Baghouse Dust collector collecting from hammer mills
Dust; rubber, synthetic fibers, steel wire, carbon

Occasionally, the filters in a shaker or pulse cleaning dust collector will lose their efficacy because of moisture or other liquids entering the collector.
1) The base filter media will be coated with moisture. In general if the media does not absorb moisture the filter elements can be recovered if the pressure drop has not reached levels over 15-29 inches depending on dust characteristics. The load of dust must be then stopped and the flow through the collector is reduced to between 20% to 50% of the normal gas flow. These conditions should continue until the moisture is evaporated. The media cake will have been restored enough to allow reverse air to flow through the media and increase the permeability. Repeated cleanings will restore the media in the cleaning process so that it will be able to support an effective filter cake. This normally requires 10 to 40 cleaning cycles. The pressure drop should be monitored as it is possible to over clean the bags so that dust may come out the clean side and cause problems if the exhaust air is re-circulated to a work area. This procedure is effective if the media has not been physically or chemically attacked, (cellulose media will swell and lose its permeability when subject to repeated wet and dry cycles). Chemical or solvent attack presents more complex problems where the section of the media becomes critical. If the pressure drop is too high before the above process has begun, there is good possibility the dust can have imbedded itself in the fibers where the cleaning system can not restore the permeability enough to recover the filter elements.
2) The next possibility of blinding is due to a “painting” phenomenon. In that case the dust reaches the cleaning element surface in a paste or liquid paint form. It does not form a permeable cake. Usually there will be some powdered granulated dust that will be mixed with the paste. The ratio of liquid to powdered or granulated powder is critical. If the ratio of dust is very high, the collector will operate normally. Many processes do have occasional loads of liquid dust such as when dew points are high. The pressure drop rises and falls temporarily but the operations are satisfactory.
Our examination and testing of the bags revealed that in this process in which automotive tires were recycled to recover the metal and the rubber through hammer mills there was a dust/paste mixture reaching our continuous cleaning pulse jet collector that had a high proportion of liquid paste compared to the granulated powder. The bags were completely blinded. We sent them through normal fabric laundering process that removed a lot of the dust but were still badly blinded. We concluded chemical attack but the base fabric was as strong as a brand new bag.

We then subjected the bag to some deep cleaning procedures that are used to clean grease and sludge from some commercial and industrial processes. It was a last ditch effort to investigate the problem. We were applied our technology from related fields to develop a procedure to clean the bags and restore the media to virgin conditions. Cleaning the bags is challenging and can be somewhat messy. In addition it has some other hazards notably those involving the risk of fire and toxicity when inhaled. To address all these concerns we are proposing a three step cleaning process which addresses these concerns. These we shall refer to as coarse cleaning, medium cleaning and fine cleaning.

These are listed below.
1) Coarse cleaning is meant to remove the most of the dirt (by weight) clinging on the bags. Our first step was to separate the bags into groups of five and lay them on a concrete floor with three inches between the bags when they are laid side by side. The worker starts at the open top of each bag with his feet standing on the bags with about a foot between his shoes to brush the dirt into the space between the bags until he comes to the closed end.
2) This procedure is repeated for all five bags in the set. The dust is then swept with a hand brush and dust pan to be disposed of or recycled to the process.
3) The bags are turned over and the same procedure applied again.
4) Medium cleaning is meant to remove much of the remaining dust. Tie the open ends of the bags with a cord and put them in a laundry tub with clothes washing detergent. Swish them around in the tub and then rinse.
5) Final cleaning will get out the hardened paste by dissolving the hardened paste {latex type paint} it also can be used to strip latex paint from a piece of canvas. We use a surfactant compound that we use in cleaning grease in commercial [restaurant grills.] and high speed industrial machining to collect submicron smoke and oil mist. We use these in our other product lines. The surfactant powder can be ordered from our lab facility in Charlotte, NC.
6) After the bags are rinsed in step 4 above they are again put in a laundry tub with one cup of the surfactant powder in each tub, if multiple tubs are used. They are swished around and rinsed as in the medium wash.
7) Finally, they are hung up to dry.
We can check one of the cleaned bags to be sure that they have recovered enough to use. You can send the bag to our office in Waterloo to test permeability.
9) Give one pint of Canadian whisky to each of the workers who cleaned the bloody bags.

Retrofit and Dust Collector Consulting