Modifications to boiler combustion systems allow Fayette Power Project Units 1 and 2 to meet new NOx emissions limits in Eastern Texas
In May 2005, the Texas Commission on Environmental Quality began requiring all coal-fired power generation units in the eastern half of the state to limit yearly average nitrogen oxide (NOx) emissions to no more than 0.165 lb/MMBtu. The Fayette Power Project, a three-unit coal-fired electricity generation facility located east of LaGrange, Texas, and operated by Lower Colorado River Authority (LCRA), fell under this requirement. Based on industry experience with similarly sized tangentially-fired power boilers fueled by Powder River Basin (PRB) coal, LCRA staff recommended that the Fayette Power Project (FPP) modify its boiler combustion systems to meet this emission requirement.
FPP selected Alstom to conduct these modifications. Unit 1 was modified during a planned outage in October and November 2002 and Unit 2 modifications were executed during an outage in March and April 2004. An overall reduction in NOx emissions of almost 69 percent was attained for each unit. FPP’s future efforts will be to optimize NOx emissions at reduced loads as well as continue fine tuning emissions at full load.
Fayette Power Project Unit 1 and 2 Boilers
FPP Units 1 and 2 are nominal 600 MW units co-owned by LCRA and Austin Energy. Both units are identical, with Combustion Engineering (CE) subcritical boilers, each rated for 4,199,000 lbs/hr steam flow at a superheat outlet steam pressure of 2,620 psig and 1005 F temperature. Each is a tangentially-fired single-furnace, balanced draft unit.
Employees install overfire air ductwork. Photo courtesy of Alstom.
The original windbox configuration for each unit consisted of seven coal elevations and two close-coupled overfire air (CCOFA) elevations. The top six coal elevations were used to fire PRB coal, and only five elevations needed to be in service to achieve design steam flow. An unused bottom elevation was included in the design as a future coal elevation in the event that the unit switched to lignite. On Unit 1, the original auxiliary air compartments were retrofitted with fixed concentric firing system (CFS) air nozzle tips in the fall of 1986.
Combustion System Modifications
Alstom supplied its low NOx concentric firing system, LNCFS Level 3, for both Units 1 and 2. For units that fire PRB coal, three major components are integrated into the LNCFS Level 3 system to lower NOx emissions with minimal effects on boiler performance: precise furnace stoichiometry control, initial combustion process control and concentric firing via the CFS air nozzle tips.
Precise furnace stoichiometry control is achieved by using multiple levels of overfire air (OFA). The LNCFS Level 3 system utilizes a combination of close-coupled overfire air (CCOFA) and separated overfire air (SOFA) to achieve the stoichiometry control necessary to reduce NOx. To minimize increases in carbon monoxide (CO), the OFA compartments have manual yaw adjustment, allowing each OFA compartment to be directed to maximize mixing during the burnout process.
Initial combustion process control is established by using Aerotip coal nozzle tips. A special design feature of these coal nozzle tips allows enhanced flame front position control under tilted conditions. This allows the flame front to be established near the exit of the coal nozzle tip, allowing for early devolatilization of the coal to further reduce NOx. Establishing early coal ignition improves flame stability and minimizes increases in unburned carbon.
Horizontally adjustable CFS air nozzle tips direct a portion of the secondary air in the main windbox away from the fuel stream in a circle concentric with the main firing circle. The CFS offset air provides an oxidizing boundary layer along the furnace waterwall and also reduces lower furnace waterwall slagging. Adjustable yaw CFS tips replaced the existing auxiliary air nozzle tips in both units.
FPP Unit No. 1
Tuning and Testing
Post-outage activities by Alstom and FPP personnel focused upon conducting a series of parametric tuning tests. The objective of the Unit 1 tuning was to characterize and optimize NOx and CO emissions, as well as the unburned carbon (UBC) in flyash, to achieve Alstom’s contractual guarantee obligations.
Over the course of the tuning, the following operating parameters were varied and evaluated for their effects on NOx, CO and UBC in fly ash:
• SOFA and CCOFA damper positions
• SOFA, CCOFA and CFS nozzle tip yaw positions
• SOFA and main burner tilt positions
• Fuel air damper position
• Excess air
• Windbox-to-furnace differential pressure
• Mills in service
Based upon the results of the tests conducted on Unit 1, new control curves were developed for the SOFA dampers; and modifications were made to the existing CCOFA and fuel air dampers, excess air, and windbox-to-furnace differential pressure control curves. Final yaw settings for the SOFA, CCOFA and CFS secondary air nozzle tips were also established.
Full load NOx and CO emissions ranged from 0.117 to 0.275 lb/MMBtu and 3 ppm to 539 ppm (corrected to 3.0 percent oxygen (O2)), respectively, depending on the operating conditions (tilt position, windbox-to-furnace differential pressure, excess air, yaw settings, SOFA and CCOFA dampers, etc.) being evaluated for each test. Figure 1 illustrates the NOx and CO results for each of the full-load tuning tests. Figure 1 also lists, adjacent to the tuning test number, the combination of mills in service during each tuning test. Fly ash samples were collected for select tuning tests and yielded UBC levels of 0.13 percent or less.
After operating for several weeks, a third-party testing company conducted two guarantee tests. One test was conducted with the bottom five mills in service and the second test with the top five mills in service. The results indicated that Alstom’s LNCFS Level 3 system met all contractual obligations. Post-modification guarantee test results are listed in Table 1, as are the pre-modification operating and emissions data for comparison.
Optimizing Furnace Air Distribution
After initial tuning and testing, FPP personnel elected to continue tuning in an attempt to further reduce NOx emissions. First, they continued with combustion air distribution adjustments in the windbox and SOFA zones.
Initially, adjustments were made to mill combination 1-2-4-5-6. For this mill combination, average NOx emissions were about 0.135 lb/MMBtu. By continuing to adjust furnace stoichiometry, NOx emissions were further reduced to 0.101 lb/MMBtu, while CO emissions averaged below 100 ppm. After implementing changes for this set of mills, FPP personnel studied the adjustments to determine if they could be implemented for the other mill combinations.
Because there are six burner elevations and only five are required for design steam load, six burner (and mill) combinations are available to achieve full load operation. As part of their tuning process, Alstom technicians developed individual CCOFA and SOFA damper curves for each mill combination. Each combination, however, produced different NOx emission levels. (It should be noted that for each combination NOx levels remained comfortably below the 0.165 lb/MMBtu limit.) With a change in mill combination, CCOFA and SOFA dampers will adjust to a different position relative to load.
FPP technicians considered whether developing one set of curves could level out these fluctuations in NOx emissions. After the optimization of the first mill combination, work was started in developing standardized curves. Upon completing many trials, standardized curves were developed for all mill configurations that would produce low NOx emissions at fairly uniform levels when optimum conditions occurred.
Optimizing Excess Oxygen
For the FPP Unit 1 boiler, the boiler fuel and air requirements are set by functions that are programmed into the boiler controls. Before the tuning effort, the amount of combustion air was “trimmed” by the O2 controller using a load-excess O2 function.
A total of eight excess O2 analyzers, located in the boiler ductwork between the economizer section and the air preheater, are used to measure excess O2. These analyzers are located after a bifurcation in the flue gas ductwork; therefore, each of the two ducts has a group of four analyzers. The measurements from the group of four analyzers per duct are averaged; and, of the two averages, the lower value is used for control.
It is well known that operating boilers with lowered excess O2 levels, while improving heat rate, also results in lower NOx emission levels. However, operating at too low an excess O2 level is detrimental because “wet” slag builds up on furnace walls. This wet slag is difficult to clean using conventional wallblowers. CO emissions also increase when excess O2 levels are reduced. Therefore, for good control of NOx, CO and slagging, optimal excess O2 control is essential.
Historically, two issues made operating at optimal excess O2 levels difficult for this unit: (i) an unequal distribution of excess O2 at the locations where the O2 analyzers were located, and (ii) changes in the distribution of excess O2 when mill configuration and burner tilt angles changed.
Before the combustion system was modified, significant differences in excess O2 levels between analyzers and average excess O2 values between ducts occurred frequently. This phenomenon was due to stratification of excess O2 in the flue gas, which prevented true measurements of excess O2 levels. Personnel sometimes observed average excess O2 differences as high as 2 percent to 2.5 percent between ducts. Certain mill combinations and burner tilt positions exacerbated the differences. These erroneous measurements, along with the original control scheme, caused high CO excursions with certain mill combinations. Other mill combinations resulted in excess O2 levels being biased down to reduce NOx emissions without incurring CO spikes. With the goal of operating the unit with even lower NOx levels, this problem with excess O2 imbalance had to be resolved. Personnel adopted two operational changes to mitigate these issues.
First, FPP personnel modified the excess O2 control scheme. Instead of using excess O2 levels as the primary “trim” control, they developed a three-tier control scheme. Like the original control scheme, they used a load-excess O2 function for “trimming” airflow. However, they also added a boiler stoichiometry limit and CO limit. The boiler stoichiometry limit, based on “trial and error” empirical data, was set using a ratio of measured combustion airflow and fuel flow that produced good combustion characteristics. A multiplier was used to set the ratio, so that the optimum ratio equaled one. The controls allowed this ratio to operate above one, but operating below this ratio caused an increase in airflow. A CO limit was also set up, so that CO was maintained below 100 ppm. An increase in CO above this value causes an increase in airflow. Depending on mill configuration, excess oxygen at full load can now vary from 2.5 percent to 3.2 percent. This change in the primary control parameter significantly reduced the fluctuation in NOx emissions.
Secondly, FPP personnel implemented a novel change in burner tilt control. The team determined that tilt position relative to each windbox had a first order effect upon O2 imbalance. Before modifications, the burner tilts operated at the same relative angle; now the burner tilts at each corner of the boiler operate at different tilt angles above and below the horizontal position. Since implementing this change, excess O2 “splits” have remained at less than 1 percent in most cases.
Improved Maintenance of the Sootblowing System
It is well known that NOx production increases with higher furnace gas temperatures. Slag coatings on boiler walls reduce heat transfer to the boiler water, producing higher furnace gas temperatures. Maintaining clean furnace walls results in lower NOx emissions.
Prior to the outage, wallblower maintenance was not a high priority. Several out of service wallblowers reduced the furnace wall cleaning system’s effectiveness. With the emphasis now on maintaining low NOx emissions, wallblower maintenance has become a higher priority. Flow transmitters were added to the wallblower steam lines to help diagnose when blowers are in need of repair. This renewed emphasis on maintaining all wallblowers, as well as maintaining optimized boiler airflow, has kept NOx emissions from climbing due to higher furnace gas temperatures.
FPP Unit No. 2
Tuning Results – Unit 2
Alstom and FPP personnel relied on their Unit 1 experiences when they began tuning Unit 2. Operating curves developed during the initial Unit 1 tuning phase were used as a starting point for the Unit 2 effort, but instead of immediately implementing the final Unit 1 control changes, an effort was made to determine the influence of adjusting SOFA yaws only. Ultimately, the team determined that the Unit 1 control settings were most applicable for Unit 2.
During Unit 2 tuning, full load NOx ranged from 0.103 to 0.147 lbs/MMBtu. Concurrent CO emissions ranged from 13 ppm to 185 ppm (corrected to 3 percent O2). Similar to the procedure followed for Unit 1, a series of tests that evaluated the impact of varying the operating conditions were conducted. Figure 2 illustrates the NOx and CO for each of the tuning tests. When comparing these tuning results with those shown in Figure 2 for Unit 1, the benefit of starting with the previously developed Unit 1 operating curves is clear.
The same third-party testing company once again conducted two guarantee tests for Unit 2. The two tests were also conducted with the top and bottom five mills in service, respectively. The gathered data indicated that Alstom’s LNCFS Level 3 system met all contractual obligations for Unit 2. Guarantee test results are presented in Table 1. Because both units are identically configured, Unit 1 pre-modification data was used as the baseline for Unit 2.
Long-Term Operating Results
As Figure 3 illustrates, average monthly NOx emissions from Unit 1 have generally followed a decreasing trend line from month-to-month in 2003, leveling out in 2004. From March 2003 to March 2004, NOx emissions averaged 0.12 lb/MMBtu. From January 2004 through August 2004, NOx emissions averaged less than 0.10 lb/MMBtu. May 2004 was the lowest monthly average for NOx emission from Unit 1, when NOx emissions averaged 0.096 lb/MMBtu. NOx emissions increased in September 2004, which has been attributed to Unit 1 having been dispatched at lower loads. Tuning efforts up to that time focused upon full load operation only.
FPP personnel are continuing to fine-tune Unit 2 across the load range. As Figure 4 shows, from July 2004 through December 2004, NOx emissions averaged 0.102 lb/MMBtu. NOx emissions in December 2004 increased from previous months, due to an inoperable fuel air damper. This clearly illustrates the need for maintaining proper air distribution and periodic inspection of key components. Similar to Unit 1, the lowest monthly average NOx emission from FPP Unit 2 averaged 0.096 lb/MMBtu in October 2004. p
James D. Mathis, P.E., is a senior engineer with LCRA’s Wholesale Power Services engineering department. Yan Lachowicz is manager of strategic products/services development with Alstom’s Performance Projects business in Windsor, Conn.
Larry Ohnheiser, Sr. and Scott A. Matus, both with LCRA, also contributed to this article.