San Juan Shaves Costs From FGD Retrofit
By Henry S. Taylor, Public Service Company of New Mexico, and Walter Nischt, Babcock and Wilcox Co.
San Juan Generating Station (SJGS), Operated by Public Service Company of New Mexico (PNM) for a
group of nine owners, is an 1,800 MW coal-fired plant in the northwest corner of New Mexico. The station has two 350 MW Foster-Wheeler boilers and two 550 MW Babcock & Wilcox boilers with three General Electric turbine/generators and one Westinghouse turbine generator. The plant is close to large deposits of low sulfur (0.9 percent), high ash (over 24 percent) subbituminous coal. Two open pit mines operated by Broken Hills Properties Company Ltd. supply coal by truck to the station.
Sulfur dioxide emissions are currently controlled with a Wellman-Lord (W-L) regenerative system, and particulates are removed by hot side electrostatic precipitators. Cooling water and service water are supplied to the plant from the San Juan River. SJGS has been a “zero water discharge” location since 1983, meaning that no water discharge is allowed from the plant site. All wastewater must be recycled and the final brine must be stored in evaporation ponds.
In 1994, SJGS owners became concerned over the comparatively high operating and maintenance (O&M) costs for the flue gas desulfurization (FGD) system. The W-L system was originally chosen in the late sixties because it offered competitive O&M costs and because of concern about sulfur waste being a hazardous material. Furthermore, at the time, competing lime and limestone FGD systems had been encountering technical difficulties. In the past 20 years, however, the W-L system at SJGS has aged and now incurs comparatively high maintenance costs, while limestone systems have become reliable and efficient. In early 1995, SJGS compared the costs for the W-L system with those of the most common modern systems. Results from a study using the Electric Power Research Institute`s FGD Cost Model confirmed the comparative high cost of the W-L system and identified a limestone forced oxidation (LSFO) system as the best economic choice for SJGS.
EPRI`s FGD Cost Model estimated the cost of a LSFO retrofit for SJGS at $110 million. However, collaborative efforts between PNM and Babcock & Wilcox (B&W)–aimed at design refinement, utilization of existing equipment, utilization of plant personnel and facilities, value engineering and risk sharing–reduced the budgeted cost to $80 million. The LSFO FGD system presented an opportunity for drastic reductions in O&M and energy costs. Economic justification based only on the FGD retrofit, however, was not satisfactory. A conversion to LSFO could greatly simplify the wastewater treatment system by using cooling tower and boiler blowdown directly in the limestone slurry, eliminating the need for more than one-half of the wastewater treatment equipment (including several evaporators, a reverse osmosis system and an oxidation system). The additional savings afforded through the wastewater treatment simplification provided final justification for the project.
The SJGS project is unique in three ways. First, it is justified on the basis of economics, not regulatory requirements. Nonetheless, the new system will be more efficient and remove more SO2 than the old system. Second, it is a true retrofit project. Much of the existing equipment will be reused and the station will remain in compliance with the air emission regulations during conversion. Third, the unique teaming arrangement between PNM and B&W simplifies project organization, reduces cost and involves the operator more deeply in the project.
Old and New
The existing W-L system is based on the chemical reaction of sodium sulfite with SO2 to form sodium bisulfite. The W-L absorbers (four per unit) are countercurrent multi trays with recirculating solution on each tray. Prescrubbers remove residual flyash and chlorides and cool the flue gas. The low pH scrubbing liquid is in the absorber sump and is kept separate from the absorbing solution, which is drawn off from the lower tray. The W-L system is equipped with demisters between the sump and the lower tray and at the top of the absorber. The product solution is sent through two-stage evaporators to produce the SO2 and regenerate the solution. To remove any oxidized sulfate solution, which cannot be regenerated, crystallizers and evaporators are used, resulting in a sodium sulfate product for sale. None of the byproducts are currently economic. The W-L system incurs very high energy costs because of the high absorber pressure drops and the use of boiler steam for regeneration. In addition, more than 200 pumps are used in the W-L process.
In contrast, the LSFO system (see figure) is simple and uses much less energy. Limestone slurry is prepared in Svedala ball mills and pumped to the absorbers, where it is sprayed at two levels. A B&W patented tray, designed to improve gas liquid contact, replaces the existing tower internals. Two sets of demisters are located at the top of each absorber. Spent slurry in the sumps or reaction tanks is oxidized to gypsum and is drawn off of the absorber sumps and dewatered. Primary dewatering is accomplished with hydroclones and secondary dewatering is done with vacuum drum filters. Dewatered gypsum (85 percent solids) is conveyed to a pile where it is hauled back to the coal mine with flyash and bottom ash for disposal. Reduced absorber pressure results in lower booster blower power, and the simpler process drastically reduces the number of pumps required. No steam is required for the LSFO system. The LSFO system also provides a less corrosive environment than the W-L system, resulting in less expensive materials and fewer maintenance problems.
In order to retrofit a LSFO system into the existing W-L system, all equipment had to be evaluated for possible reuse. The new design is based on the reuse of the existing absorber towers and all of the inlet and outlet ductwork. Since less power is required in the LSFO system, new smaller motors were added to six of the absorbers and smaller fan rotors were added to the other six absorbers. Two new recirculation pumps were added to each absorber, abandoning the old scrubber pumps and tray pumps. Some existing pipe racks were used to tie the limestone preparation/dewatering area to the absorbers. Several tanks and a silo were converted for limestone storage. All of the electric power supply was reused and the switchgear and breakers were converted to limestone system use. Controls were simplified and updated to a distributed control system (DCS). Three control rooms were consolidated into one and two were abandoned.
Phased Construction and Operation
PNM and B&W mutually developed a detailed budget for the project and then value engineered improvements to lower the cost. All design issues–including the reuse of existing equipment and new equipment selections–were mutually agreed upon.
Three of the four absorber vessels per unit will be converted from W-L to LSFO. Each LSFO absorber will have the capacity to handle 33 percent of the design gas flow and each absorber is designed for 90 percent SO2 removal. SJGS must meet New Mexico air emission regulations, which require approximately 75 percent SO2 removal or 0.46 lb SO2/MMBtu on an annual rolling average. To keep the plant in compliance during construction, only one absorber per unit could be converted at a time. The project, therefore, is being constructed in three phases.
PNM released B&W to begin engineering in May 1996. Babcock & Wilcox Construction Co. mobilized on the site in May 1997 and completed Phase I in March 1998. Phase I consisted of housing the limestone preparation and gypsum dewatering areas in a new building. It also included the conversion of the first absorber on each unit from the old W-L type to the new LSFO design. The first four LSFO absorbers were started up as a system in March 1998 and are currently being operated in conjunction with two W-L absorbers per unit. Phase II of construction, to be completed in July 1998, is conversion of the second absorber on each unit to LSFO absorbers. After Phase II commissioning, two LSFO absorbers will be operating on each unit and the W-L system will be shut down. Existing compliance will be attainable with two limestone scrubbers per unit because it will be possible to pass more flue gas through the limestone scrubbers than was possible through the W-L absorbers. The final construction phase will be conversion of a third absorber on each unit to limestone. Final project completion is scheduled for January 1999.
SJGS justified the FGD retrofit project on the basis of O&M and power/fuel savings. A large percentage of the savings came from the reduction of personnel by 110 positions. The old W-L and waste water systems required 67 people in operations, 70 in maintenance and 16 providing miscellaneous support; the new limestone/waste water system will require 26 operations personnel and 17 multi-craft maintenance personnel. Most of the power savings for the new LSFO system come from reduced pressure drop. The old W-L system had a prescrubber and five sets of internal trays. The LSFO system requires only about 60 percent of the blower power. Additional power savings will be realized through equipment elimination in the W-L regeneration process and wastewater treatment system. Overall, the auxiliary power requirements are expected to be reduced by almost 30 MW.
The LSFO system after Phase I start-up is operating with four absorbers or one for each unit. Removal efficiencies have been measured and absorbers are performing according to design. One positive indication is that the limestone absorbers are processing more flue gas than the existing W-L absorbers. The limestone preparation system and gypsum dewatering equipment are currently operating at one-third capacity.
Phase I was completed under budget and savings of $6.6 million have been declared. B&W and the subcontractors shared in part of the savings, which shows up as a project cost. This conversion project, with the realized savings to date, is now seen as a $77 million project that will provide O&M fuel cost savings of $16 to $23 million a year (depending on the way fuel and power are valued). Additional savings are expected in phases II and III through labor efficiency. Ongoing savings are expected when the old W-L FGD system and one-half of the waste water treatment system are shut down and cleaned up for disposal.
Most of the start-up problems concerned instrumentation and controls. For example, the absorber sump level indicator plugged frequently because of the thick ceramic absorber walls, necessitating redesign of the pressure taps with an air purge. Ironically, the advanced DCS, designed to be as automatic as possible, has also caused problems. Eliminating operator discretion has occasionally led to system upset conditions. p
The 1,800 MW coal-fired San Juan Generating Station is converting its 20-year-old Wellman-Lord scrubber system into a limestone forced oxidation scrubber system.
The retrofit project relies on as much ofthe original equipment as possible. This photo shows the limestone preparationand dewatering area, where tanks andpipe racks were reused.
Henry S. Taylor is project engineer with Public Service Company of New Mexico.
Walter Nischt is a project manager in Babcock and Wilcox Co.`s Utility, Environmental and Industrial Equipment Division.