Taking Care of the Long Term via Spinning Reserve

Issue 4 and Volume 115.

By Joe Zwers, freelance writer

SaskPower, the electric utility for the Canadian Province of Saskatchewan, is engaged in a plan to add more than 175 MW of wind generation to its portfolio without compromising growth or reliability.

Part of that strategy calls for building gas-fired peaking plants near remote industrial loads that can act as synchronous condensers to support the grid when they are not generating power. In December 2010 the first such plant, the Yellowhead Power Station, opened in North Battleford.

“Wind power is an intermittent source of electricity and therefore must be supported by alternate fuel generation,” said SaskPower President Robert Watson. “The Yellowhead Power Station will play an important role as the company doubles its wind power capacity over the next few years.” 

Planning for Growth 

Forty percent of Saskatchewan’s one million residents live in the two largest cities; the rest are scattered across an area nearly the size of Texas. It is a continual challenge to provide reliable power to residents far from the population centers, as well as to the oil field and mining operations in the western part of the province. As of late 2009, SaskPower owned and operated 3,172 MW of generating capacity, purchased another 469 MW and distributed its electricity over 156,000 km of power lines. Forty-five percent of that power comes from coal, 23 percent from hydro, 13 percent from gas and 5 percent from wind, with the remainder purchased from a variety of natural gas, wind and heat recovery sources

The two units at Kerrobert during construction.

To meet rising demand, SaskPower has embarked on a plan to add or replace 4,100 MW of generating capacity over the next 20 years and upgrade the transmission system. The details of the plan are laid out in a document titled Powering a Sustainable Energy Future – The Electricity and Conservation Strategy for Meeting Saskatchewan’s Needs submitted to regulators in October 2009. The plan is broken down into three parts: a short term plan covering 2009 to 2014 with a supply requirement of 1,091 MW; a midterm plan covering 2015 to 2022 with a supply requirement of 1,017 MW; and a long-term plan covering 2023 to 2032 with a supply requirement of 1,985 MW. Each of the stages includes generation and transmission targets, as well as demand-reduction programs that incorporate energy efficiency and load management and conservation activities.

SaskPower’s generation strategy is source neutral. In the last two years alone, it has added 400 MW of gas-fired generation, with more on the way, and is in the process of doubling its wind capacity. It is also spending C$354 million (US$364 million) revamping Unit 3 at the Boundary Dam Power Station. The coal-fired unit was scheduled for decommissioning in 2013, but instead will continue operation with a new Hitachi turbine designed for carbon capture. The long-term plan also includes examining the value of adding a nuclear power plant to the fleet. 

New Gas Plants 

While much of the new capacity, other than wind, consists of baseload plants, SaskPower is also adding peaking turbines. This is not simply a matter of providing reserve capacity, but also reducing the need to expand its transmission system.

One of the clutches during installation prior to alignment.

To meet the power needs of the mining and oil and gas industries, SaskPower runs dedicated lines to these facilities. For example, recent projects include a temporary 138 kV line for construction at a BHP Billiton potash mine, a 230 kV line to support a proposed Shore Gold Inc. mine and a pair of transmission lines to support TransCanada Pipeline’s Keystone pipeline to deliver oil from the oil sands to refineries in the U.S. These lines are all just a few kilometers long and tap into existing transmission lines nearby. A larger problem is how to get the power out to an area that is remote from the generating stations. It is more economical to have fewer and larger power plants than a large number of smaller distributed generation sources. The need to increase transmission capacity coupled with line losses over large distances can offset those savings.

SaskPower’s strategy to minimize the need to add transmission capacity, while also meeting peak power demands is to place peaking generators out near the remote loads. During non-peak hours, the customers draw power from the baseload plants and the peaking generators operate as synchronous condensers, providing the reactive power needed to boost the capacity of the transmission line. Then, during peak demand, they generate additional power locally. 

Peaking and Regulation 

SaskPower has already implemented this strategy at two peaking plants.

“Both plants are in the west side of Saskatchewan where there is a lot of oilfield activity and pipeline compressor activity and our transmission system was taxed to the maximum because of the growth happening out there,” said Terry Scott, who recently retired as a plant manager for SaskPower. SaskPower decided that rather than build transmission lines to support the voltage in that area it would deploy fast start, simple cycle gas turbines instead.

The first to go on line was the 92 MW Ermine Power Station near Kerrobert that went commercial in 2009, followed by the 138 MW Yellowhead Power Station in North Battleford, in December 2010. The $250 million Yellowhead plant has three GE LM6000 simple cycle turbines operated by satellite remote control from the province’s capital city, Regina. Upon receipt of a signal from headquarters, the quick-start turbines can be generating power in 10 minutes.

During non-peak times, the generators act as synchronous condensers, providing the reactive power necessary to support the voltage power coming from the baseload stations. Components of AC systems produce and consume both real and reactive power. Real power accomplishes useful work and is a product of the voltage and current. The voltage and current, although operating on the same frequency, are not in phase with each other. There is, however, an average power level which is measured in volt-amperes or watts.

Reactive power, on the other hand, has a zero average value, so it does not produce work nor does it travel over long distances. Since it has a zero average value, it is measured in volt-amperes reactive (VArs) which express the maximum value over a cycle, rather than the average. The VARs can be positive or negative, depending whether the current peaks before or after the voltage. Although reactive power doesn’t perform actual work, it is needed to provide voltage support and system reliability. Consuming reactive power lowers the voltage, while supplying it raises the voltage. In addition to preventing catastrophic voltage collapses, it also increases transmission efficiency, boosting the amount of real power that can be transmitted over a line.

When acting as a synchronous condenser, the generator draws current from the grid in order to keep spinning. Spinning both the turbine and the generator, however, is a waste of energy. Not only is there the additional mass of the turbine, but the windage losses from the blades. To keep power usage to a minimum, SaskPower installed clutches from SSS Clutch Co. between the turbine and the generator. SSS has provided hundreds of these clutches to the power generation industry for applications ranging from 500 kW to 300 MW.

For this type of application, the turbine is used to bring the generator up to synchronous speed, at which point it connects to the grid. If power is needed, the turbine continues to drive the generator. During off peak times, the turbine is shut down. As the turbine slows below the speed of the generator, the clutch automatically disconnects the turbine and generator and the generator draws just enough power from the grid to keep it synchronized and generating reactive power. Later, when peaking power is needed, the turbine is again brought on line and when it reaches the speed of the generator, the clutch links the two and the turbine again provides the power needed for generation.

Between Yellowhead and Ermine, SaskPower now has five LM6000s with SSS clutches in operation, with two more on the way for a Northland Power, an independent power producer under contract with SaskPower, which is installing the LM6000s at a plant in Spy Hill. 


EAM Tools Enable Just-in-Time Maintenance at Southern Power 

By Karen Threlkeld, Condition-based Maintenance and Predictive Maintenance coordinator, Southern Power

Southern Power, a subsidiary of Southern Co., is among the largest wholesale energy providers in the Southeast, meeting the electricity needs of municipalities, electric cooperatives and investor-owned utilities. The Birmingham, Ala., company owns and operates a fleet of generating facilities in Alabama, Florida, Georgia and North Carolina totaling 7,500 MW and has an additional 820 MW coming online in North Carolina and Texas in 2012.

Cohesive Information Systems helped ease the transition from manual processes to automated processes by creating schedules to maximize manpower and minimize material costs at Southern Power’s Plant Franklin.

All Southern Power Co. (SPC) plants, with the exception of a Texas facility under construction in Nacogdoches, house natural gas-fired simple or combined cycle combustion units providing supplemental base and peak load power. Nacogdoches will be the company’s first large-scale biofuel plant generating about 100 MW from wood fibers to serve the Austin metro area.

SPC’s Generation Support group recently completed a project to extract maximum value from maintenance dollars by automating predictive or preventive maintenance work orders using condition-based and run-time based triggering techniques. 

Hitting the Numbers 

Whether the metric is equivalent operating hours or starts, tracking turbine use to monitor wear and corrosion in a plant started up hundreds of times a year will have different preventive maintenance (PM) intervals for its pumps, valves, turbines and supporting components than a peaking plant running sporadically. Southern Power was having issues with the manufacturer’s recommendations for maintenance that were geared to base load operations. The service manuals didn’t fit how its plants operated—neither the combined cycle units that cycle almost daily, nor the peaking facilities that don’t get a lot of run hours.

The challenge was to devise a robust maintenance management system that could be applied uniformly across all locations but that factor in the different operating characteristics of each generating asset. This would enable PM service intervals based on run hours, starts or combinations of data translating to reliable operation and getting the full lifespan from each asset.

SPC already had real time data visualization of its plant assets through an operation information system (OIS) based historian. What it needed was a system that could automatically interpret this data into work orders for scheduled PMs to avoid unplanned downtime and reactive maintenance. Attempts at reaching this end were made using several off-the-shelf products but results weren’t satisfactory. Turning to the IT group for assistance, Generation Support considered building a custom solution that would be easy to use but not cost a lot to maintain.

Maximo enterprise asset management (EAM) software had been in use at Southern Power operations for years to manage aspects of maintenance, inventory, purchasing and other tasks, so it was logical to build a tool off the platform for this project, designated the “OIS to Maximo Interface.” 


Maximo has flexibility to integrate various enterprise systems, but creating a custom web-based interface between two different systems is complex and time consuming. Southern Power knew it could be done but wasn’t certain about execution. The task was assigned to Cohesive Information Solutions of Kennesaw, Ga., known for its knowledge of Maximo and business processes.

Goals, system requirements and timetables were documented ahead of the physical work. Build-out and configuration proceeded over a period of months, following a schedule that accommodated the creation of multiple tags for each piece of equipment and putting the new interface through a series of predefined test cases. The interface passed and went into production after about a month of testing.

To build the interface, Cohesive engineered a series of web services to query tag information accumulated in OIS, such as starts and hours of operation. The interface, or “cross walk” allows all of these condition code attributes from the asset ID (tag points) to update a corresponding EAM “meter” attached to various equipment records. The seamless data exchange populates tables set up in Maximo that can generate appropriate preventive maintenance tasks for each piece of equipment, determined on condition based information or time intervals.

This is helpful for managing a diverse generating fleet and supporting infrastructure. Meter data, for example, can create a PM to inspect a combustion unit’s R0 blades every 50 starts. That’s important for equipment in regular use. But at a peaking plant that might be only used a few days out of the year, regularly checking the oil at certain time intervals is a more critical parameter.

Don’t Overdo It 

Cohesive worked with SPC to transition from a manual process to an automated system for getting the most out of existing manpower and controlling material costs. Equipment tolerances were reviewed to devise a more pragmatic schedule that avoided “over maintaining” assets regardless of their actual condition. Too much preventive maintenance can zero out time and cost savings that would have otherwise been achieved with automation, without significantly reducing risk. The automated preventive maintenance program based on efficient, just-in-time service helps ensure reliability while removing the potential for human error to counter balance risk.

Using existing products also helped SPC avoid the expense of license acquisitions, training and maintenance, one of the project goals from the outset. Technicians now log into the familiar, secure web portal for automated work orders across the Southern Power system. Worker efficiency and utilization are enhanced by the removal of guess work in scheduling or the work being performed. Personnel know the data is trustworthy.

Careful planning and good communication among all parties led to successful completion of the project for real-time data exchange between the OIS and Maximo. The net result of joining the two platforms is a holistic architecture producing a preventive maintenance program that accounts for how Southern Power combustion units are used. Leveraging meter data in terms of PM frequency helps them reduce the likelihood of a forced outage/reactive maintenance event, which in turn reduces cost.

Author: Karen Threlkeld is the CBM/PdM coordinator for Southern Power in Birmingham, Ala. 

Redefining Valve Zero- Leakage Standards 

By Kevin Hunt, President, ValvTechnologies

One of the biggest stories of 2010 involved valve failure. From April to July, engineers were scrambling to stop the hundreds of millions of gallons of crude being released into the Gulf of Mexico after the blowout preventer on the Deepwater Horizon rig failed to seal the pipe. Of course, major disasters are typically preceded by known, but uncorrected errors, and this was not the first time BP had trouble with isolation valves. As The Washington Post reported in June 2010, a 2001 operational integrity report covering BP’s North Slope operations found that, “Workers believe internal leak-through of isolation valves is a significant problem and under certain circumstances may pose a potential hazard to workers and equipment.”

A leaking globe valve.

Such incidents attract a lot of attention, but most leaks are well hidden somewhere deep inside the equipment and piping that cover the grounds of a power or petrochemical plant. These valves gradually eat away at performance and profits, a problem that is particularly critical with severe service isolation valves (SSIV). These issues, though, are preventable by using properly designed valves that allow zero leakage. The problem is that “zero leakage” does not always mean zero leakage. The usual definition actually means acceptable slow leakage internally, with no visible external leakage. By applying best available technology and adopting new standards, however, zero can equal zero, both internally and externally. 

Slow Erosion 

SSIVs are isolation valves that are used in high energy conditions including elevated temperatures and/or elevated pressures. They are used for numerous applications throughout power plants. These temperatures and pressures exceed the normal operational limits of thermoplastic seals, so the seal needs to be made by the metal components of the valve. This requires precision manufacturing beyond the standards generally applied to valves that can use a flexible or compressible gasket to provide the seal. Further complicating the matter, the fluids involved may contain some solid content or abrasive materials that erode the seal surfaces, thereby producing leakage paths.

With SSIVs, the cost of the leakage is far greater than the cost of the valve. High temperatures and pressures coupled with erosive substances entrained in the fluid means that even minor leaks can grow into major ones. This results in unscheduled shutdowns and frequent equipment repair or replacement as well as wasted fuel/process liquids. To protect people from injury and equipment from damage, it is essential to achieve zero leakage with SSIVs.

Valve leakage can be either internal or external. It isn’t difficult to discover external leakage, evident, for example, by the steam plume escaping from the packing or gasket areas.

Internal leaks are entirely different. Take an isolation valve on a bypass line between the boiler and the steam turbine that redirects steam to the condenser. Any leakage though that SSIV lowers generator output while increasing fuel consumption. Since leakage is internal, it may only show up as a gradually increasing heat rate, requiring additional fuel expenditures and accompanying emissions remediation expenses to produce the same amount of electricity. The cost of replacing a faulty valve, on the other hand, is minimal compared to the lost output.

Keep in mind that there can be hundreds or thousands of SSIVs in a power plant. The overall losses are not from a single leaky valve, but the aggregate losses from each of them leaking a tiny amount. Together they add up to millions of Btu’s never reaching the turbines. Zero leakage SSIVs can typically improve a plant’s heat rate performance from 1 to 2 percent to as high as 5 to 6 percent. 

Loss of Power and Profits 

To see how that impacts the bottom line, consider two scenarios: A 1,000 MW coal plant (annual fuel cost $150 million) and a 1,000 MW natural gas plant (annual fuel cost $300 million), each with 1,000 SSIVs. Those valves together, since they provide imperfect seals, cost the plant about 3 percent of its efficiency, adding $4.5 million to the price tag for operating the coal plant and $9 million to the gas plant. Given an average replacement cost of $4,500 per SSIV, it would cost $4.5 million to replace all valves, a payback period of 12 months for coal and 6 months for natural gas.

A worker performing the Ultrasonic Emissions Testing program (Total Valve Management ‘TVM’) on a valve which measures internal leakage rates.

Following the 80/20 rule, if you replace 20 percent of the valves and eliminate 80 percent of the leakage, the replacement cost for 200 valves would be $900,000. The payback time would be 3 months at the coal plant and 1.5 months at the natural gas plant. These figures do not include the additional benefits of increased output, reduced emissions, smaller amounts of downtime and lower repair costs. Similar arguments will pertain to the nuclear industry with the proper information.

These losses may seem extreme at first glance, but are validated by other well-established research in areas such as leakage through an orifice in a pressurized pipeline, as well as heat and pressure losses in steam traps. Per the ANSI/FCI 70-2 leakage specification, a Class V valve should have a maximum seat leakage of 5 x 10-4 ml per minute of water per inch of seat diameter per psi differential (5 x 10-12 m3 per second of water per mm of seat diameter per bar differential). Buying Class V valves, therefore, would seem to eliminate the leakage losses. However, those standards apply only at the point of installation.

Over time, the continuous steam leakage past the plug seat erodes the seal, causing steam cutting and wire drawing. What was once a Class V valve will eventually degrade to the point where the valve completely ceases to perform its isolation function. To try to seal against the high pressures (a 2” ANSI 4500 globe valve is subject to up to 19,623 pounds of force), a hammer-blow handwheel is sometimes used. This method uses up to 10 times greater torque to drive the plug against the seat, but the method can damage the valve parts and won’t stop leakage through an eroded seal. Additional damage comes from vibration, flashing, cavitation and internal erosion. 

Creating a New Standard 

Just as utilities must apply Best Available Control Technologies (BACT) to eliminate excess emissions, so should they adopt Best Available Isolation Technology (BAIT) design features to eliminate the problem of erosion and gradually rising losses. Here, for example, are some of the elements that make up BAIT for ball valves:

  • An Integral Seat: The integral seat is part of the valve body rather than a slip-in seat ring. A slip- in provides a leak passage behind the seat, which doesn’t exist with an integral seat.
  • High Strength Belleville Seat Springs: Belleville springs are cone shaped washers that apply a constant high thrust to create a mechanical preload on the ball and seat, and the packing. By stacking several of the washers, you can increase the deflection, while keeping the load on each washer constant. Use of Inconel 718 produces a high tensile strength (155,000 psi) with a high yield strength (125,000 psi) and high creep strength. This allows a seat spring compression of several hundred PSI to fully position the ball against the seat, preventing ball misalignment or vibration and restricting particles from entering and damaging the seal.
  • Full Alignment/Positioning of Ball and Seat
  • RAM (HVHF) Hardcoating: RAM is a High-Velocity Hydrogen Fuel (HVHF) coating process that uses a hot, high-velocity gas jet to spray a coating of molten particles on to the ball and seat surfaces. Traditionally disks and seats of carbon, alloy or stainless steel are hardfaced by welding on an overlay of Stellite, a cobalt-chromium allow with good wear and corrosion resistance properties. However, above 800 F, Stellite begins to soften at an alarming rate and is subjected to galling and heavy wear and tear of the valve seating surfaces. Rocket Applied Metallic (RAM) is both harder than Stellite and maintains its hardness at high temperatures. At room temperature, RAM 31 has a Rockwell C hardness of 72, vs. 39.8 for Stellite 6. At 1,400 degrees F, RAM has a Rockwell C hardness of 62, compared to just 8 for Stellite. RAM is also self-repairing in operation, where one million leak-free valve cycles are possible.
  • Mate Lapping of Ball and Seat: The ball and seat must be precisely mated to each other to form a perfect seal. This is accomplished by lapping the ball and integrals seat for several hours on a rotating fixture. The final step involves using a 3-micron diamond compound and moving the ball in a figure-8 motion.
  • Blowout Proof Stem: A typical valve stem is externally inserted and uses a slip-on collar held in place by a pin. This is a weak link that can possibly pose a hazard to life, limb and property. A blowout proof stem is internally inserted and has an integral shoulder rather than a collar. Since it is internally retained, it is virtually 100 percent blowout proof. 

A leaking HRSG drain valve.

Adopting BAIT makes it possible to actually achieve absolute zero leakage on SSIV. Therefore, it is time for a new valve classification that goes beyond the FCI 70-2 Class VI standard. Class VII would be zero visible leakage for three minutes using the ValvTechnologies VQP-10 hydro test and VQP-10 gas test. With this new standard, zero means zero, not something that we hope comes pretty close. 

FGD Reagent Supply 

By Greg Andersen, FGD Commercial Development Manager–Reagent Technology Services, Mississippi Lime Co.

Back in the mid 2000’s coal fired electric utilities, primarily in the eastern half of the United States, basked in an era of limited regulatory uncertainty. The Environmental Protection Agency’s Clean Air Interstate Rule (CAIR) had not yet been overturned by the courts and the train wreck of other regulatory initiatives was but a faint and cloudy worry on the utility planner’s horizon. Most of these utilities were already several years into a major but methodical process of evaluating and implementing sulfur dioxide (SO2) and nitrous oxide (NOx) control strategies on those units of their coal fleet where it made most economic sense. More often than not the result was a multi-year effort to install wet limestone flue gas desulfurization (FGD) systems and selective catalytic reduction (SCR) systems on a number of their larger coal-fired plants.

Mississippi Lime milling facility.

However, because of the demand for specialized expertise and exotic materials and the then-booming economy, the cost of these systems was escalating. In addition, regulated utilities knew these costs would be scrutinized by rate regulatory bodies.

A number of utilities, including Southern Co., began to explore a non-traditional design strategy whereby they could eliminate the cost of the entire limestone receiving, storage and wet milling facilities at several plants within the same geographic region and replace them with pre-ground limestone from an outside supplier. This reagent would be produced, delivered and stored at the power plant on a dry basis and only mixed into a slurry as needed.

This idea was not entirely new. One utility in the upper Midwest had successfully used such a strategy at two plants near the Great Lakes nearly 15 years before. Depending on the number and size of the plants involved, the savings in capital could range anywhere from the low 10’s of millions of dollars to nearly $100 million. Clearly this was an idea that merited additional study.

Reducing capital expenditures is not the whole story. The pre-processed limestone reagent would cost more than the crushed limestone received under the more traditional design. However, other operating, energy and maintenance savings associated with eliminating these major subsystems would work to offset this price differential. There were other benefits and concerns that had to be addressed as well. To properly and prudently evaluate such an alternative would require a more sophisticated level of economic and operational analysis and review.

Southern Co. was one of the utilities that conducted such an evaluation. They had already decided that some of their largest plants would best be served by installing wet limestone scrubbers. As part of their design process, they conducted an analysis to determine the most economic method of supplying limestone slurry to these scrubbers. Southern’s project engineers and procurement analysts compared the traditional wet ball mill approach with the pre-pulverized option for various plants in their system, at times concluding that pre-pulverized reagent was the better operational and economic choice. As a result, they entered into pulverized limestone supply agreements.

Southern also concluded that they needed to install scrubbers at two and possibly three coal plants on the Gulf Coast. Unfortunately, because suitable limestone is virtually non-existent in this area and there are no corresponding grinding facilities nearby, it appeared that these plants were candidates for the traditional wet ball mill approach.

Mississippi Lime Co. and its Reagent Technology Services (RTS) Division came into the picture. The RTS model was to locate regional grinding facilities where it made the most sense for the customer(s). To do so it chose not necessarily to locate the grinding plant at the stone source as had been traditionally done but rather in the location that best met the utility’s needs.

When RTS began to discuss Southern Co.’s needs for the Gulf Coast, it was not long before they concluded that a regional grinding facility in the Mobile, Ala., area would allow RTS to provide pulverized limestone supply.

Mobile’s excellent logistics infrastructure, particularly as it relates to bulk water transport by barge and ship, made it feasible to supply crushed FGD-quality stone from relatively remote locations.

Before deciding that this was the best alternative, Southern insisted that a number of key operational issues be addressed in detail. For example:

How would RTS design the plant to withstand the threat of a major hurricane and what would RTS do to keep a Category 3 hurricane surge from taking the plant out of service?

  • How would RTS ensure sufficient inventory was available to meet sudden increases in demand and what impact would that storage capacity have on the ability to get the material out of a silo?
  • How would RTS design the plant to ensure RTS could meet demand if a catastrophic failure were to occur to the milling device?
  • How did RTS plan to ensure consistent and reliable delivery and unloading of reagent supply? 

Some issues involved solutions that required actions on the part of the power plants. Therefore, RTS committed to detailed, ongoing collaboration and coordination.

Ultimately, RTS assured Southern that all these concerns could and would be successfully addressed. In November 2008 RTS began construction of the Mobile Regional Grinding Facility.

The milling facilities are built at an elevation of 14.6 feet above sea level. Most of the city of Mobile would now be under water before the plant would ever see a drop of flood water.

The plant began startup activities in October 2009 and initiated routine deliveries on Jan. 1, 2010. 

Benefits Analysis 

Intended and unintended benefits associated with a non-traditional pulverized limestone supply are described in the paragraphs below.

Capital Savings: The most obvious benefit is the reduction in capital required by the utility. Depending on the size, complexity and redundancy of the traditional design, capital requirements often can be reduced by $20 million to $30 million or more per plant. When more than one plant is being served, these savings are correspondingly multiplied. Dry roller mills themselves, including ancillary equipment are less costly to install. A recent study estimated that capital costs for a dry roller mill system are approximately 20 percent lower than a comparably sized wet ball mill system.

O&M Cost Savings: With the reduction in capital equipment there is a corresponding reduction in operating and maintenance costs, including labor. Operating a wet ball mill system requires ongoing diligence to ensure the system produces quality pulverized limestone slurry. Operators must pay attention to the entire system and make adjustments for ongoing wear to or malfunction of the equipment. Something as simple as plugging of a hydroclone can cause reductions in efficiency that are hard to detect. Startup and shutdown is much more involved and piping systems must be routinely flushed to guard against line pluggage. Grinding media must be routinely replaced. Ongoing maintenance and periodic major rebuilds and ball mill liner replacement are required.

Reduced Parasitic Power Consumption: Eliminating the limestone receiving and processing system also results in a significant reduction in station service. This, in effect, results in a capacity increase of 2 to 4 MW for each plant, which can be incredibly valuable in times of extreme customer demand where high cost generation alternatives must be used to meet the demand.

Equal or Greater Reliability: Numerous utility evaluations have shown that, by a combination of milling equipment, backup sources of reagent and storage of finished product both on-site and at the milling location, a pulverized limestone supply can offer at least equivalent and in some cases greater reliability than an on-site wet ball mill. Some utilities are now considering backup supply of pulverized limestone at their wet ball mill plants for emergency backup and operational flexibility.

Superior Quality Control and Increased Flexibility: The ability to analyze key parameters in advance allows the scrubber operator to know the physical and chemical characteristics of the reagent before it is received. If a delivery fails to meet requirements it can be rejected. Once the pulverized reagent is on site all that remains is a simple mixing of reagent with water, potentially on a just-in-time basis, to produce reagent slurry. Frequent testing ensures that the limestone will perform as desired within the scrubber to minimize the stoichometric ratio, maximize SO2 capture and meet all gypsum product requirements. The pulverized limestone reagent is delivered bone dry so the utility does not pay for the 3 to 4 percent or more moisture typically present in crushed limestone.

Ability to Serve Multiple Facilities: There are other economic advantages to utilizing dry roller mills. Dry roller mills can be installed at the best location to optimize the logistics costs, whether at the plant, the quarry or somewhere in between. Also, one regional grinding center can supply multiple power plants, eliminating unnecessary redundancy.

Limestone Supply Independence: It is possible with a customer oriented regional grinding facility to have access to multiple limestone deposits, rather than relying on a single limestone source. The Mobile plant has several limestone sources that are approved for use by Southern Co.

Reduced Environmental Impact: Eliminating the limestone receiving, storage pile, reclaim and conveying systems will reduce overall particulate emissions. Pulverized limestone is typically pneumatically conveyed at much lower emissions. In some cases these reduced emissions can be important in a permitting context.

Reduced Footprint: To utilize pulverized limestone simply requires a receiving silo and mixing system, which has a dramatically smaller footprint.

The availability of an economic reagent supply alternative that can reduce project capital requirements could play an important role in future control decisions. 

Boiler Derates Caused by Inadequate Fabric Filter Performance: Lessons Learned at a Midwest Utility Plant 

By Michael Johnson, Senior Product Application Manager, GE Energy Environmental Services and Michael McMenus, Environmental Compliance Administrator, Kansas City Power & Light

The Kansas City Power and Light (KCP&L) Hawthorn Station Unit 5 is a nominal 550 MW, Powder River Basin pulverized coal-fired boiler equipped with an SCR, air-preheater, two spray dry absorbers (SDA) and a B&W pulse-jet fabric filter (PJFF) to maintain PM10 emissions below 0.018 lb/MMBtu and SO2 emissions below 0.12 lb/MMBtu (30 day rolling average).

The original filter bags supplied by the manufacturer were a nominal 16 oz/yd² scrim supported PPS needle felt with a dipped Teflon coating and experienced a bag life of approximately 3 ½ years, which is beyond the normal industry bag life for this filter media in this application. Nominal bag size was 150mm diameter by 8 meters.

Between the 30th to 38th months of operation the system pressure drop had gradually increased to over 10” w.g. and even with constant, rapid pulse cleaning at 100 psi, the unit could not recover sufficiently, which resulted in the boiler being de-rated on several occasions.

Several original filter bags were sent to an outside laboratory for analysis where it was determined that the manufactuer’s needle felt PPS filter bags had developed a heavy and tenacious dust cake on the surface which resulted in a significant decrease in fabric permeability causing the high differential pressures. The condition made the filters unrecoverable and in late 2004 the plant decided to change out all 13,520 filter bags.

KCP&L determined that the original PPS filter bags met all regulatory requirements for PM (filterable and condensable) and SO2 emissions but wanted to evaluate the opportunity to increase filter bag reliability and decrease total operational costs, while continuing to meet and exceed the regulatory requirements.

KCP&L and GE met to discuss the potential to improve the overall operational reliability of the PJFF system with an emphasis on methods to increase the original filter bag life. Operational data and independent lab tests were reviewed.

After this full system evaluation, GE recommended that KCP&L consider upgrading the filter bags to a 15.5 oz/yd², scrim supported, proprietary blended PPS filter media laminated with a high durability ePTFE to the collection surface.

What is ePTFE Membrane? 

Expanded polytetrafluoroethylene (ePTFE) membrane is a technology that provides some of the highest efficiency of any available filtration media. Through a controlled manufacturing process, PTFE resin is expanded into a membrane film, composed of millions of microscopic pores in a three-dimensional web-like structure. These micro pores are small enough to capture sub-micron ash, yet large enough for the passage of airflow. The membrane can be laminated by various methods to most available needle-felt and woven filtration medias.

To confirm the efficiency of the suggested filter media, test data was reviewed with KCP&L comparing both the original OEM needle-felt PPS and PPS with ePTFE membrane. Testing was done internally using the same protocol used in the Environmental Protection Agency’s Environmental Technology Verification (ETV) program. 

Test Protocol and Results

Testing of KCP&L’s original OEM filter media along with GE’s recommended 15.5 oz/yd² PPS with ePTFE membrane filter media was performed in-house following testing protocol used in EPA’s ETV program using the same type of equipment utilized by EPA/ETV and in accordance with ASTM Test Method D6830-02 along with the test specifications and conditions as detailed in Generic Verification Protocol for Baghouse Filtration Products (BFP). The protocol was adopted from German VDI Method 3926 and modified for the ETV.

A 6-inch-diameter fabric filter sample is challenged with a standard dust (particulate matter) under simulated baghouse conditions at specified rates for air and dust flow. The test consists of three test runs. Each run consists of three sequential phases or test periods during which dust and gas flow rates are constantly maintained to test specification. The test phases are:

  • A conditioning period of 10,000 rapid pulse filtration cycles (every 3 seconds)
  • A recovery period to allow the test sample to recuperate from rapid pulsing where the filter is pulsed only when the differential pressure reaches 4” w.c.
  • A 6-hour performance test period during which measurements for particulate emissions are determined by gravimetric measurement of the particulate matter that, passes through the sample. Particulate used for the test is 1.5 micron mass mean diameter with at least 50 percent less than 2.5 microns. 

Test Conditions Throughout the Test Were as Follows:

  • Test dust: Pural NF Alumina (1.5 ± 1 micron mass mean diameter)
  • Inlet dust feed rate: 8.0 ± 1.6 gr/dscf (18.4 ± 3.6 g/dscm)
  • Filtration velocity: 6.6 ± .5 fpm (120 ± 6 m/hr)
  • Gas temperature: 78 ± 4 F (25 ± 2 C)
  • Pulse cleaning pressure: 75psi
  • Testing was conducted to determine the filter sample’s performance with respect to the following parameters:
  • Outlet particulate emissions (PM 2.5)
  • Outlet particulate emissions (total mass)
  • Initial residual pressure drop
  • Increase in residual pressure drop
  • Average residual pressure drop
  • Mass weight gain of the filter sample
  • Average filtration cycle time
  • Number of filtration cycles 

In December 2004, KCP&L made the decision to rebag the entire PJFF with the recommended 15.5 oz/yd², scrim supported 100 percent PPS with ePTFE membrane laminated to the filtration surface. 

Filter Bag Media Performance 

The Hawthorn Unit 5 PJFF operation was reviewed over time comparing the pressure drop, cleaning frequency, gas flow and other parameters between the two styles of filtration media. The most obvious improvement for KCP&L was a filter bag life from 38 months to 62 months while still compliant with all regulations.

In April 2005, a small leak in a filter was detected in compartment No.5. When this compartment was isolated to change the filter bag, the pulsing system was inadvertently left off on the entire baghouse for a prolonged period of time and the pressure drop across the baghouse system exceeded +15”w.g. before the fan tripped out (fan set to trip at 16.0”w.g.). Once the pulsing system was turned back on, the filters were able to recover back down to the 6.0” to 7.0”w.g. established as the target PJFF system pressure drop between KCP&L and GE Energy.

On several occasions there were ammonia slips upwards of 10 ppm typically as a result of SCR catalyst erosion. Ammonia slip can have a substantial negative impact on filter bag performance in the form of high differential pressure and difficult recovery. The ammonia slip forms ammonium bisulphate and creates a dense and sticky cake structure as shown in the picture below.

When these events occurredthe differential pressureincreased across the PJFF unit, but came back down within a reasonable time to a normal operating differential pressure simply because the membrane surface isa slick surface that sheds the dustcake much better than the original OEM, non-membrane PPS filter bag.

PJFF system pressure drop averaged between 6.0”w.g. to 7.0”w.g. over the life of the membrane filter bags compared to over 10”w.g. at the end of the life cycle of the non-membrane PPS filter bags. This consistent differential pressure across the PJFF system allowed KCP&L to operate continuously at the required load without any derates of the boiler due to the PJFF system.

Pulse cleaning pressure used to clean the filters was reduced from up to 100 psi with the original supplied non-membrane filter bags to 75 psi during the life cycle of the membrane filter bags. The baghouse cleaning cycle, which is defined as the amount of times the baghouse cleans each filter per day, averaged between 45 to 55 cycles. During the last 30 -to 38-month life of the original filters, the baghouse was in a continuous cleaning mode, pulsing approximately between every 30 seconds down to 10 seconds.

And finally, PM10 emissions remained well below the 0.018 lbs/MMBtu regulation limit through out the 62-month life cycle. 

Operational Results 

PM10 emissions test data was performed annually on both types of filter medias and documented. The ePTFE membrane PPS filter media performed well below Kansas City Power & Light’s PM10 regulation of 0.018 lbs/MMBtu and at a fairly level rate over the entire 62-month life cycle.

Data on differential pressure measured across the PJFF system was acquired from KCP&L DCS. Data is not available from years 2000 through 2002 due to the unavailability of the data from the DCS. During the last two years of data collection from the original PPS filters, the adverse effect of filter age (especially 30th to 38th month), ammonia slips and filter media bleed through causing an average of +10”w.g. across the system resulting in de-rating of the boiler on several occasions.

In comparison, the PPS membrane filters show a more consistent system pressure drop over the 62-month life cycle with an average between ~6.0”w.g. and ~7.0”w.g. During the evaluation of filter media candidates, KCP&L took into consideration the energy cost reduction with the potential lower system pressure loss as part of the economic evaluation between the two medias as well as expected life expectancy and PM10 efficiencies. 


Long-term operational results conclude that upgrading the original PPS media to PPS with laminated ePTFE membrane filter media provided KCP&L Hawthorn 5 PJFF system a more reliable system. Filter life increased from 38 months to 62 months, lower operational costs were realized, PM10 regulations were well below limits and no boiler de-rates were experienced due to high differential pressure drops across the PJFF. 

More Power Engineering Issue Articles
Power Engineerng Issue Archives
View Power Generation Articles on