Scrubbing coal for emission control.

Scrubbing Coal For Emission Control

The demand for electric power in the United States will grow at an annual rate of 2.4 percent throughout the 1990s, according to government forecasters. With coal already comprising 80 percent of our fossil fuel resources and coal-fired facilities supplying

57 percent of our electric power, it is safe to assume that coal will remain a viable option for meeting this demand well into the next century. At the same time, electric utilities are likely to have to comply with increasingly stringent emissions standards. As in the early 1970s, when passage of the Clean Air Act caused research on pollution control technologies to accelerate, electric utilities lately have been stepping up their search for effective and economical ways of further reducing emissions.

A portion of this work is directed towards refining a system for cleaning combustion gases known as the wet scrubber - a processing facility built at the back of the main power plant that traps sulfur dioxide before the flue gas reaches the atmosphere. Although the process underlying wet scrubbing was developed just after the turn of the century, the first devices were not built until the 1930s in Great Britain.

Most of the scrubbers in use today are based on concepts worked out during the 1960s. In the United States, the first full-scale scrubber began operating in 1967 at a coal-fired power plant owned by Union Electric of Missouri. Since then, utilities, suppliers, architectural and engineering firms, and the federal government have taken part in developing the technology.

The earliest scrubbers were dogged by numerous problems, such as corrosion and plugging, whose persistence reflected the slow development of the technology. And though many of these operational problems were eventually solved, most of today's scrubbers remain expensive and difficult to operate and maintain. For 90 percent of scrubber operations in the United States, handling and disposing of the waste product - a wet, pasty sludge - is a costly and complex undertaking. But above all, there is the question of efficiency: running a scrubber system typically drains away between 3 and 5 percent of the power produced by a generating plant.

Scrubbers come in several varieties, differentiated by three major characteristics: whether the process results in a wet or a dry product, whether it employs a slurry or a solution to absorb the [SO.sub.2], and whether it produces a salable or a throwaway end product. Whatever the type, scrubbers belong to a technology known as flue-gas desulfurization (FGD), the traditional method by which pollutants are removed from the coal after combustion. In the past few years, refinements in FGD have resulted in more efficient and reliable scrubber designs that remove more [SO.sub.2] at lower cost and produce a salable, or at least a more easily disposable, end product.

It was to encourage the development of advanced FGD and other techniques that the Department of Energy initiated the Clean Coal Technology Program in 1986. This cost-shared effort by government and industry to demonstrate innovative coal-burning processes at a series of full-scale facilities around the country is expected to finance more than $5 billion in projects before it is completed later in the decade. Under the program, the federal government will provide up to 50 percent of the total cost of the demonstration projects. In the first two rounds of solicitation for proposals, the DOE selected 29 projects for funding. In the second round, held in the summer of 1988, seven of the 16 successful proposals involved the use of both wet and dry scrubber systems. These projects included two full-scale demonstrations of a limestone wet scrubber that produces gypsum as the end product, and a unique kiln-dust wet scrubber that produces potassium sulfate and distilled water as byproducts.

Lime and Limestone Systems

Ninety percent of the scrubbers currently in operation at coal-burning facilities in the United States are wet scrubbers, the type that produces a wet end product. Of these, between 60 and 70 percent use a lime or limestone slurry to absorb the [SO.sub.2]. (Spray dryer scrubbers that produce a dry end product have been developed, but have been most effective on plants burning low-sulfur coal.) In general, lime and limestone systems have lower capital, operating, and maintenance costs than other wet scrubbers. Widely available reagents, minimal equipment requirements, and comparatively low energy requirements have kept costs down compared with other systems. In addition, because wet scrubbers were the first systems to be developed and are the most widely used, a large body of operating experience has led to improved designs. The most advanced wet scrubbers can now be expected to remove up to 90 percent of [SO.sub.2] from the flue gas stream. Consequently, they are the type usually chosen by utilities that burn high-sulfur coal.

The process underlying the operation of lime and limestone systems is basically the same as that of other scrubbing systems. [SO.sub.2] is oxidized to form sulfite and sulfate molecules inside an absorption vessel. In conventional scrubbers, the absorbent, a mixture of water and a chemical reagent, is sprayed down from nozzles set into the sides and across the top of the absorber. The flue gas enters the bottom of the absorber and rises through it. Most of the [SO.sub.2] is absorbed in midair by the droplets of absorbent. Fresh absorbent is continually pumped into the system, while expended absorbent drains into a holding tank where solids form from the reacted [SO.sub.2].

Meanwhile, the clean flue gas continues to rise. Water droplets caught in it are removed at the top of the absorber. Since the flue gas cools as it passes through, some systems reheat it in order to minimize condensation on the sides of the flue and prevent corrosion of the liner. Finally, the treated gas is released to the atmosphere.

In conventional systems, the used absorbent and the solids formed from the reacted [SO.sub.2] are disposed of as waste. In lime and limestone systems, the reagent is not completely water soluble, and the result of mixing it with water is an absorbent slurry. This waste sludge, made up of calcium sulfate, calcium sulfite, unreacted lime or limestone, and water, can leach into groundwater unless it is impounded or treated before being used as landfill. The quantities produced are enormous. Over the course of its lifetime, a 500 MW coalfired power plant equipped with conventional wet scrubbers will produce enough sludge to fill a 500-acre disposal pond to a depth of 40 feet.

To a large degree, the consistency of a scrubber's performance depends on how well the utility operates and maintains it. In the absence of careful supervision, equipment can become plugged with solids deposited in nozzles, ducts, and drains, and surfaces can become corroded by the precipitation of thin layers of hard solids called scale. Utilities also can help improve both burner and scrubber performance by insuring that only coal of a specified sulfur content is burned.

Finally, there also are cost considerations. Wet scrubbing systems require large quantities of water, and since the reagent is used only once, replacement costs can add up to a considerable economic burden.

A variation on the traditional wet scrubber is one that produces gypsum as an end product. In this process, calcium sulfite, which is a waste product in lime and limestone systems, is oxidized to form calcium sulfate, or gypsum. This oxidation step also provides the considerable advantage of eliminating problems with plugging and scale.

Single Module

This spring, a joint venture company called Pure Air (Allentown, Pa.), formed by Air Products and Chemicals Inc. and Mitsubishi Heavy Industries America Inc., began construction of a gypsum-producing wet scrubber system at a Northern Indiana Public Service Co. coal-burning facility near Gary, Ind. At a cost of about $150 million, the Bailly Station retrofit is the most expensive of the seven flue-gas desulfurization projects co-funded by the DOE under the Clean Coal Program. Its purpose is to demonstrate that an advanced FGD system, in combination with sale of the gypsum byproduct and efficient plant operation and maintenance, can reduce [SO.sub.2] emissions by 90 percent at approximately half the life-cycle cost of a conventional system. In addition, the system will generate no solid waste and will minimize liquid waste.

Although single scrubbing units of up to 700 MW have been installed in Japan, most systems in the United States employ absorber modules capable of handling the gas stream from coal-burning units with a capacity of no more than about 125 MW. Large boilers or power plants consisting of a number of smaller boilers use multiple modules. Hence, when scrubber systems are installed at larger plants, the use of these multiple modules tends to increase both capital and operating costs.

The scrubber at Bailly will process the combined flue gases from two boilers rated for a total of 528 MW, making it the largest-capacity module in the United States. The plant, which currently burns coal with a sulfur content of 3.1 percent, will test coals of 2 to 4.5 percent sulfur throughout the demonstration period. The system is expected to reduce sulfur dioxide emissions at Bailly by 90 percent; energy to run the scrubber will consume about 1.5 percent of plant output.

Because the New Source Performance Standards (NSPS) of the Clean Air Act require that an FGD system operate whenever fuel is being burned, most utilities in the United States have opted to incorporate spare modules to increase reliability. Nevertheless, with today's designs, 98 percent reliability can easily be obtained by larger, more cost-efficient modules, and operation without a spare is common outside the United States. Because the project at Bailly is a retrofit, as are most installations that employ advanced forms of FGD, the NSPS do not apply. The spare module will therefore be eliminated to reduce costs.

Single Loop, Single Vessel

Pure Air's project at Bailly will be based on Mitsubishi's basic wet limestone FGD process, but it will demonstrate several advanced features. In addition to employing a single module, the system cuts costs by consolidating the equipment used in [SO.sub.2] removal and gypsum production.

In a wet limestone scrubber, [SO.sub.2] reacts with limestone to produce calcium sulfite, which may then be oxidized to form usable gypsum. In conventional gypsum-producing systems, the oxidation of calcium sulfite takes place in a vessel separate from the absorber. This is known as ex situ oxidation. Another common feature of conventional wet scrubbers is the prequencher, which removes chlorides and particulates from the flue gas before they enter the absorber. This prequencher/absorber arrangement is referred to as a dual loop.

The combination of dual-loop scrubbing and ex situ oxidation means that for most gypsum-producing systems, separate reaction vessels for absorption, quenching, and oxidation are necessary. In the Pure Air system, oxidation of calcium sulfite and absorption of [SO.sub.2] occur in a single vessel. This is called in situ oxidation. In addition, the prequencher is eliminated so that particulate and [SO.sub.2] removal take place simultaneously in the absorber. In this so-called single-loop process, the flue gas is quenched with a recirculating limestone slurry as it enters the top of the absorber. An intermittent spray of fresh water prevents the formation of deposits.

While conventional wet scrubbers use a countercurrent flow, in which the flue gas rising from the base of the scrubber passes upward through the falling slurry spray, the Pure Air version uses a high-velocity, co-current flow. In this design, both gas and scrubbing slurry are introduced at the top and pass downward through the scrubber. The pipes in the absorber tower are arranged in an open grid. By providing a large surface area for liquid/gas contact and more uniform distribution of flue gases, the design helps improve overall [SO.sub.2] removal efficiency.

The advantage of co-current flow is that it permits a significant increase in gas velocity. Judging from work done at a pilot plant in Japan, the design at Bailly should result in velocities as high as 20 ft/sec, compared to about 10 ft/sec for counter-current flow. As a result, the absorber can be half the size of currently available FGD absorbers and still handle the same amount of flue gas. This can be an especially important advantage in retrofits such as Bailly, where space is limited.

In the United States, co-current flow has been used at the Tennessee Valley Authority's 10 MW Shawnee Plant and at Hoosier Energy's 125 MW Merom Station. Worldwide, co-current flow scrubbers treat the flue gas from power plants whose combined capacity totals 12,000 MW. However, the gas velocities at these plants are significantly lower than the 20 ft/sec proposed for Bailly.

Gypsum, Gas, and Water

When oxidation is carried out in a separate vessel, fixed spargers are usually employed. A large array of these submerged, perforated pipes distributes the air, and mechanical mixers or pumps recirculate the slurry. At Bailly, another new feature to be demonstrated is an air rotary sparger, designed to combine the functions of oxidation air distribution and mixing in the absorber (some fixed sparging will also be used to insure complete oxidation). The device has been tested at a 50 MW Two absorber bleed pumps will transfer the slurry to the centrifuge feed tank for further processing. Basket centrifuges will reduce the slurry to a dewatered gypsum cake containing 8-10 percent moisture by weight. The gypsum can then be placed iin storage to await either sale or burial in a landfill.

The flue gas, meanwhile, after falling through the absorption grid, will turn, pass over the reaction tank, and proceed upwards towards a mist eliminator located in the outlet ducting. Here, entrained fluid will be separated from the gas and returned to the absorber tank for recirculation. A spray system installed in front of the mist eliminator will intermittently wash down its surfaces to prevent buildup of deposits. After passing through the mist eliminator, the scrubbed flue gas will continue through the outlet duct into the exhaust stack for discharge into the atmosphere.

In conventional systems, any wastewater from the centrifuge operation that cannot be recycled back to the process must be sent to a disposal pond or treatment plant before being discharged. This increases operating expenses and can be especially costly for plants burning any of the many high-chloride U.S. coals. At Bailly, a portion of the filtrate water from the centrifuge operation will be returned to the absorber vessel as process water. In addition, a system that has been demonstrated on a 125-MW, oil-fired boiler in Japan will enable waste heat in the flue gas to evaporate the remainder.

First, the wastewater will be fed into a pH-adjustment tank, where it will be neutralized by hydrated lime. As a result, impurities such as chloride and sulfate ions will be stabilized and prevented from evaporating. The wastewater will then be pressurized and pumped to evaporators located in the flue gas ducting and fans section at the beginning of the process. It will be atomized there by a pressure nozzle in the duct, mixed with the flue gas, and dried through evaporation. Flue gas containing solids will then join the main gas stream at the evaporator outlet and the dry solids will be removed by an electrostatic precipitator. If the wastewater evaporation system functions as planned, the Bailly plant will generate no liquid waste for treatment and disposal.

In Japan and West Germany, suppliers of limestone in the powdered form required for scrubbing have proliferated. In the United States, on the other hand, the limestone is generally milled at the facilities themselves. Whether it is cheaper to purchase limestone from a supplier or grind the material on-site depends on the plant's size and location. When space is limited, as it frequently is in the case of retrofits, eliminating the milling operation can be an advantage. At Bailly, the FGD system will use pulverized limestone purchased from a supplier.

By far the largest consumer of gypsum is the wallboard industry, followed by the cement industry. In Japan and Germany, where natural reserves are scarce, virtually all FGD systems produce gypsum, and 100 percent of it is sold. In the United States, the market for gypsum is about 30 million tons a year, of which about 20 million tons go to wallboard manufacturers. It is anticipated that the gypsum produced at Bailly will be sold to a single wallboard company, supplying about half the manufacturer's yearly demand for the material, or about 150,000 tons.

Managing the System

The gypsum-producing wet limestone system that will be demonstrated at Bailly includes many components typical of today's designs. The equipment has been tested in pilot plants or employed in commercial operations at low-sulfur coal and oil-burning facilities similar to Bailly. In this project, however, many of the system's features - the single, large-capacity absorber module and single-loop, ex situ oxidation - will be combined for the first time in the United States. Similarly, the high-velocity, co-current absorber and the waste evaporation system both represent comparatively new technologies. Under the circumstances, it will be especially important to insure that the facility is operated as carefully as possible.

When a utility decides to install a scrubber system at one of its plants, the usual practice is to contract with several firms to design and build the equipment. The utility is then left to operate on its own a pollution control system notorious for its dependence on highly trained operators and continual maintenance.

The project at Bailly is taking place under a novel business arrangement devised to relieve the utility of this burden. Under the contract negotiated by Northern Indiana Public Service and Pure Air, the latter is engineering and constructing the scrubber system, which it will then operate during the demonstration period. After the project ends in 1995, Pure Air will continue to own and operate the facility for an additional 17 years.

Recovery Scrubber

A scrubber of a quite different sort, developed by Passamaquoddy Technology (Thomaston, Me.), will be demonstrated under the Clean Coal Program at the Dragon Products Co. cement plant (formerly owned by the Passamaquoddy tribe) in Thomaston. At a cost of about $10 million, it is the least expensive of the advanced FGD projects selected by the DOE. The technology is notable for employing a waste product from the cement manufacturing process as the reagent, for recovering and recycling a component of that waste product back to the manufacturing process, and for producing two other potentially salable byproducts.

The cement kiln at Dragon Products produces approximately 470,000 tons of cement a year, burning about 90,000 tons of 2.5-3.0 percent sulfur coal in the process. In addition to cement, the plant generates 10.4 tons per hour of dust from the kiln. This dust contains too much potassium and sodium to be reused directly in the kiln feed and, because of its high pH, must be disposed of in a secure landfill.

While investigating ways to make this dust suitable for recycling to the kiln by reducing its potassium content, researchers at the company made the serendipitous discovery that a solution containing the dust removed 90 percent of the [SO.sub.2] from the kiln's flue gas. In addition, at the end of the process the kiln-dust slurry contained calcium solids sufficiently free of potassium to be recycled back to the kiln. Finally, the process generated potassium sulfate, which is a major constituent of fertilizer, and distilled water.

Cement is produced by heating a mixture of minerals - primarily limestone, but also clay, sand, and iron ore - in a kiln. The kiln is a 200-ft, refractory-lined, cylindrical vessel that is elevated at one end to create a slight slope. The entire vessel slowly rotates to mix the contents while the feed is fed into the elevated end. A burner is located at the opposite end, and the combustion products pass through the kiln to heat the feed materials. The exhaust gases exit the kiln and, after passing though a dust collector, are discharged to the atmosphere through a stack.

Because the burner end of the kiln operates at high temperatures, some of the potassium and sodium salts contained in the feed minerals vaporize and condense as oxides in the form of a fine dust in the cooler part of the kiln. The dust leaving the kiln is therefore relatively high in potassium and sodium. Only limited quantities of these substances are permitted in cement production, and for this reason the kiln dust is not suitable for recycling.

In the recovery process, the dust removed from the flue gas is transferred to a mixing tank, where it is combined with recycled, process-derived water to form a slurry. The slurry is then pumped to the reaction tank. Meanwhile, the gas leaving the dust collector is cooled as it passes through the crystallizer/heat exchanger. Next, it is bubbled through the slurry. The reaction of the potassium oxide in the slurry and the [SO.sub.2] in the flue gas causes potassium sulfate to form. The desulfurized flue gas then leaves the reaction tank and enters the atmosphere.

Slurry is continuously removed from the bottom of the reaction tank and pumped to the settling tank, where suspended solids - mostly calcium carbonate, which makes up 95 percent of limestone - are separated from the potassium sulfate solution. The settled solids are pumped to a dilution tank and mixed with process-derived distilled water, which dissolves any residual potassium sulfate. The resulting dilute potassium sulfate solution is pumped back to the mixing tank, where it is used to prepare the kiln-dust slurry. The settled solids, by now rich in pure calcium carbonate, are ready for recycling to the kiln.

Meanwhile, the potassium sulfate solution, free of solids, is further processed to produce potassium sulfate crystals and distilled water. In the heat exchanger/crystallizer, it is heated indirectly by the flue gas. The hot solution is then fed to a flash tank in the evaporation section, where a portion of the water is vaporized and the potassium sulfate solution is cooled. The combination of cooling and evaporation causes crystals of potassium sulfate to form. These are removed from the remaining solution and dried. The water vapor leaving the flash tank is condensed and a portion of the distilled water is recycled to the dilution tank.

If buyers can be found, both the excess distilled water and the potassium sulfate crystals formed at this last stage in the process can be sold. While it is unrealistic to expect distilled water to generate much cash, even if the market for it were larger, potassium sulfate is a promising source of income. Dragon Products anticipates that all the potassium sulfate produced at its plant will be sold to a single fertilizer manufacturer.

Commercial Prospects

Although the flue-gas feed rate to the pilot plant where this technology was developed was only 2000 cu ft/min, all the equipment was of the same size and type used in a commercial operation. Tests at the pilot plant confirmed that the scrubber's [SO.sub.2] removal efficiency was at least 90 percent, that is, as good or better than most other scrubbers. In addition, there were indications that the process reduced emissions of nitrogen oxides and carbon monoxide as well as [SO.sub.2]. One of the purposes of the demonstration program at the cement plant will be to measure levels of [NO.sub.x] and [CO.sub.2] in the flue gas after scrubbing. Even if it turns out that the system does not appreciably reduce [CO.sub.2] emissions, the fact that it does not augment them will set it apart from other systems.

Several other potassium-based scrubbers are currently under consideration, but they are still in the early stages of development. A significant obstacle to commercializing this technology is that potassium is relatively expensive, and a regenerable system is therefore desirable to minimize costs. However, regeneration facilities are generally more complex and expensive than those that generate a throwaway product. By employing waste from its own manufacturing process for the absorbent, the technology developed by the Passamaquoddy tribe thus eliminates a major expense associated with these scrubbers.

In addition to generating its own scrubbing reagent, the system at Dragon Products creates no waste sludge for disposal. All the unused scrubber reactants, both solid and liquid, are recycled. Moreover, the amount of coal required to heat the kiln will remain unchanged by incorporation of the system, and the fresh limestone needed to feed the kiln will be reduced by the recycled calcium carbonate.

Research conducted at Dragon Products suggests that the technology can be applied outside of the cement industry by substituting biomass ash for kiln dust as the source of potassium. Although the reacted ash that would constitute the byproduct of this process could not be recycled, it is a more benign substance to dispose of than other scrubber waste products. The Passamaquoddy tribe, which owns the patent for the recovery scrubber, is pursuing the possibility of expanding its use to other cement plants and to the utility industry.

The technology of flue gas scrubbing has undergone continual improvement since it was discovered nearly a century ago, and opportunities exist to refine it further. Many advanced FGD systems have proved more reliable and less costly than their conventional counterparts, eliminating more [SO.sub.2] and consuming less of the plant's total energy output. In many cases, disposal problems have been overcome by eliminating the waste sludge created by conventional scrubbers, an improvement sometimes attended by the commercial benefit of a potentially salable byproduct like gypsum or potassium sulfate. The ideal scrubber, however, would be capable of reducing [NO.sub.x] and [CO.sub.2] emissions as well as [SO.sub.2]. Reduced water requirements and the use of a recoverable absorbent that can be reused or marketed for other purposes would also provide important advantages over today's systems.

In the United States, about 160 FGD units are currently installed in boilers with a total power capacity of about 70,000 MW. Most of these were built between 1975 and 1986, in response to the Clean Air Act. Since then, few new FGD systems have been installed, because of a slowdown in new power plant construction. Now, however, demand for electricity-generating capacity is expected to increase, and more stringent emissions reduction requirements are anticipated under the new clean air legislation being debated in Congress.

Together, these factors are likely to contribute to a growing market for FGD systems, some of which will undoubtedly include advanced systems such as those funded under the Clean Coal Technology Program. Apart from the use of low-sulfur coals, these may offer the most promising alternative, in terms of cost and efficiency, to the most common emission-control method in use today - the wet limestone scrubber.

PHOTO : Advanced scrubber. This Mitsubishi flue-gas desulfurization (FGD) system is one of 75 the company has built around the world. Mitsubishi Heavy Industries America Inc., in partnership with Air Products and Chemicals Inc., has formed a joint venture called Pure Air to build an FGD system at Northern Indiana Public Service Co.'s Bailly Station facility.

PHOTO : Wet process. A schematic of the $150 million gypsum-producing wet scrubber process being retrofitted onto the Bailly Station facility near Gary, Ind. Its objective is to reduce [SO.sub.2] emissions by 90 percent at approximately half the life-cycle cost of a conventional system.

PHOTO : Recycling waste. A schematic of the $10 million potassium-based scrubber, developed by Passamaquoddy Technology, that will be demonstrated at the Dragon Products Co. cement plantin Thomaston, Me. The system uses a waste product from the cement manufacturing process as the reagent to recover and recycle part of the waste back into the manufacturing process, and for producing two other potentially salable byproducts.