HRSG Users: The only user meeting focusing on Rankine-cycle solutions

The HRSG User’s Group must be doing something right. It was the only user group in the gas-turbine-based gen­eration sector of the electric power industry to ride through the eco­nomic downturn last spring without a noticeable drop in attendance. The 2009 meeting at the Hyatt Regency Jacksonville Riverwalk hosted 350 participants, only seven fewer than last year’s record 357.

The 83 exhibitors at the vendor fair, which was open on each of the meet­ing’s three days, topped the 2008 total of 69. There were seven sponsors.

Organization of the confer­ence program hasn’t changed much in the last few years. A day-long pre-conference seminar is offered on Mon­day. Theme this year was supervising outage contrac­tors. A reception and dinner beginning at 6 pm, when the exhibit hall opens, is the offi­cial start to the meeting.

Four formal presentations anchor the open-forum dis­cussions on the second and third days. This year’s spe­cial presentations:

  • “Digital asset intelli­gence: Bottom up meets top down,” Jason Makansi, Pearl Street Inc.
  • “Using real-time software tools and training to make rapid and informed plant decisions,” John Koza, General Physics Corp.
  • “Chlorine dioxide: A unique solu­tion to cooling system biofouling,” Gary Laxton, Baker Hughes Inc.
  • “Stack damper retrofit: A case study,” Yogesh Patel, Tampa Elec­tric Co, and Bill Kitterman, Brem­co Inc.

Topics for the open-forum discus­sions are the same each year, but subject matter addressed differs in response to audience interests. The discussion sessions are:

  • Heat transfer equipment.
  • Water treatment.
  • Piping systems.
  • Controls, ductwork, dampers, and stacks.
  • Valves and supplementary firing.
  • Environmental systems and bal­ance of plant.

Formal presentations

Makansi got the meeting rolling after the Tuesday morning coffee break with an overview of powerplant con­trols, automation, monitoring, diag­nostics, and software and where they’re all headed. He began with the plant’s viewpoint on software:

  • CMMS (computerized mainte­nance management software) and task management software is “a pain in the ass, but a necessary evil.”
  • The data historian, usually OSI­soft LLC’s PI System™, has become the communications “con­duit” and typically is praised by plant staff.
  • User interfaces and user friend­liness reigns supreme when it comes to software features and functionality.
  • Integrating different software packages to facilitate plant use often presents challenges. Conse­quently, there are “islands” of auto­mation, digital applications, and software at virtually every plant.
  • Plant “models” (first principles, data analytics, statistical, etc) have value at the deck-plates level, but few plant personnel understand the fundamentals and benefits, and know how to use them.
  • Central/corporate IT is increas­ingly driving plant software deci­sions.

Makansi offered his views on the state of software development for combined cycles—heat-recovery steam generators (HRSGs) in par­ticular. He believes HRSGs have been “neglected,” compared to gas and steam turbines, when it comes to electronic tools for assist­ing in O&M decision-mak­ing. Makansi said there are opportunities in the areas of condition assessment, online corrosion monitoring, water chemistry, and elsewhere.

Some headway already is being made in these areas. For example, shortly before the conference, Vogt Power International Inc told the editors about the positive beta test results for its Life Consumption Assessment and Monitoring Program (LCAMP™) at The Southern Company. This new software tracks life remaining in criti­cal HRSG parts and is valu­able for its ability to avoid forced outages, assist in plan­ning boiler work for planned outages, minimize life con­sumption during startups, etc. For details, see the sidebar begin­ning on p 146.

Makansi then took a top down view from the executive suite, highlight­ing some of the challenges facing the leadership of generating companies and how they might influence plant life. Mentioned were relentless cost pressures, the looming need for cur­tailing CO2 emissions, postponement or cancellation of new plants based on revised demand forecasts, the need to improve the performance of genera­tion assets to shore up balance sheets, coexisting with renewables on the grid, and staffing.

Among the solutions suggested was a digital asset intelligence system distinct from the physical assets. This would include the deployment of formal knowledge management processes for capturing, managing, and propagating human intelligence. Makansi stressed the need for more software demonstrations and validations before deployment, more and better sensors and final control devices, communication standards.

To dig deeper into some of the ideas and concepts offered by this speaker, listen to the webcast, “The Per­fect Power Plant,” at Background on the emerging field of knowledge management is avail­able in “Actionable Intelligence,” a new book by Dr Rob­ert Mayfield, a plant manager at an F-class combined cycle and member of the 7F Users Group steering com­mittee (details on p 118).

Informed decision-making

Koza’s presentation after lunch was the perfect comple­ment to Makansi’s “scene setter.” It offered a close look at some software tools for enabling plant-level decision-making. Koza began by discussing some performance challenges facing managers of combined-cycle plants, including:

  • Lean staffing compared to conventional steam/electric plants.
  • Fewer operator controllable parameters, which Koza attributed to plant design and sophisticated control systems.
  • The need for personnel to quickly and confidently evaluate operational changes and performance defi­ciencies to assure plant profitability.

Koza stressed the need for a systematic approach to assure success. First step, he said, was to establish the process: Identify key performance indicators (KPIs), develop design and tuned models, use software tools to monitor KPIs in real time, and integrate training for all involved.

Next, implement a performance- and condition-mon­itoring methodology. Doing this properly requires that you identify potential problems and define their relative impact on goals management expects you to achieve. Then you have to decide on the data to evaluate and to identify the actionable information. Monitoring of KPIs in real time, Koza continued, demands accurate, tuned models of plant processes. Fixed-deviation and rate-of-change alerts are valuable management resources.

A series of case studies illustrated how the informa­tion provided by the software tools can be used to gain competitive advantage. The first illustrated the use of data analysis to identify “zero-cost” improvements—such as control changes, process-setpoint adjustments, changes to operating procedures that improve perfor­mance, etc.

The example was an evaluation of evap cooler opera­tion at part load in a western plant. Staff learned that below a certain load point the evap cooler could be turned off and the loss in GT performance would be more than offset by gains in HRSG steam production and steam tur­bine/generator output.

Another case study illustrated the value of data analy­sis for scheduling routine maintenance with maximum ROI (return on investment). The example: Cleaning of an air-cooled condenser to improve heat rate. Data on the performance gain and the cost of cleaning enabled plant personnel to develop an optimal cleaning schedule.

Case study 3: Optimum scheduling of outages to restore lost performance. Calculated condenser cleanli­ness decreased over a period of several months. During a brief outage, plant maintenance personnel inspected condenser waterboxes and found the inlet waterbox and tubes clean, but the return and outlet waterboxes fouled.

The fouling was microbiological growth not treatable with bleach. Proper scheduling of cleaning elimi­nated derating of turbine backpres­sure during periods of peak load and increased the bottom line by $1.5 mil­lion annually.

Case study 4: Comparing actual operating data with that predicted by the plant model indicated a devia­tion in HP turbine steam flow and pressure. The O&M staff tracked the gremlin to the HP stop valve, which had a clogged basket strainer that was costing $300,000 in lost perfor­mance on an annual basis.

Biofouling control

Chlorine dioxide might have received more than its fair share of podium time at a meeting that serves only GT-based cogeneration and combined-cycle plants, many of which don’t have significant biofouling issues.

Perhaps a backgrounder on the subject of biofouling would be of greater value. Biofouling takes two forms: microbiological fouling of heat exchangers and macrobiological foul­ing of intake and discharge canals. The former is caused by both plant and animal organisms, such as algae and bacteria, while the latter stems primarily from invertebrate life—mussels, clams, etc.

Protection against biofouling falls into three broad categories: mechani­cal, thermal, and chemical. Mechani­cal methods for preventing organism entry into cooling-water systems are long-standing, and effective primar­ily against macroscopic life forms. They include trash racks, bar and traveling screens, strainers, scrap­ers, brushes, etc.

Biota getting through to the con­denser may be removed (1) by online flushing and flow reversal, (2) with scrapers and brushes, and (3) by elevating the temperature of of the circulating water. Holding tempera­ture above 105F for a couple of hours will kill at least some problematic shellfish.

Chemical treatments—that is, microbiocides—are effective against microorganisms; the only questionable aspect is the possible environmental effect of treatment chemicals that persist beyond the plant boundaries. Some microbiocides alter the perme­ability of microbe cell walls, thereby interfering with vital life processes. Other chemical agents are designed to inhibit microorganism growth.

Chemicals used are either oxidiz­ing or non-oxidizing. Of the oxidizing chemicals—those that irreversibly oxidize protein groups, resulting in loss of normal enzyme activity—chlo­rine is the most familiar and effective. Alternatives include chlorine dioxide, bromine chloride, chlorinated bromine salts, and some others. Interestingly, at the 2009 Southwest Chemistry Workshop, which reflects experience at nuclear and conventional fossil-fired powerplants, as well as GT-based generating units, chlorine was the most popular biocide (p 66).

However, chlorine has limited effectiveness in high-pH (above 7.5) water and in reclaimed water contain­ing ammonia. Bromine-release agents are preferred by some chemists under these conditions. Bromine oxidant is generated by reaction of bromide salts with chlorine gas or hypochlorite liq­uid, producing sodium hypobromite or hypobromous acid.

Chlorine dioxide also is effective at high pH and does not react with ammonia. It is generated onsite. A claim made in the HRSG User’s Group presentation was that a new fully automated process for making chlorine dioxide eliminated the need to mix chemicals and that the bio­cide was more cost effective in many instances than chlorine, ozone, bleach, bromochlorodimethylhydantoin, and hypochlorous acid and sodium bro­mide.

The bottom line: There are several alternatives for controlling biofouling. The best one for your application will require evaluation of relative effec­tiveness (be prepared for at least some laboratory-scale testing) and cost.

Stack damper retrofit

The benefits of stack dampers, which include faster startups and reduced corrosion on the gas side of HRSGs, have been discussed at the group’s annual meetings for about as long as anyone can remember. But until this year, relatively little first-hand expe­rience and hard data found their way into the floor discussion.

A formal presentation on the ret­rofit of dampers in stacks serving seven F-class GTs at Tampa Electric Co’s Bayside Power Station offered user attendees sufficient information to decide if a similar value proposi­tion existed at their plants. A sum­mary of the presentation is available in the sidebar that begins on p 150.

Heat transfer

The first time you attend an HRSG User’s Group conference and head off for the opening session, you wonder why a boiler meeting is starting with feedwater heaters, lube-oil coolers, etc, because that’s what most people would think when you put “Heat Transfer Equipment” on the program. Not here, though. “Heat Transfer Equipment” refers to the reheaters, superheaters, evaporators, economiz­ers, and other parts of the HRSG.

The first session is the catalyst for this meeting, and while it’s scheduled only for an hour and a half, discussion on the HRSG proper usually remains lively into the afternoon. It’s really tough to keep Chairman Bob Ander­son, principal, Competitive Power Resources, on the official schedule; he lives and breathes boilers, as do most of the attendees.

Even Rob Swanekamp, who man­ages the HRSG User’s Group on a day-to-day basis, waving wildly from the back of the room “to move on” has no impact. Anderson, microphone in hand, is moving faster than a coy­ote across the front of the room and up and down the aisles asking and answering questions. No one is going to catch him.

But it’s not all Anderson all the time. He has a “team” of boiler experts in the audience to handle the tough questions. You don’t see this at first, but after a couple of meetings you realize that there’s no question asked that someone on Anderson’s boiler “swat” team can’t answer. These experts include:

  • Mike Pearson, principal, J Michael Pearson & Associates Co Ltd.
  • Joe Schroeder, senior VP of engi­neering, Nooter/Eriksen Inc.
  • Dr Barry Dooley, Structural Integ­rity Associates Inc.
  • John Briggs, GM steam genera­tors, Florida Power & Light Co.
  • Akber Pasha, technical director, Vogt Power International Inc.
  • Jim Witherow, executive chemist, Scientech LLC.
  • Bob Krowech, president, HRST Inc.
  • Jeff Henry, president, Energy Solutions.

It can be difficult to fire up an audience of 350 from a cold start and Anderson and Swanekamp leave nothing to chance. They spend con­siderable time on the phone with reg­istrants a couple of weeks before the meeting to learn about some of the problems their colleagues faced dur­ing the year and what sticking points may still be standing in the way of a solution. That information is used to develop questions to keep the discus­sion moving in high gear.

You might think that the first few questions would be easy ones, to make everyone in the room comfort­able. Not so. The first one was a hum­dinger from a user who was warned against using P23/T23. This prob­ably was out of left field for many attendees, who were still coming up to speed on P91/T91.

One speaker said the ASME task group for creep-strength-enhanced ferritic steels was investigating the interrelationship among material specs, heat treatment, and hardness as part of a Code case (No. 2199) and that this work was ongoing at the time of the HRSG meeting. Then another person, one close to this activity took the microphone and offered the following report:

A boiler OEM had been qualifying a new vendor for Grade 23 material and in the course of that effort had done some heat-treatment studies. The investigator believed he was not getting the hardness response expected based on the published con­tinuous cooling transformation dia­grams for the material.

First thought was that the dis­crepancy might have been caused by the vendor’s heat-treatment practice as was often the case in the early days of P91/T91 development. But follow-on studies showed it was not the heat-treatment procedure, but a result of the material’s response to the normal processing of the steel.

The participant continued, saying that to achieve the promised creep strength for this material, you must achieve a very specific condition of the microstructure. What the results of the OEM’s analysis showed is that it is possible to meet the chemical requirements of the material as stat­ed in the Code case without achieving that necessary microstructure.

Working closely with two sup­pliers of the Grade 23 material, the ASME task group modified the chemistry requirements to ensure that the alloy would have sufficient hardenability and imposed hardness requirements that make it necessary to demonstrate the proper harden­ability is there.

Now, for the part of the response that no one wanted to hear: If you have existing Grade 23 material that was not supplied by a major supplier (two mentioned were Vallourec & Mannesmann Tubes and Sumitomo Corp) before the Code case changes were implemented, review the mate­rial thoroughly. There is a high prob­ability that your material has a defi­cient chemistry and will not have the creep strength necessary to support the Code-allowable stresses.

How’s that for a real-world eye-opener first thing in the morning? Now you understand why some com­panies are loathe to buy anything without a blue-ribbon pedigree and years of problem-free performance.

Attendee adrenaline was pump­ing by this time and Anderson did not have to ask everyone to stand up for deep-breathing exercises to get the attention he believes boilers deserve. He can be a drill sergeant at times.

The discussion continued with this advice: If you have deficient heats of Grade 23 material, assume that you have creep strengths comparable to Grade 22 (worst-case scenario). A question on where P23/T23 would be used in an HRSG. Answer: super­heaters and reheaters. What’s the upper bound of the metal tempera­ture? Possibly up to 1100F responded a boiler OEM’s representative, but the scaling temperature (the temper­ature at which a metal begins to form iron oxides) is only 1025F.

An HRSG manufacturer volun­teered that it believes there are too many unknowns with Grade 23 and his company has decided not to use it at this time.

Then came the obvious question, “What’s so great about Grade 23, anyway?” A member of the Anderson “swat” team replied that the attrac­tiveness for an OEM is that, unlike Grade 91, you are not necessarily required to do post-weld heat treat­ment with this material.

With Swanekamp waving and mumbling “23 skidoo already,” it was time to move on. Next topic: solid-particle erosion (SPE) of steam-turbine blades at a five-year-old 3 × 1 combined cycle that cycles regularly. Typically, two GTs/HRSGs are used, so the steamer rarely operates at full load. The third train can be out of service for a month or so, then rein­stated and another train shut down. User thought that the source of the problem might be exfoliation of tube material on startup and was looking to the group for advice.

Another user reached for the “mike” and talked about a similar experience. Unit operated base-load for more than 10 years, coming down for maintenance two or three times annually. No problems in the steam turbine. Then the plant was rel­egated to cycling service. Steamer was opened for inspection and plant personnel found “the IP section was just disappearing” from SPE. A big literature search ensued to help iden­tify the root cause.

One thing this user and his col­leagues learned was that while Grade 91 is prone to forming oxides, the par­ticles are not as thick as those from exfoliating T22 and cause less damage. To keep metal particles out of the tur­bine, a fine-mesh screen was installed at the steamer inlet; two months later it was almost plugged solid.

New software tracks life remaining in critical HRSG parts; field tests verify expected results

Most US combined cycles are unregulated assets and oper­ate in challenging merchant power markets. Competition is keen and a plant’s pedigree must be char­acterized by high efficiency, availabil­ity, and starting reliability to meet pro forma expectations.

In this type of environment, you can’t manage a generating station by the seat of your pants and expect to have any measure of success. You need top-notch personnel and sophis­ticated analytical tools to make your numbers.

At the center of the combined-cycle universe is the gas turbine (GT). No surprise there. The steam turbine is respected for its reliability, having more than a half century of operational experience with the same working fluid at comparable pressures and temperatures. The heat-recovery steam generator (HRSG) gets the least respect and sometimes is thoughtlessly referred to as “the dumb hunk of steel” between the gas and steam turbines.

GT operation is carefully monitored onsite and at remote M&D (monitor­ing and diagnostic) centers, data are recorded and trended, the remaining lifetime of critical hot-gas-path com­ponents (in particular) is assessed online using the latest software, etc. Attention befitting a beauty queen.

Steam turbines are far more robust than the GTs. If water/steam chemistry and steam quality are held within spec­ified limits, lube/control oil is main­tained in good condition, and proper warmup, ramp, and shutdown proce­dures are followed, the unit should be able to operate for its design lifetime with a minimum of issues.

What about the HRSG? Although the health of rotating equipment is actively monitored, and the lifetime remaining in critical parts continually updated, little comparable intelligence is available for the HRSG. Stack tem­perature traditionally has been used as an indicator of deterioration and fouling of heat-transfer surfaces, steam temperature as an indicator of proper spray-valve operation, makeup flow as an indicator of tube leakage, water analysis as an indicator of waterside metal health, etc. These are qualitative assessments, not quantitative.

Inspections of the HRSG on the gas side, and on the water side to the extent possible, are conducted dur­ing plant outages and that information traditionally has been used to guide repairs and replacement of pressure parts, such as module harps.

Prior to the introduction of LCAMP™ (Lcamp, for Life Consump­tion Assessment and Monitoring Program) by Vogt Power International Inc, Louisville, there was no commer­cially available monitoring tool with the ability to track the life remaining in critical boiler parts—at least not one known to the editors. The proprietary software package was developed over several years by a passionate engineering team headed by Technical Director Akber Pasha and supported by Dr Peter Sun and Lin Shian Tsai.

Lcamp imports and analyzes infor­mation gathered by a plant’s data acquisition system and calculates the estimated life consumption of the specified boiler component. Having an accurate estimate of end of life allows timely ordering of replacement parts and proper planning of the outage required to install the new components. Operating until a pressure part fails gen­erally is unacceptable for a merchant plant because of the cost and outage time incurred for repairs and extensive lead times for replacements.

The software also can be used for predicting crack initiation and for measuring the effects of various boiler operating modes on the life­times of critical pressure parts. The latter enables better operational deci­sions to keep maintenance expenses within budget. Lcamp also suggests when and where to conduct detailed inspections.

Real-world experience

Beta testing by The Southern Com­pany, which worked closely with Vogt during Lcamp’s development, verified that the program is relatively easy to install and use and provides meaning­ful operational intelligence.

The editors visited with Jimmy McCallum, manager of CCP and CT technical support for Southern Company Generation, and two of his senior engineers—Andre Prewitt and Dean Sheffield—in Birmingham last spring to learn more. Not present was McCallum’s engineering analyst, David Jenkins, who wrote the code to merge Lcamp and Southern’s OIS software (Operational Information System from AspenTech). Vogt engineers partici­pated in this activity as well.

Obvious from the discussion was that McCallum and the two engineers had intimate knowledge of the soft­ware and were comfortable using it. Input data for calculations comes from instrumentation typically installed on the HRSG (pressure and temperature), plus drum-metal thermocouples. Drum TCs must be installed if not supplied with the unit. At Southern, Lcamp extracts the data required from the DCS via the company-standard OIS, validates that information, and pre­pares the input for its use.

Important to note is that no sub­jective interpretation or statistical inferences are required to use Lcamp, neither is experience with stress analysis. Output can be tailored to the specific needs of any plant and displayed using standard commer­cial software—such as Excel. Lcamp works for any OEM’s HRSG.

Southern’s applications thus far have not run Lcamp in real time. Plant McDonough, which is scheduled for commercial operation in mid 2011, will be the company’s first real-time application.

The online version will have the ability to calculate the cost of a start­up—incorporating fuel price, the sale of power to the grid, and impact on the life cycle of critical components—as the unit is being brought online. This allows the operator to increase or decrease startup time based on cur­rent conditions. The present version of Lcamp calculates startup cost after the cycle is completed.

McCallum said analytical tools like Lcamp are particularly valuable to owner/operators today because fuel costs and electricity prices continually impact dispatch schedules. Citing the boiler as an example, he said that mar­ket conditions can cause the HRSG’s operating profile to deviate significantly from that assumed by designers.

Consequently, some boiler com­ponents may deteriorate faster than others and may require repair or replacement much sooner than antici­pated by the design team. A reduction in expected life caused by unforeseen operating conditions would not be bad, McCallum added, were it static and fixed in time at a certain stage. However, life-expectancy estimates change as operations progress and operating requirements change in response to market requirements.

A core value of Lcamp is that it assesses the impact of changing operating conditions on the lives of critical components and continuously updates the life consumed so plant personnel can schedule inspections and repairs or replacements before a failure occurs.

Information provided by the Vogt software program includes HP drum and superheater cycles consumed based on fatigue calculations, and HP superheater and reheater hours con­sumed based on creep calculations.

McCallum and his team ran through a series of “results” charts developed for one of its plants with two 2 × 1 F-class combined cycles that had been in service since 2000. Life consumption was estimated from 2000 through 2007, before LCamp was installed.

Life-consumption estimates compiled at the central engineering center on Inverness Center Parkway are shared with the plants on a quar­terly basis to avoid overwhelming station personnel with data. Reports generated include identification of “significant” events, recommenda­tions for changes in operating proce­dures to extend component lifetimes, and a suggested inspection plan for critical components. Interestingly, the data gathered also can be used as a shift “report card” and point to deficiencies that can be corrected through focused training.

To illustrate: A data set for the start of one unit indicated that if the same conditions and actions were repeated 162 times, the HP-drum downcomer nozzles would be at end of life—as estimated by Lcamp. If units were cycling upwards of 200 times annually, a number like this would get attention quickly and deficiencies would be addressed. Data for another start revealed ideal conditions and a projected lifetime of nearly 9000 starts.

Summary charts for management are easy to read at a glance. Detailed curves of temperatures and pressures over time also are produced to help operations managers and staff engi­neers assess the root cause of off-spec data, which might be something as simple as a defective instrument.

The summary graphics include the cumulative charts from 2000 through the current quarter for component life consumption (Fig A) and for cycles by type (Fig B). A dollar value for life con­sumed is displayed in another chart. Plus, the same data are presented on a series of charts for the current quarter.

The numbers in parentheses in the charts show the range in values across the four HRSGs serving the two com­bined cycles at this plant. The HRSG data presented in Fig A was the best of the four units. The reason is evident in Fig B, which profiles the same HRSG. It had the fewest starts, fewest cold starts, and the most hot starts among the site’s four boilers.

To illustrate how dramatically the dispatch schedule can change over time, the Southern engineers com­pared starts data for one unit during 3Q/2008 with data for 1Q/2009. There were more than 60 starts in first peri­od, only five during the second. Rea­son: Low gas prices made GT-based power less expensive than that from coal-fired units and the combined cycles were operating many more hours on a per-start basis.

Lessons learned

The Lcamp experience has helped The Southern Company improve some of its traditional design and O&M prac­tices. A few lessons learned:

  • Specify HP drums for future HRSGs with thermocouple wells to accurately and reliably measure drum center and skin metal tem­perature. Southern engineers were reluctant to drill existing drums to install TC wells and used other methods for TC attachment.
  • Lcamp suggested that the thick nozzle welds on the HP drum’s two downcomers would fail earlier than predicted because of cold-condensate impingement attrib­uted to daily weekday cycling. That prompted plants to remove insulation from downcomers prior to boiler inspections.

    Inspectors found cracks where Lcamp predicted they would be. Those identified soon after crack­ing started could be ground out and re-welded relatively easily.

  • One of Southern’s “Golden Rules” is to keep HP drum skin tempera­ture above 250F to protect against material damage. The company engineers now assure that new plants have a reliable backup source of steam to keep drums hot during unit shutdowns.

Now this user had a question for the group: Can we expect exfolia­tion to decrease to a point where it levels out? He said station person­nel considered doing a chemical clean on the T22 reheater bank, but determined this course of action could not be justified economically. Another bit of information: This user found that the scale liberates when it achieves a thickness of about 7 mils.

Question from the floor: Are you only having the problem in the reheater? No, said the user who had asked the original question; there’s also SPE in the HP section of the tur­bine but it’s not as serious as what we’re finding in the IP. This made sense to him because the reheater sees larger temperature gradients and flexes more than the superheat­er, facilitating exfoliation.

Another participant with deep experience in SPE and its origin contributed to the dialog saying that exfoliated scale from T22 was, in fact, the erodant of greatest concern. He offered a more comprehensive explanation of the scale formation and release process: The oxide grows as a duplex layer and won’t exfoliate until it breaks down into a multi-laminated structure, the formation of which is a time- and temperature-de­pendent process. It could take 30,000 hours for this to occur, depending on temperature.

Then, when the unit shuts down, exfoliation occurs because of the strain energy in the oxide. During ensuing restarts of the unit, steam picks up the exfoliated scale and car­ries it into the turbine. The expert confirmed that Grade 91 oxidizes and exfoliates faster than T22 but is less likely to cause damage because the particles have less mass.

Now, to answer the questions asked. “Absolutely, you can expect exfoliation to continue if you don’t do anything about it.” There is history to prove this, the expert continued. Regarding chemical cleaning of the superheater and/or reheater, it’s generally not eco­nomical. The cost can be astronomical. Two ways to prevent SPE:

  • Use erosion-resistant alloy blades on the first few turbine stages. The materials have been developed and proven effective. But they cost more than conventional materials. Owner/operators must anticipate problems, such as exfoliation, at the design stage and mitigate them in the specifications.
  • Redesign the steam flow path. The leading turbine manufacturers know how to do this, the attendees were assured.

The exchange of ideas never seems to dead-end at an HRSG User’s Group conference. Another participant offered what appeared to him a more practical approach. If the material is removed by steam during startup, he surmised, you could blow it into the condenser through the turbine bypass systems for a short period to protect turbine.

The previous contributor agreed, saying that this is the practice at many conventional steam plants in Europe, which generally are equipped with bypass systems. They have much lower erosion rates than plants in North America.

Confirmation that bypassing steam to the condenser during start­up is a viable strategy for minimiz­ing SPE also was offered by experi­ence from Asia. Two large coal-fired plants in the same country—one with European-style bypass systems, one without. Former had insignificant SPE, the latter serious problems. Reminds you of the adage: You can pay me now or you can pay me later.

Unfortunately, the financial peo­ple who make the final decisions on what features a plant will and won’t have, were not in the audience to relearn that “an ounce of prevention is worth a pound of cure.” But it prob­ably wouldn’t have made any differ­ence if they were.

Life-cycle cost also impacted answers to the next question on forced cooling of the HRSG. A user asked: What procedures are recom­mended for safe-cooling of the HRSG serving an F-class GT to permit quick internal access for inspection while protecting personnel.

A boiler manufacturer’s rep was quick to reply. “I think your HRSG manufacturer should have studied that issue and given you approved cooling rate because it impacts the OEM’s life-cycle study.” The rep con­tinued, “Be aware that the old ways of cooling HRSGs (such as by opening the doors and forced-cooling them) should never be attempted, because that method will harm at least part of the superheater and reheater sec­tions.” The bottom line: There is no standard rate of cooldown that works for all HRSGs.

Another user offered an opinion based on three decades of experience. Rapid cooling does a great deal of fatigue damage to the unit and it seri­ously compromises the boiler’s abil­ity to meet lifetime expectations. He believes forced cooling simply should not be attempted. Readers are referred again to the sidebar on p 146. The soft­ware described there could provide a reasonable estimate of how much unit life is lost by forced cooling.

Drain system design was the next discussion point, which went on for quite some time. If you are having difficulty draining condensate in time­ly fashion, all that might be required is a relatively simple change in piping size, layout, and/or valve type and/or arrangement. This subject is cov­ered well in the quintessential refer­ence for HRSG owner/operators, the “HRSG Users Handbook,” compiled by Swanekamp and available through the group’s website at

Economizer cracking. Just when you think you have seen all there is to see because you’ve been designing, operating, and/or maintaining HRSGs for three decades or so, something “new” pops up on the radar screen. One consultant reported that he and his colleagues have encountered a new problem (new to them at least)—cracking in the cold end of economiz­ers within the finned area. This is not a pandemic by any means, but it is being found “here and there.”

The cracks are circumferential and “fairly tight.” Sometimes there’s a small amount of leakage associ­ated with the condition which causes stripes on the tubes. Some people refer to this as “raccoon tails” because of striping on the fins. The mech­anism is stress corrosion cracking (SCC); nitrates are the corrodant. The question: Is anyone else seeing this?

A user took the microphone and said, “We have seen this problem in one of our units. In fact, we were in an outage, water was still in the tubes and you could see 30 or 40 plac­es where there were leaks occurring.” The leaks were never big enough to impact water inventory, nor so problematic to require a shut down.

Stack dampers: The value proposition

The benefits of stack dampers for gas-tur­bine-based cogen­eration and combined-cycle plants—including faster startups, reduced corrosion on the gas-side of heat-re­covery steam generators, etc—are discussed at vir­tually every meeting of the HRSG User’s Group. How­ever, until this year’s conference, most information transfer on the subject offered relatively little first-hand experience and hard data.

A joint presentation by Tampa Elec­tric Co’s Yogesh Patel and Bremco Inc’s Bill Kitterman on retrofit of stack dampers at Teco’s Bayside Power Station contributed to the industry’s collective knowledge by quantifying the benefits and describing the work involved to assure project success.

Patel’s portion of the presentation was supported Paul Lofton, a key player on the plant’s maintenance team; Kitterman’s by Phil Shand, Bremco’s project manager on the job, and Mike Wheeler, the project super­intendent. Patel is a member of Teco’s headquarters engineering staff, Kitter­man is Bremco’s VP/general manager.

Bayside began life as the coal-fired F J Gannon Station, Patel said. It was repowered in 2003 and 2004 and now consists of seven F-class gas turbine/HRSG trains (Fig A) and the two original steam turbines. Total out­put is a nominal 1800 MW. The plant was designed as a base-load facility, but during the latter stages of design it became apparent that the facility would have to cycle.

The run-up in the price of natural gas during the repowered plant’s early operating period normally required operators to remove from service daily five of the seven GTs for between six to 12 hours. Natural convection and the “stack effect” caused a strong draft through the offline GT/HRSG trains, quickly cooling the equipment.

Result: More fuel was required on the ensuing start to achieve combined-cycle opera­tion. Teco engineers decided that installation of stack damp­ers was the best way to reduce the rate of cooling and the associated fuel consumption. Calculations revealed that stack dampers would save the plant nearly $500,000 annually based on gas at $8/million Btu and a nominal 190 annual starts per unit. Engineers figured that a hot start would be reduced by three minutes, saving $344 in fuel.

Patel said other benefits of stack dampers include the following:

  • Less NOx (tons per year basis) is produced because DLN (dry, low NOx) operation is achieved more quickly.
  • Thermal cycles are less severe, reducing maintenance costs and extending the lifetimes of critical pressure parts.
  • Convective cooling of the GT/HRSG train is virtually eliminated during shutdown periods. This reduces gas-side corrosion caused by condensation of water vapor in the exhaust gas. Heat retention can be improved by insulating the stack below the damper and by painting damper blades black to minimize radiation losses. A ben­efit of reduced fouling of finned heat-transfer surfaces by corrosion products is higher efficiency.
  • Prevent the ingress of water dur­ing rainstorms, which can be sig­nificant in Florida—especially with Bayside’s 19-ft-diameter stacks.

Data from physical tests at the plant following installation of the stack dampers showed that the time it took the high-pressure (HP) drum to decay to 53 psig was 31:05, or 8:33 longer than the 22 hours 31 minutes without dampers. Also, prior to the damper retrofit, it took 21:38 to achieve DLN operation; now it takes 16:44.

Kitterman followed Patel at the podium to review the details of stack-damper design and installa­tion. He began by extending the value proposition offered by stack dampers. Kitterman, who lives in New Hamp­shire, noted that during long periods offline in northern climes, a damper contributes significantly to freeze protection provided space heaters are installed as well.

Next, he talked about stack geom­etry and emissions monitoring. Most new large-frame-powered plants have traditional round stacks and they were the focus of Kitterman’s remarks. Some installations have rectangular ones, he acknowledged, but Bremco has not yet installed dampers on them. He did not see why there would be any problems in outfitting rectan­gular stacks with dampers.

Kitterman cautioned that rules gov­erning the location of monitoring ports for continuous emissions monitoring systems (CEMS) may impact damper placement. Regulations differ among the states, he said, but ports typically must be located about two and a half stack diameters from any obstructions both above and below. At Bayside, the centerline of the damper was about 4 ft above the top of the stack breech­ing connection to the HRSG with the CEMS ports located above the damper.

Installation. Most of Kitterman’s two-dozen slides illustrated how Bremco installs stack dampers with a high degree of safety. Fig B shows that the Bayside stacks had plat­forms where the dampers would be installed. This is not always the case, he said. Installation of a stack door was one of the first steps, thereby allowing access to the work platform installed inside the stack along with the scaffolding (Fig C).

Holes also had to be cut in the stacks to accommodate the damper drive shaft (Fig D). Damper blades are assembled, bolted, and welded into position in Fig E. The vertical flanges are bolted and stitch-welded to join adjacent blades. Damper drive end is installed in Fig F, support bearing for the other end of the shaft is in Fig G.

How the damper operates. The Bremco/Bachmann Dampjoint Inc (Laval, Que, Canada) damper is opened prior to gas-turbine startup by an elec­tric actuator like that shown in Fig F. A sensor signals the control system when the damper is in the full-open position, allowing unit startup to proceed. View ports are provided for redundant visual verification of damper position.

When the GT is in operation, damper blades are held in the full-open position by the high-ratio (self-locking) mechanical gearbox. After the GT is shut down, the actuator closes the damper with an assist from blade weight and gravity which force the blades to follow the actuator’s closing motion. The pressure differen­tial across the closed damper is mini­mal, so leakage by the metal-to-metal sealing surface is minimal.

The self-locking gearbox assures that the damper will remain wide open under all expected operating condi­tions. An added safeguard: The actua­tor is designed to fail open if power is lost. Also, damper design is such that a nominal flue-gas pressure of 10 in. H2O will open it in the unlikely event the GT starts when the damper is in the closed position—this to prevent a unit trip on high backpressure.

A calibration system is provided with the damper to assure its open­ing and closing at given flue-gas pressures without in-service testing. Weights can be added to increase the gravity force at the full-open position without changing the desired gravity closing force. Calibration should be verified annually when the linkage and shaft-packing tightness are checked.

Kitterman said owner/operators usually want to know if any stack structural changes will be required to support damper installation. Bremco’s experience is that a belly-band might be required, as it was at Bayside, but that’s about it. Calculations can be made to support decision-making. This portion of the presentation closed with a review of information required in support of a proposal.

The user said he and his colleagues believed that the cracking was initi­ated externally and caused by a con­centration of acids at the cold portion of the tube bundle.

All HRSGs so affected did not have SCRs, so their exhaust streams were relatively high in NOx. One opinion offered: If there’s a slight amount of sulfur in the fuel, it elevates the dewpoint; once you have a wet con­dition the NOx is absorbed into the condensate thereby creating an SCC mechanism between the nitrates and the carbon steel. ccj