Mercury control on trial21 September 2001
Gaston unit 3, with a unique combination of baghouse downstream of an ESP is one of four test sites being used in a US DoE programme on mercury control.
A great deal of research (much of it funded by the US DoE, EPA, and EPRI) has been conducted during the past decade to characterise the emission and control of mercury compounds from the combustion of coal. This research anticipated the long-awaited December 2000 announcement by the US Environmental Protection Agency that it was going to regulate mercury emissions from US coal-fired power plants. Findings so far suggest that control of mercury from utility boilers will be both difficult and expensive.
Because of the large volumes of gas to be treated, low concentrations of mercury, and presence of difficult to capture species such as elemental mercury, some estimates show that 90 per cent mercury reduction could cost the industry as much as $5 billion per year. The main burden of these costs would fall on plants not equipped with wet scrubbers and which burn low-sulphur coal.
Injection of dry sorbents such as powdered activated carbon (PAC) into the flue gas and further collection of the sorbent by ESPs and fabric filters represents the most mature and potentially most cost-effective control technology for power plants.
However, all of the work to date has been conducted using bench-scale and pilot experiments. Although these reduced-scale programmes provide valuable insight into many important issues, they cannot fully account for impacts of additional control technology on plant-wide equipment. It is therefore necessary to scale-up the technology and perform full-scale field tests to document actual performance levels and determine accurate cost information.
Under a DoE/NETL co-operative agreement, ADA-ES is working in partnership with PG&E National Energy Group, Wisconsin Electric, Alabama Power Company (a subsidiary of Southern Company), EPRI, and Ontario Power Generation on a field evaluation programme of sorbent injection upstream of existing particulate control devices for mercury control.
Other organisations participating in the programme include Apogee Scientific, URS Radian, Energy & Environmental Strategies, Physical Sciences, Inc, Southern Research Institute, Hamon Research-Cottrell, Environmental Elements Corporation, Norit Americas, and EnviroCare International.
The NETL programme
The Department of Energy's National Energy Technology Laboratory (NETL) is the primary funding agency for an industry cost-shared test programme to obtain the necessary information to assess the costs of controlling mercury from coal-fired utility plants that do not have scrubbers for SO2 control. The method for mercury control evaluated in this programme is the injection of dry sorbents, such as activated carbon, upstream of the existing particulate control device. The economics will be developed based on various levels of mercury control at four different host sites. The four sites, listed in the table "Coal fired plants", burn coal and have particulate control equipment that is representative of about 75 per cent of the coal-fired generation capacity in the USA.
Gaston unit 3 was chosen as one of the test sites because COHPAC represents a cost-effective retrofit option for utilities with electrostatic precipitators (ESPs). COHPAC is an EPRI patented concept that places a high air-to-cloth ratio baghouse downstream of an existing ESP to improve overall particulate collection efficiency. The advantages of this configuration are:
• Sorbents are mixed with a small fraction of the ash (nominally 1 per cent), which reduces the impact on ash reuse and waste disposal.
• Pilot plant studies and theory indicate that compared with ESPs, baghouses require one-tenth the sorbent to achieve similar removal efficiencies.
• Capital costs for COHPAC are less than other options such as replacing the ESP with a baghouse or larger ESP.
The Gaston plant
In the test at Gaston unit 3, carbon-based dry sorbents were injected upstream of COHPAC, downstream of the ESP over an eight week period.
The Gaston plant is owned and operated by Alabama Power Company. It is located in Wilsonville, Alabama. The plant has four 270 MW balanced draft and one 880 MW forced draft coal fired boilers. All units fire a variety of low-sulphur, washed, Eastern bituminous coals.
The primary particulate control equipment on all units is hot-side ESP. Units 1 and 2 and units 3 and 4 share common stacks.
In 1996 Alabama Power contracted with Hamon Research-Cottrell to install COHPAC downstream of the hot-side ESP on unit 3. This COHPAC system was designed to maintain unit 3 and 4's stack opacity levels below 5 per cent on a 6 minute average.
The COHPAC system is a pulse-jet cleaned baghouse designed to treat flue gas volumes of 1 070 000 acfm at 290°F (gross air-to-cloth ratio of 8.5 ft/min with on-line cleaning). The COHPAC baghouse consists of four isolatable compartments, two compartments per air-preheater identified as either A- or B-side. Each compartment consists of two bag bundles, each having a total of 544, 23-foot long, Ryton felt filter bags, 18 oz/yd nominal weight. This results in a total of 1088 bags per compartment, or 2176 bags per casing.
The hot-side ESP is a Research-Cottrell weighted wire design. The specific collection area (SCA) is 274 ft2/1000 acfm. Depending on the operating condition of the hot-side ESP, nominally 97 to 99+% of the flyash is collected in the ESP. The remaining flyash is collected in the COHPAC system. The average inlet particulate mass concentration into COHPAC between 1/97 and 4/99 was 0.0413 g/acf.
Hopper ash is sent to a wet ash pond for disposal. A hydrovactor system delivers the flyash to the pond.
Design parameters for Gaston unit 3 are presented in the table to the right.
The overall objective of testing at Alabama Power's Gaston unit 3 was to determine the cost and impacts of sorbent injection into the COHPAC baghouse for mercury control.
The evaluation was conducted on one-half of the gas stream, nominally 135 MW. The side chosen for testing was B-side. A-side was monitored as the control unit.
To achieve the overall objective, the programme was designed with an extensive field evaluation and an equally extensive laboratory testing and analysis effort.
The critical elements of the programme were the actual field tests and measurements, which relied upon accurate, rapid measurements of mercury concentration and an injection system that realistically represented commercially available technology.
Near real-time vapour phase mercury measurements were made using a Semi-Continuous Emissions Monitor (S-CEM) designed and operated by Apogee Scientific. This instrument was developed with EPRI funding to facilitate EPRI research and development efforts. The S-CEMs operated continuously for over seven weeks providing speciated, vapour phase mercury concentrations at the inlet and outlet of COHPAC.
Norit Americas supplied a portable dilute phase pneumatic injection system that is typical of those used at Municipal Solid Waste (MSW) facilities for mercury control with activated carbon. ADA-ES designed the distribution and injection components of the system.
The test plan was designed to allow for the evaluation of two sorbents during the parametric tests. The test protocol required that one of the sorbents be a lignite-derived powdered activated carbon (PAC). The second sorbent was selected based on results from fixed-bed screening tests.
The field tests were separated into four different test phases:
• pre-baseline and sorbent screening;
• parametric; and
• long term.
Pre-baseline measurements and screening
The first field measurements were made prior to installing the injection equipment. The objectives for the pre-baseline tests were to:
• Measure vapour phase mercury concentrations at several locations using the S-CEM to compare results with Ontario Hydro measurements (draft wet method used by EPA) done in 1999 (these measurements were made across the hot-side ESP on unit1);
• Document mercury emissions across COHPAC; and
• Perform screening tests for mercury adsorption characteristics of several activated carbons that were candidate sorbents for the full-scale tests.
Five carbon-based sorbents, three variations of ash from Gaston, and one non-carbon based sorbent were screened by URS Radian in a laboratory mercury adsorption test fixture.
Six of these sorbents were then evaluated in a similar test device on a slip stream of flue gas at Gaston.
After equipment installation and checkout, a set of baseline tests were conducted immediately prior to the first parametric test series to document current operating conditions. COHPAC and plant operating data were collected.
During the baseline test, boiler load was held steady at "full-load" conditions during testing hours, nominally 7 am to 7 pm. Mercury across the B-side of COHPAC was measured using two separate methods: S-CEMs; and Modified Ontario Hydro Method.
In addition to monitoring mercury removal, it was also important to document the performance of COHPAC during sorbent injection, which is critical to the success of sorbent injection for mercury control. The primary performance indicators are:
• Pressure drop/drag. Pressure drop and drag are both used to monitor the permeability of the filter and dustcake. Pressure drop is a direct measurement of pressure loss across the fabric filters. Drag is a calculated number that normalises pressure drop to flow by dividing pressure drop by the air-to-cloth ratio. These values are a function of inlet grain loading, filtering characteristics of the particulate matter, and flow and time between cleaning. Of particular interest is the change in rate of pressure drop increase with sorbent injection and whether pressure drop/drag returns to baseline levels when injection is stopped.
• Cleaning frequency. Pressure drop/drag is controlled in a baghouse by the cleaning frequency. It is expected that cleaning frequency will increase with the increased particulate loading from sorbent injection. Cleaning frequency was monitored before, during and after sorbent injection.
• Opacity/emissions. Cleaning frequency and particulate matter characteristics can affect collection efficiency across the baghouse. Most emissions occur immediately following a clean, so increasing the cleaning frequency can increase outlet emissions. The emissions could also increase if the particulate does not form a high efficiency filter cake, but tends to work through the fabrics.
• Bag strength. The filter bags in COHPAC are made from Ryton felt. The Ryton bags at Gaston have experienced very little loss in fabric strength, as measured by Mullen Burst tests, in the four years of operation. To ensure that carbon injection will not adversely affect fabric strength, samples of both old and new bags were pulled periodically throughout the test. Prior to the baseline tests, several new bags were installed in both the A- and B-side to monitor short term strength loss.
A series of parametric tests was conducted to determine the optimum operating conditions for several levels of mercury control up to 90 per cent mercury removal, for several activated carbon products. To minimise permitting issues, only coal-based sorbents were considered at this site. Norit Americas lignite-based PAC, Darco FGD, was chosen as the benchmark sorbent. Darco FGD is Norit's standard product for mercury removal at MSW and incineration sites. Sorbent and injection concentration for the long term tests were chosen based on results from these tests.
Long-term performance tests
Long-term testing at "optimum" plant operating conditions (lowest cost/highest mercury removal), as determined from the parametric tests, aimed to gather data on:
• mercury removal efficiency over time;
• the effects on COHPAC and balance of plant equipment of sorbent injection; and
• operation of the injection equipment to determine the viability and economics of the process.
During these tests, carbon was injected continuously 24 hours per day, for 10 days. Darco FGD activated carbon was chosen as the sorbent for these tests. The injection rate was determined taking into consideration both mercury removal and the projected increase in COHPAC cleaning frequency.
As in the baseline test series, mercury was measured by both the S-CEMs and manual methods (Ontario Hydro). COHPAC performance, coal and ash samples, and plant CEM data were collected.
Applying the results
Based on results from the parametric tests, Darco FGD powdered activated carbon was chosen as the sorbent.
A significant increase in the cleaning frequency of the COHPAC baghouse occurred with the injection of activated carbons. At Gaston 3, the maximum acceptable cleaning frequency and pressure drop limited the amount of sorbent that could be injected.
Results from the S-CEMs showed 80 - 90 per cent mercury removal at an injection rate of 1.5 lb/Mmacf. This injection rate was selected to minimise the COHPAC cleaning frequency, which varied between 1.0 and 1.3 pulses/bag/hour.
Using data obtained from these tests, future COHPAC (TOXECON) baghouses can be designed to operate acceptably with carbon injection. The next step is to apply the findings to the design and costing of a full-scale commercial system for mercury control using sorbent injection.
TablesGaston unit 3 - the basic data Coal fired plants