FINAL TECHNICAL REPORT

September 1, 1997, through August 31, 1998

Project Title: A STUDY OF CHLORINE IN HIGH TEMPERATURE CORROSION OF ALLOYS IN AN AFBC SYSTEM

ICCI Project Number: 97-1/2.2A-1

Principal Investigator: Wei-Ping Pan, Western Kentucky University

Other Investigators: John T. Riley, Western Kentucky University

Ian G. Wright, Oak Ridge National Laboratory

Project Managers: Ken K. Ho, ICCI and Arun K. Mehta, EPRI

ABSTRACT

The possibility of corrosion of power plant boiler components and ash deposits on heat transfer surfaces are always major concerns when high sulfur and high chlorine coals are used. For this reason the chlorine content in some Illinois coals is a concern for potential end users of these coals. The major advantage of fluidized bed combustion (FBC) of fossil fuels is its ability to absorb sulfur oxides by limestone in the combustor. Limestone can also capture hydrogen chloride to form liquid or solid phase calcium chloride under relatively low temperatures in the freeboard area of an AFBC system, which can suppress the corrosion of boiler components caused by chloride compounds. The major purposes of this research project is to help the coal industry solve the corrosion and ash deposit problems that may be associated with the utilization of high chlorine coals, and to provide technical support for the operation of the 160 MWe atmospheric fluidized bed combustion (AFBC) boiler at TVA's Shawnee Steam Plant near Paducah, KY.

During the past year, the 0.1 MW_th AFBC facility at WKU was modified extensively, including the combustor, electrical system, gas and solid sampling systems, and the cooling system. The coal (15 tons) and limestone needed for the first 1000-hour test burn, which will begin the first week of September, 1998, were acquired and prepared. Samples of cooled tubes and uncooled-boiler tubes (210-C) were designed, constructed, and suspended in the AFBC system, as well as in an electrical furnace, to test the corrosion rates caused by chloride in the AFBC system and metal wastage caused by oxidation. Forty-two test coupons (14 each of 304SS, 309SS and 347SS alloys) were acquired and analyzed (composition, thickness, etc.). for tests in the AFBC system. Nineteen ash deposits (from previous test runs) were collected from different regions of the AFBC system and were analyzed for the elements and specific compounds present. A comprehensive study was carried out in a TGA and a small scale AFBC to better understand the interaction between chlorine and sulfur during combustion processes in AFBC systems. The results showed desulfurization efficiency can be improved significantly with the existence of HCl in the flue gas. Also, the retention of chlorine by limestone is increased when SO_x is in the flue gas.

EXECUTIVE SUMMARY

At least half the known economically minable coal reserves in Illinois have a chlorine content around 0.3%, which has been a concern for potential end users of these coals. Between 1992 and 1993, TVA's Shawnee Plant observed that the boiler tubes in the primary superheater region of the atmospheric fluidized bed combustion (AFBC) system had wastage/corrosion problems. The system had been firing some Illinois coal and in 1993 TVA declined to renew an Illinois Basin coal (0.4% Cl) contract for approximately 4 million tons per year. The decision was made solely on the chlorine content of the coal. This concern is reflected in the suggested coal chlorine limits used by boiler manufacturers and in the application of coal chlorine content limitations in coal contracts.

It is known that AFBC systems can absorb sulfur oxides with limestone in the combustor with high sulfur retention efficiency. Limestone can also capture hydrogen chloride at relatively low temperatures in the freeboard region of the AFBC system. The scope of this project is to study the behavior of chlorine during coal combustion in an AFBC system. The results from this research project could help reduce the fuel costs for the 160 MWe AFBC unit at the Shawnee Steam Plant near Paducah, KY, if raising the specification limit for chlorine content is possible after completion of the study. A successful project demonstrating the utilization of high chlorine coals may open new markets for Illinois high chlorine coals. It will also reduce the risk of serious damage to commercial units using high chlorine coal, if power plant operators know the limits for utilization of such coals.

In order to honor the requests of TVA for corrosion tests under conditions simulating those of the 160 MWe AFBC system at the Paducah Shawnee Plant, the 0.1 MW_th AFBC facility at WKU was extensively modified by replacing the refractory layer using a much harder and stronger inner section, installing a new cooling system, upgrading the control system and installing two sets of secondary air nozzles. With these modifications, the oxygen concentration in the flue gas, the temperature, and the residence time of fine particles in the combustor can be adjusted by changing the primary/secondary air ratios.

Coal and limestone for the first 1,000-hour test burn was acquired, air dried, crushed to -4 mesh, and stored. The coal (15 tons) is a blend of two coals in order to achieve the chlorine (0.2%), sulfur (3%), and ash (10%) contents needed for the study. Analytical data from the two source coals and the blend showed that the alkali contents in all three were nearly the same, and the analytical values for the source coals are very close to those of the blend. The first 1,000-hour test burn will begin the first week of September, 1998.

In order to achieve both high capture efficiencies for SO_x and HCl in the three scheduled 1,000-hour test burns with the AFBC system, a study of limestone decomposition was carried out in a drop-tube furnace using a fast heating rate. It is known that the optimum temperature for capturing hydrogen chloride by limestone in an FBC combustor is less than 600^oC, whereas the optimum temperature for capturing SO_x is around 850^oC. Thus, the decomposition and use of limestone should be looked at in two stages: large particles which stay in the fluidized bed zone capture SO_x at relatively high temperatures, while middle-sized and fine particles that stay in the fluidized bed zone a very short time for decomposition escape from the bed and enter the freeboard area to react with HCl, thus reducing the HCl concentration in the flue gas. This consequently inhibits corrosion and lowers ash deposits. In this study the effect of particle size, residence time, temperature, and different types of limestone on the decomposition rate were studied. The results indicated two factors that may affect SO_x and HCl capture efficiencies. First, the decomposition temperature shifted to higher value with an increase in particle size of limestone. A higher conversion rate from limestone to calcium oxide is obtained with the smaller particles and longer residence time. Second, the decomposition of limestone is very sensitive to temperature. The higher the temperature, the faster it decomposes. The conversion rate for limestone at 830^oC is much faster (almost double) that at 800^oC within the first minute.

It was determined in a study of 19 ash deposits, resulting from previous test runs and collected from various sections of the WKU AFBC system, that the temperature in an AFBC system plays a key role in the retention of sulfur and chloride in the ash. While most of the SO_x was captured in the fluidized bed zone at high temperatures around 850^oC, the chloride content in the ash deposits increased along the entire height of the combustor, even with small alkali metal contents. This indicates that the reaction between HCl and CaO occurs under relatively low temperatures (<650^oC). This information also will be useful for choosing the kind of limestone and particle size to get high absorbance efficiency of SO_x and HCl in the AFBC system.

A study of the interaction between HCl and SO_x captured by a sorbent in an AFBC system was carried out in a TGA system and a lab-scale FBC, which has a 0.05 meter I.D and 1.0 meter height. The major tasks for this investigation were to (1) study the mechanism of the effect of HCl on SO_x emissions from an FBC system, and (2) find the optimum sorbent and conditions for the capture of HCl and SO_x. A goal of the study was to determine whether or not HCl promotes the conversion reaction between SO₂ and SO₃, and whether or not HCl improves the absorbing surface area of sorbent. In this study, limestone (or calcium oxide) was continuously fed into the top of the combustor using an auger feeder, while reaction gases, including HCl, SO_x and oxygen supplied by cylinders, were injected through a distributor at the bottom of the combustor. Nitrogen was chosen as an adjustable reagent. The concentration of each reaction gas was controlled by a calibrated flowmeter. The test results indicated that HCl can influence SO_x emission remarkably. The desulfurization efficiency increases sharply with an increase in HCl concentration in the reaction gases.

Forty-two test coupons (14 each of 304SS, 309SS and 347SS alloys) were acquired and prepared for the three 1,000-hour test burns. Each circular coupon has a 5.08 cm outside diameter, a 1.587 cm hole in the center, and is 0.3175 cm thick. A layer of corrosion-resistant metal (25% Cr, 75% Fe) was coated around one-side rim of each specimen.

Measurement to a high degree of precision of the final surface relative to the original surface protected by the resistant layer can now be made using an optical microscope. The procedures for assessing changes in the coupons requires that specimens be removed after each 250 hours of a test burn. Complete testing of the metal coupons requires destructive examination. However, only eight coupons can be installed in each run. In order to obtain as much corrosion data as possible for every 250 hours of exposure, a special procedure was adapted. The coupons will be removed in a fixed sequence after each 250 hours, so that coupons of each alloy will be exposed for 250, 500, 750, and 1,000 hours during each test burn. In order to get reference data for future tests, two coupons of each alloy with three random points for each were examined by SEM-EDS. The SEM analysis was performed with a JEOL JSM-5400 SEM with a KEVEX Sigma 1 EDX System with a Quantum detector for elemental analysis down to carbon. A small piece, the analysis sample, was cut from each coupon, mounted, ground, polished and coated by carbon. The analytical results indicate that the profile of each element is very uniform. Also, four points were chosen and marked to provide the same point to measure after every 250-hour test period. The initial thicknesses of all alloy specimens were carefully measured.

A 1000-hour oxidation test with a 304SS metal coupon was carried out in an air atmosphere in an electrically heated furnace. The results will be used as basic reference data to determine what type of metal alloys are helpful in suppressing corrosion. Similar tests with flue gas conditions similar to those in three 1,000-hour runs will be conducted.

For the second year of the project, completion of the first 1,000-hour burn with the 0.20% chlorine coal will be followed by two 1,000 hour burns using coals with higher chlorine contents. Coals with 0.35% and 0.45% chlorine contents containing about 3% sulfur and 10% ash will be used in these tests. Negotiations for the acquisition of these two test coals are underway. Extensive analysis of the test coupons, the exposed boiler tubes, and deposits from the AFBC system will follow the test burns. Comparison of the test results from each test burn, along with results from previous test burns, should provide adequate information about the role coal chlorine may play in the corrosion of boiler components.