Acceptance of the process called flue gas conditioning (FGC) as an engineering solution to a common environmental problem has come slowly. To start with, the process is mis-named. It is not the gas which is conditioned, but the particles of dust in the gas, or in the most common application, the fly ash. Adjusting the resistivity of fly ash particles by application of a dilute acid solution to their surfaces is an easily understood and straightforward task. The very idea that the behavior of a gas stream might be adjusted in any useful way by addition of a few parts per million of anything boggles the mind.

In spite of an inappropriate name and some bungled efforts, early trials of FGC showed benefits which justified continued experimentation, and encounters with several dozens of ways of doing it wrong eventually developed a set of reliable rules for doing it right. One result is that use of sulfur trioxide (SO₃) for adjustment of the resistivity of fly ash from low sulfur coal has been widely applied (See Table 1) and has become an automatically accepted part of the option of switching to low sulfur coal for compliance with the Clean Air Act of 1990. NERC estimates that approximately 45,800 megawatts of generating capacity will utilize coal-switching plus flue gas conditioning for this purpose. Guarantees that this equipment will be available to operate at least 98 percent of time it is called upon are routinely fulfilled.

With very few exceptions all commercial SO₃ FGC systems installed to date are based on catalytic conversion of sulfur dioxide (SO₂) which is either supplied in liquid form or obtained by burning elemental sulfur. Elemental sulfur is the preferred feedstock for long-term permanent operations because its operating costs are lower, but liquid SO₂ is used for trials and small or short-term situations where reduced capital costs can offset increased feedstock expense. The catalytic conversion design was chosen over other methods by which SO₃ may be made available partly because it is flexible and easily controllable, but mainly because it minimizes the quantity and the difficulty of handling hazardous materials resident in the system.

Figure 1 is a diagrammatic representation of a typical sulfur-burning system for SO₃ conditioning. Molten sulfur, a common item of trade in the chemical industry available throughout the U.S., is delivered by thermally-insulated tank trucks fitted with steam coils for melt-out. In locations where sulfur cannot be delivered in molten form, bagged or bulk solid sulfur may be supplied for melting on site. The sulfur grade is designated as "Bright Yellow" which contains very low levels of contaminating materials. The exact analysis varies slightly among suppliers, but completely lacks chemicals which could act as catalyst poisons and contains only very minute quantities of hydrocarbons. Storage in insulated steel tanks with steam-blanketing provisions for fire suppression is standard, but concrete-lined pits are sometimes used. Tanks and molten sulfur piping are heated by steam controlled to a saturation temperature of approximately 290^oF at which the sulfur has ideal flow characteristics. Steam tracing is strongly preferred because of the ease with which controllable highly uniform temperatures can be maintained throughout the system. Sulfur metering pumps are supplied in duplicate so that one may be serviced while the other is in operation. Even though extensive precautions are taken to maintain cleanliness of the sulfur and piping, it is impossible to eliminate all possible chances of contamination and therefore impossible to guarantee continuous unimpeded pump operation over long periods of time. Inasmuch as emission compliance often depends upon continuous operation of the flue gas conditioning system, it is usually considered that the cost of an installed spare pump is justified.

Combustion of the sulfur to generate SO₂ is obtained by introducing sulfur into an air stream which has been electrically preheated for startup purposes to the temperature at which the catalyst becomes active. Since this temperature exceeds the autoignition point of molten sulfur, burning is initiated immediately upon the introduction of sulfur and SO₂ is delivered to the catalyst for conversion to SO₃. Roughly 4000 BTU's of heat are generated per pound of sulfur burned. This replaces a portion of the startup electrical heat input, and at full system rating all the required heat to maintain the catalyst at operating temperature is supplied by sulfur combustion. Clearly, operation of the system in this manner allows the generation of any quantity of SO₃ from zero to full system rating as a function of any selected control signal used to determine the rate at which sulfur is delivered.

Prior to the last couple of years controls for SO₃ FGC installations consisted of safety and sequencing devices plus load-following apparatus to modulate the rate of delivery of SO₃ in proportion to boiler load. The level of SO₃ to suit a given coal was determined experimentally by the operator by manual adjustment of a proportioning factor between the boiler load signal and the SO₃ generation rate while observing the effect on ESP operation and emissions. It was generally found that the setting for a given coal, once determined, was seldom subject to much variation so it was possible for most stations to make initial determinations for the various coals they fired and use those values with little need for subsequent adjustments. These control functions were implemented with relay logic and analog controls, and permitted operation with a minimum of attention from station personnel. They were reasonably satisfactory for most applications.

There were, however, instances (especially overseas) where the coal being burned at any given time might be unidentified or mis-identified so that previous history was of no value in determining the injection rate setting and there were some stations burning "spot" coals received in small quantities where constant determination of new settings was required. There were also locations where difficulty was encountered in determining optimum settings through unfamiliarity with ESP operation. For these reasons, as well as the desirability of freeing the operators from the task of identifying which coal was being fed to the burners, and especially the times at which a particular coal started and ended, effort was directed to automatic determination of the optimum SO₃ rate. A first attempt was based on the idea that the rate would be a function of the sulfur content of the coal, and that quantity could be inferred from measurement of flue gas SO₂. This proved to be partially workable, but did not adequately account for ash chemistry or ash quantity differences between coals with the same sulfur contents. Another aborted attempt relied on the increase in ESP power input accompanying SO₃ injection, but seemed to have an unbreakable habit of running to maximum system rating whether that injection rate was optimum or not. Excess injection of SO₃ will tend to make the ash resistivity in the ESP too low, decreasing its collection efficiency and causing severe rapping puffs. If SO₃ injection is ahead of the air preheater acid condensation and corrosion may occur in the cold end baskets.

Controls are now available implemented with solid state apparatus which digitally store the operating characteristics of the ESP as a function of the injection rate setting and, for any coal being fired, determine the characteristic in use and proceed to the optimum point of that curve. Periodic perturbations of the setting are made to determine that the set point remains optimum. If not, a new optimum is determined, and the new characteristic is stored for future reference. Besides freeing the operator from necessity for closely monitoring the fuel and adjusting the set point, these controls retain a history of operation which can be retreived for information on control and system performance.

Those who have studied factors affecting the resistivity of fly ash in a flue gas atmosphere are very familiar with the hump-backed characteristic of resistivity as a function of temperature. Generally, resistivity has a maximum value at a temperature around 300^oF with decreasing values above and below the maximum point as illustrated by Figure 2. In determining the amount of SO₃ to be injected to reduce the ash resistivity to a desired lower value, one would think that the maximum rate would be required at the maximum unconditioned resistivity point, with decreasing amounts at higher and lower temperatures. That is, a hump-backed curve of injection rate reflecting the unconditioned resistivity characteristic would be expected. An interesting discovery made in the development of SO₃ flue gas conditioning is that the amount of SO₃ required to attain a desired level of resistivity follows the expected dome-shaped curve with respect to temperature only up to a point, after which it breaks off to a rapidly rising characteristic as seen in Figure 3. The inflection point between the two portions of the curve is a function of the surface chemistry of the ash, occurring at relatively low temperatures for acidic ashes and at higher temperatures for basic ashes. The range of variation of the inflection point temperature appears to be approximately from 250^oF to 400^oF for coals available world-wide. If the flue gas temperature is above the inflection point, the portion of the injected SO₃ which is greater than the level of the dome-shaped portion of the curve does not attach to the ash and will be passed through the ESP. In the rare cases where a wet scrubbing system follows the ESP the excess SO₃ will be captured; otherwise an objectionable blue plume will be formed if the emitted SO₃ concentration approaches 10 ppmv or more. In either case SO₃ and the feedstock used to produce it are being wasted. Of course, this problem may be avoided by changing to a coal which produces an ash having its inflection point temperature above the flue gas temperature, or by reducing the flue gas temperature to a value below the inflection point. If neither of these is possible, as is often the case, it has been found that the simultaneous injection of ammonia (NH₃) with the SO₃ will condition the ash surface to accept resistivity adjustment without excess SO₃ being required. As an example, at the Hudson Station of Public Service Electric & Gas of New Jersey where operating temperatures in portions of the ESP exceeded the inflection point, injection of 15 ppmv of SO₃ resulted in particulate emission compliance, but an objectionable blue SO₃ plume was highly visible. Introduction of 4 ppmv of NH₃ with the SO₃ level reduced to 8 ppmv eliminated the blue plume and materially improved particulate collection.

The catalyst for conversion of SO₂ to SO₃ may be chosen from any of the types developed for the manufacture of sulfuric acid and similar applications. Vanadium pentoxide is the active ingredient in most of these, and is classified as a hazardous material. It appears that catalyst life in this service is on the order of ten years so that frequent handling is not required, but prevention of dust dispersion and ingestion of or contact with the material, retention of the material in sealed containers, and disposal by qualified handlers are necessary. In spite of its hazardous classification, servicing of the catalyst bed is neither particularly difficult nor expensive, but it is nevertheless desirable to do it as infrequently as possible. To this end care should be taken to exclude dust and water from the air intake to the maximum extent possible, and to service the air intake filter on a regular basis. Dust entering the system will tend to plug the inflow end of the catalyst bed and decrease the available air flow, eventually causing overtemperature tripouts. Water has the effect of breaking down the physical structure of the catalyst pellets, so that interstices in the bed become plugged and air flow is restricted. Air intake locations which might be subject to overspray from cleaning hoses or windblown rain, for instance, should be suitably shielded or relocated to an absolutely dry position.

The combustion air stream bearing the generated SO₃ will issue from the catalyst at temperatures from 750^oF to about 1000^oF, depending on the rate of SO₃ production. It is essential that this stream be held above its acid dew point temperature all the way through the delivery manifold and injection probes until it actually issues into the flue gas. If the temperature goes below the dew point, the acid will condense out in the manifold or the injection probes. This is undesirable for two reasons. First, none of the acid will reach the flue gas to do the intended conditioning job, and, secondly, the condensed acid will corrode the piping and nozzles. Since the SO₃ concentration is a few percent in this stream the acid dew point will be on the order of 460^oF, but maintaining the delivery end of the system in a condensation-free state requires that the calculated gas temperature as it issues from the injection probe nozzles not be less than 500^oF. For this reason the distribution manifolds are heavily insulated, and in addition the injection probes, if installed on the cold side of the air preheater, are thermally insulated from the flue gas. Also, calculations of the expected temperatures at the most distant (or coldest) injection nozzle are made for every cold side installation.

It is probably obvious that avoidance of acid condensation in the manifold and probe areas requires preheating of the system with hot air when being started up from cold. It is not quite so obvious that a hot purge of the system during shutdown is required to remove sulfur products from the catalyst so that acid will not be present in the warm-up gases at the next startup. When proper startup, operating and shutdown procedures are observed as described here, long manifold and probe life is obtained. Some early systems are approaching 20 years on line with original probe and manifold parts still in service. Steel sheathing over the insulation of cold-side probes is subject to the same erosive fly ash attack as other internal parts of the ductwork, and requires similar periodic maintenance and repair.

The length of travel of the flue gas in the ductwork after the injection point required to provide essentially complete mixing with the SO₃ is often glibly cited as "one second mixing time" or "ten times the nozzle spacing". These rules of thumb are derived from experiments showing that in turbulent flue gas flow complete mixing with another gas injected through a bank of nozzles arrayed as a uniformly spaced grid occurred at a distance downstream of the grid equal to about eight to ten times the nozzle spacing. In FGC systems a nominal grid spacing of three feet has been found to be a reasonable compromise. Ten times this spacing gives a mixing distance of 30 feet, which at typical duct velocities of 60 ft/sec allows only one-half second mixing time, and this is adequate providing you have a straight run of uniform size that long. Ordinarily you don't, hence the conservative quest for one second. Expansions, contractions and bends in the ductwork interfere with neat mixing patterns. If the ESP ductwork is being modeled for flow visualization it is desirable to include an examination of the mixing behavior which may be expected. Modification of the spacing, location and gas delivery rate of the nozzles are steps which may be taken to cope with less-than-optimum mixing situations.

Because installation of injection probes is ordinarily more easily accomplished in the ESP-type ductwork on the cold side of the air preheater than in the boiler-type construction on the hot side, most of the present FGC installations inject on the cold side. It should be noted, however, that hot side installation has the advantages of lacking any close approach to acid condensation temperatures in the probes, and provides excellent mixing and contact between the SO₃ and the fly ash as it passes through the air preheater. Hence, cases where long manifold runs may make it difficult to maintain sufficiently high probe temperatures or where there is little cold side gas travel length for mixing of the injected SO₃ with the flue gas favor location of the injection probes on the hot side. In determining the desirability of putting the probes on the hot side possible SO₃ condensation in the cold side of the air preheater must be considered. This is not usually a problem because the injected SO₃ quantity establishes air preheater conditions roughly equivalent to those expected from an unconditioned boiler firing coal containing 1.5 percent sulfur. If the average cold end temperature and basket material of the air preheater are compatible with that assumption hot side injection would ordinarily be acceptable.

Probes for hot side installation are less expensive than cold side because no thermal insulation is required and the erosion-protective outer sheathing can usually be eliminated. The probe insertion point into the boiler on the hot side may move more than a cold side point when the boiler is started up and shut down, and this can add to the difficulty of design and the expense of the manifold system, possibly offsetting savings in probe costs. As far as operation is concerned, every existing hot side installation works as well as or better than equivalent cold side units.

Note that the foregoing applies to injection of SO₃. Injection of ammonia, to be discussed later, is ordinarily not allowed on the hot side of the air preheater because of the tendency to plug the preheater with ammonium-sulfate compounds unless there is little or no SO₃ in the boiler gas.

On the face of it, it seems like a good idea, considering the relatively high cost of the SO₃ generators, to use a single such generator to feed conditioning gas to several boilers. The primary objection to such an arrangement is that, assuming operation of the FGC system is necessary for compliance with emmission limits, all the boilers served by whatever portion of the system is common must be taken off line simultaneously before the common portion can be shut down for service. The high availability figures compiled by existing installations are based on periodic (at least, annual) boiler outages during which accumulated maintenance requirements of the conditioner system can be met.

If, however, measures can be taken to maintain emission compliance with the FGC system off line or if simultaneous removal from service of all boilers on the common system is not objectionable, then a portion or all of the SO₃ generator may be made common to more than one boiler. The most straightforward way to do this is by making the blower and air heating apparatus common and providing individual burners (if used), catalytic converters, manifolds and injection systems for each boiler. Control of such an arrangement is uncomplicated, and manifolding and injection requirements are identical to those for single boilers. The possible alternate arrangement where the burner and catalyst are also made common appears to offer cost savings, but is actually difficult to implement unless simplifying assumptions such as operating all common boilers at the same output level, etc., can be made. To retain full flexibility in boiler operation it is necessary to provide valving in the manifold system. Carefull heat tracing of the valves and the portions of the piping which may become dead-ended while still exposed to SO₃-containing gases is required to prevent acid condensation. Cold spots in valve bonnets are not easily avoided. Acid-resistant valves are expensive. Another less severe problem with this system concept is that fully flexible operation of the individual boilers requires a relatively complicated control system. In general, this approach is not recommended.

Ideally, the amount of conditioning gas delivered to any single injection nozzle in the flue gas stream should be proportioned to the amount of ash passing through that nozzle's treatment area and adjusted for the flue gas temperature at that point. It is entirely possible to make comprehensive sets of measurements which determine these values, at least for one set of operating conditions, and it is also possible to divide the delivery of conditioning gas among the injection nozzles in accordance with such measurements. In practice, a full scientific treatment of this type is never done, and is rarely even approximated. The reason, aside from the variability of the target conditions, is that ESP's have a considerable degree of built-in tolerance for random variations. Witness to this fact is provided by many existing installations where flue gas conditioning was neither needed nor used. It is common in these units to find that a wide spread of temperatures across the air preheater outlets is passed through to the ESP inlet along with non-uniform dust concentrations, in spite of which the ESP exhibits excellent operation. In general, inlet conditions which meet industry criteria for satisfactory ESP operation allow treatment of a flue gas conditioning system on the basis of uniform temperature and flow distribution. Another way to state this is that situations which could require special precautions for distribution of conditioning gases usually also require modifications to permit satisfactory ESP operation, and incorporation of the latter will usually eliminate the former.

The burner and catalyst chambers of FGC systems are lined with refractory materials which can be damaged by thermal shocks. Recommended procedures for starting a system from cold conditions require a gradual warm-up procedure over several hours. Some modern boilers are capable of cycling from stand-by to full load in as little as 45 minutes. To permit the FGC system to follow this type of load change, it is held in stand-by condition with internal temperatures maintained at operating levels by the startup heat source. For economy, air flow rates in stand-by may be reduced from normal operation, thereby reducing heat input. The FGC unit is capable of going from stand-by to full output in 30 minutes, and can therefore accomodate the requirements of boiler cycling. Because of the length of time required for a cold start it is normal practice to hold the FGC system in stand-by when the boiler is in a stand-by or banked-fire condition, and not go to a cold shutdown unless the boiler will be off-line for at least a few days.

During most of the history of ESP collection of fly ash it was considered that use of the full available output of the high voltage supplies indicated good operation and high efficiency collection, and this was highly desirable. In later years attempts to collect high resistivity dust without resort to flue gas conditioning lead to the construction of ESP's of monumental size, complete with row on row of hoppers, high voltage power supplies and ash handling apparatus. Whether or not these passed their guarantee test requirements, quite a few were eventually fitted with FGC which, among other things, promptly caused all the power supplies to go to full rated output. What once had been considered good ESP operation was now rightly considered outrageous consumption of station auxiliary power. Fortunately there is a cure for this situation.

Control units for the high voltage power supplies are now available from most ESP manufacturers which provide a feature called "intermittent energization" or "skip-cycle control". Although the claim that this type of control materially improves collection of high resistivity dust has proven to be somewhat optimistic, there is a definite economic justification for its use in conjunction with FGC. Once the fly ash resistivity has been adjusted by FGC to an optimum value, high collection rates may be maintained while ESP power consumption is greatly reduced by use of the "skip-cycle" feature. Power reductions to levels on the order of 20 to 30 percent of the unmodified power input have been obtained without noticeable increases in emissions.

In an effort to reduce the cost of SO₃ conditioning an experimental system which converts a portion of the SO₂ in the flue gas to SO₃ is being investigated as a possible way to avoid the purchase of sulfur or liquid SO₂ feedstock, but questions yet to be thoroughly investigated include control of SO₃ quantities to proper levels when a variety of different coals are burned producing ashes with differing conditioning requirements. Also, variation of flue gas temperatures as boiler loads vary may lead to insufficient production of SO₃ at low loads, causing fouling of the ESP with high resistivity ash which may interfere with emission compliance upon return to full load operation. For economy of installation and simplicity of operation it is preferred that booster fans not be used, which means that the system relies on the relatively small pressure drop in the main system to provide the driving force for the side loop in which the catalytic conversion of SO₂ to SO₃ occurs. This leads to large injection probe sizes and difficulty in obtaining uniform distribution and mixing of SO₃ in the downstream gas. The expense of custom ductwork design and installation for such a system is not negligible. It is expected that a considerable amount of testing and demonstration work will be required to accurately assess the suitability of this system for commercial operation. Development of another system proposed earlier in which ammonium sulfate was to be decomposed in hot boiler flue gas and the resulting SO₃ used for ESP conditioning is said to have been abandoned.

Simultaneous injection of SO₃ and NH₃, arbitrarily called dual conditioning, was mentioned earlier as assisting in obtaining attachment of SO₃ to fly ash at temperatures above the inflection point of the injection rate curve. Dual conditioning has been found to have other uses as well. It has been demonstrated that it may be used to control rapping and reentrainment losses from ESP's. It appears that the process can be crudely visualized as reacting the two conditioning gases to form ammonium bisulfate which has a melting point close to typical ESP operating temperatures. This is thought to act as a binding agent (liquid glue?), increasing the cohesivity of the ash particles so that they adhere to each other and fall into the hopper with reduced dispersion into the gas stream. A striking demonstration of this effect occurred at the Monroe Station of Detroit Edison where SO₃ alone permitted operation only up to 450 Mw before the compliance limit on particulate emission was reached, but dual conditioning allowed the full 750 Mw unit rating to be attained. In tests conducted by Ontario Hydro it was found that dual conditioning was sufficiently effective to obtain significant reduction of losses to the stack caused by high unburned carbon levels in the ash.

Some experimenters in the past found that the injection of NH₃ into flue gas streams containing appreciable quantities of SO₃ caused the ash to become "sticky" enough to cause problems in ash handling systems. Experience with dual conditioning to date indicates that this problem arises when relatively large quantities of NH₃ and SO₃ are present as compared to the amount of ash. The quantities required to provide effective prevention of excessive rapping and reentrainment losses are controlled to much lower values. None of the dual conditioning installations thus far (14,854 Mw) have had problems with ash handling from increased ash cohesivity.

From the preceding description of the inflection point which appears in the injection-rate-versus-temperature curves it is clear that one criterion for needing dual conditioning is flue gas temperature exceeding the inflection point temperature. Criteria for advance determination of a need for dual conditioning to suppress excess rapping and reentrainment losses by means of fuel and ESP data have not yet been developed, partly because the factors affecting ash cohesivity in the ESP have not been explored or defined, and partly because the mechanical and fluid dynamic characteristics of the ESP which affect the magnitude of rapping and reentrainment losses may differ radically from one installation to the next and have not been subjected to any investigative procedures capable of developing applicable loss factors. It seems odd that such a significant ESP loss mechanism should be so poorly characterized. It is expected that continued research will help to clarify this situation in the future. In the meantime the only recourse is to observation of trials with fuel, SO₃ and ESP conditions as nearly as possible the same as expected in service. Then, observation of the difference between recordings of flue gas opacity with and without rapping will immediately indicate, if the difference is more than a few percent and large rapping spikes appear on the instantaneous trace, that dual conditioning should be considered.

It has been observed that introduction of SO₃ and NH₃ into gas streams containing fly ash significantly affects the physical composition of collected layers of the material, whether the collection is by mechanical or electrostatic methods. It is not surprising then, that dual conditioning has been found to have marked effects on the operation of fabric filters. Pilot scale tests conducted at the University of North Dakota Energy and Environmental Research Center and field tests at the Monticello Station of TU Electric have shown that for a number of different coals with several different fabrics and several different cleaning modes that fine particulate emissions and baghouse pressure drop are greatly reduced by use of dual conditioning. In the Monticello tests the filter drag without conditioning was more than twice as great as that with conditioning. Inasmuch as the initial Monticello results confirmed laboratory data, it is expected that the pilot plant investigations of other ashes are reliable predictors of short-term field performance. Longer-term data from Monticello and eslewhere is needed to make sure no objectionable side effects exist. If this proves to be the case, dual conditioning should take a place in the future as a useful technique for control of baghouse performance.

The addition of SO₃ alone to boiler flue gas streams only occurs if the coal is low in sulfur, as otherwise the ash resistivity will not be high enough to require conditioning. Nearly all of the injected SO₃ is attached to the ash particles and is captured with them in the ESP. Without conditioning one to one-and-one-half percent sulfur content in the coal would typically give the same ash resistivity as that obtained by injection, meaning that the same amount of SO₃ is on the ash surface, but an additional amount would still be present in the gas stream and would be emitted from the stack. This latter amount, typically about one-half percent of the SO₂ content of the gas stream, is not present in the emissions from the conditioned unit unless operation at temperatures above the inflection point of the injection rate curve is attempted without the simultaneous injection of ammonia.

Fly ash from a conditioned system carrying acid on its surface in about the same concentration as that from unconditioned combustion of one-plus percent sulfur coal is not significantly different in handling requirements and environmental effects from the ashes dealt with in the past.

Ash handling systems which sluice SO₃-conditioned ash to a pond and recirculate the water may find that the pH of the pond will be depressed below acceptable levels for discharge to the environment, making neutralization treatment necessary.

Ammonia treatment of fly ash may cause release of objectionable odors from concrete made with cement in which the fly ash has been used as an additive. Determination of the maximum acceptable ammonia content for fly ash in this use is a trial-and-error process. The level has been found to be greater than zero in all cases to date, but the maximum appears to be influenced by cement and ash chemistry and ash handling conditions.

Ammonia and sulfur as used by conditioning systems are essentially pure chemicals which neither introduce nor generate any hazardous or toxic pollutants. Except for improved control of particulate emissions the overall impact of the boiler system on the environment is essentially unchanged by the use of flue gas conditioning.

It is reasonable to suppose that dual conditioning of both ESP's and fabric filters will develop through trial and error and gradual accumulation of pilot and full scale data in much the same way that SO₃ progressed over the past several years. Pilot scale work is under way at the present time to find out whether the beneficial short-term effects of dual conditioning on the performance of reverse-air and shake-deflate fabric filters also apply to pulse-cleaned designs. Full-scale long-term trials remain to be performed to validate dual conditioning of fabric filters as a commercial option. The demonstrated capability of conditioning to greatly improve baghouse collection of fine particles will likely become more important with increasingly severe restrictions on fine-particle emission.

There is a need for a great deal of basic work to establish a sound technical basis for the estimation and control of rapping and reentrainment losses from ESP's. Methods of measurement of dust cohesivity in the ESP environment, determination of the combination of dust cohesivity, rapping impact and plate height to attain minimum losses and design of ESP structures and operating schedules which best provide optimum cost and performance all need to be explored. ESP research, fueled in large part by EPA and EPRI, has made major progress during the past 20 years in substituting engineering for art. It appears that a complete understanding of the rapping/reentrainment process is the only significant remaining grey area.

Over the long run dual conditioning will progress to a position of predictability and acceptance now, at long last, enjoyed sulfur trioxide conditioning for adjustment of fly ash resistivity in electrostatic precipitators.

In connection with flue gas conditioning at Monroe Station of the Detroit Edison Company mentioned earlier in this article it was found possible to develop specifications for fuel procurement taking advantage of the ability of the generating units to compensate for variations in coal and ash analysis. This was expected to simplify obtaining and evaluating fuel bids, and to avoid compliance problems by use of relatively precise knowledge of equipment and fuel limits. It would appear that there is a future possibility of using on-line coal and ash analysis capability in combination with economic analysis to optimize both fuel allocation to units and generation costs. A utility using a combination of generating units having flue gas conditioning without scrubbing, other units with scrubbing, and some purchased emission allowances, for instance, could very well be in position where optimum allocation of both fuel and load would make an appreciable difference in costs and emissions.

In ESP operation dust collecting performance is often adversely affected by resistivity effects or by rapping and

reentrainment. Problems of this type reduce the performance level of the equipment compared to that which it is capable of attaining if the dust to be collected is treated to permit optimum operation of the collecting mechanisms. In many cases maximum economy may be attained by application of flue gas conditioning to an appropriately sized dust collector having fewer hoppers, less acreage, less power and much less maintenance than conventional units handicapped by being required to collect dusts which have not been properly prepared for efficient collection. Your automobile won't run well if you put crude oil in the tank - your ESP won't run well on high resistivity dust. Advantages obtainable by use of flue gas conditioning will permit better emissions control with smaller, more reliable, less expensive ESP installations. There is a possibility, yet to be fully confirmed, that fabric filter performance may be similarly subject to improvement in certain cases.

#####