AUSTRALASIAN HRSG USERS GROUP

A cleaning expert assured the questioner it was not a problem to spread the bundle and blow downward with high-pressure air to achieve the expected level of cleanliness—assuming the debris is primarily corrosion products.

Users with long-term experience may remember that the first pre-engineered combined cycles offered by GE Energy—called STAG units—incorporated HRSGs of its design and manufacture. These heat-recovery boilers had horizontal tube bundles and sootblowers (many of the early gas turbines were equipped for dual-fuel service or to burn oil only). Those still in use today typically are found in combined cycles powered by early Frame 7 engines.

Next came a general two-part question: Do attendees experience deposit formation on HRSG tubes and, if so, do they pose any operational or maintenance problems? What methods do users employ to remove the deposits, and at what frequency? First user to respond said his HRSGs are experiencing an increase in deposition and the deposits are high in sulfur. Plant personnel increased the temperature of water to the air preheater section from about 130F to 140F by use of a recirculation loop, but that hasn’t had much positive impact. The questioner asked if it was best to look for other ways to minimize deposition or to just clean at some trigger point.

Another user said sulfur-rich deposits also were accumulating at the back end of his HRSG, increasing the pressure drop through the unit. That plant dry-ice blasts periodically to clean heat-transfer surfaces.

A respected consultant advised a performance-based solution. Increasing the feedwater temperature, he said, adversely impacts overall thermal efficiency. A rule of thumb is that you reduce output by 0.0184 MW/deg F rise in stack temperature. His recommendation: run to failure/clean rather than suffer the performance drop. Fouling is worst, he continued, with units having SCRs, particularly those overspraying ammonia. While that’s true, it was a moot point for most in the room because HRSGs operating in Australia and New Zealand are not equipped with SCRs.

Another consultant suggested analysis of deposits by x-ray diffraction/x-ray fluorescence to accurately determine their composition and to track changes over time. He has identified deposits of ammonium sulfate and elemental sulfur around access doors where rain water/salt spray have leaked in; also has found ammonium sulfate on the last few rows of tubes and zinc sulfate deposits where zinc has come out of the casing paint/coatings. A user seconded this consultant’s findings, having experienced the same at his plant.

A “mail-in” question from a UK user had to do with an odd header/tube arrangement on a double-wide HRSG with an alignment-plate configuration that didn’t fulfill its intended function. Result: Wing tubes from the upper and lower headers flared out, leaving a gap at the bottom and top of the unit that allowed hot gas to bypass heat-transfer surfaces from the front to the rear of the boiler. The user said a series of short, flimsy gas baffles were really only a token gesture at filling these gaps, and that they were already starting to break up from high-cycle fatigue and relative header movement. This owner/operator’s concern was the possibility of economizer steaming and increased possibility for flow-accelerated corrosion (FAC) in the evaporator because of a higher steam fraction in wing tubes.

A representative of the OEM in the room responded with a non-answer. A couple of consultants couldn’t pass up the opportunity to talk, but they didn’t say much of value either. Another user said he had a heat-recovery unit (not a HRSG) with a similar issue. In his case, during hot operation the tubes should have expanded to “fill in the gap” but didn’t and some tubes suffered overheating creep failure.

Next, a heat-transfer expert said the “issue with gas bypass is major and can lead to major thermal transients. You have to be on top of this.” One of the consultants who had responded earlier grabbed the handheld mic after thinking more on the subject and added, “These types of issues can be really bad for forced-circulation boilers, leading to dry out and dry-out related tube failures.”

Chemistry questions

One user, in particular, had several questions related to cycle chemistry. The first had to do with the capabilities of commercial laboratories in Australia and New Zealand to analyze anions and cations down to single-digit parts-per-billion levels. He got a good news/bad news answer from his colleagues: No labs could at present, but hopefully one or more will be able to do so soon. A chemist added that bench top UV-Vis spectrophotometer methods are no good for iron; you must use methods based on atomic absorption flame spectrometer or inductively coupled plasma to actually measure total iron.

Second question: “We found FAC in our LP system, particularly in the LP drum risers. Before the outage we had carryover issues from our LP drum, so only ammonia was being dosed. The carryover problem has been rectified and tri-sodium phosphate (TSP) is being dosed. The inspector suggested the FAC was caused by steam velocities, whereas I suspect it was caused by the ammonia residing in the steam rather than in the liquid phase. What do you think?”

One of the world’s leading water chemists replied this way: “IP and LP risers are the No. 1 locations for FAC in HRSGs—all two-phase—and oxidizing chemistry is not able to deal with it. Raise water-phase pH using solid alkali; then change the risers to P11.” In two-phase locations, he continued, magnetite solubility peaks at about 350F, so knowing the temperature of the fluid in the risers is critical. A pH of at least 9.6 is necessary; carryover must be prevented to avoid steam contamination.

The chemist stressed drum performance as critical. A common mistake in FAC management, he added, is to try to fix single- and two-phase FAC issues at the same time. Deal with each form of FAC separately. The questioner replied that his plant just changed out drum internals with carryover being reduced from 5% to 0.02% and had implemented TSP treatment, so significant improvement in the FAC issue was expected.

The water chemist jumped back into the dialog, suggesting that the user be careful of links to velocity. FAC is driven by turbulence, he said, which in this case is created by a torturous path. He also recommended that the user try sodium hydroxide in place of TSP, saying it was more commonly used now.

Third question by the same user: “Our unit ramps up and down daily and we have trouble meeting steam conditions as a result. Is there any way to reduce the spikes in cation conductivity (sometimes as much as 0.6 microSiemens/cm experienced when units are changing load regularly)?

Another user offered that his plant also has issues with steam quality being a cogen facility. Solution was to improve the HRSG chemistry to help lower carryover rates; the very high TDS levels that had been experienced were reduced. He suggested that the questioner regard short spikes as cumulative damage. The questioner remarked that his steam turbine was sandblasted following a major salt-ingress occurrence and the boiler was chemically cleaned at the same time.

An experienced chemist took the mic. I get asked about “damage” all the time. Before answering, he said, I want to know the conductivity after cation exchange (CACE) by way of ion chromatography (IC). If chloride and sulfate are found, you should be concerned at the HP evaporator and the phase transition zone in the steam turbine.

Sample an HP evaporator tube to determine its condition, he suggested. If you find significant deposits then the chloride and sulfate levels are a major concern because they can concentrate up and cause under-deposit corrosion issues. In the steam turbine, crystalline deposition is a major concern because of its impact on blades. When the unit is shut down, the deposits absorb moisture and cause pitting, which, in turn, leads to stress corrosion cracking (SCC). Offline turbine protection should be implemented.

Holding onto the mic with a vise-like grip, the same user asked a fourth question: “When you return to service after a long outage during which the Rankine cycle was drained, and where work required the boiler to be open to atmosphere for a significant period, what’s the best way to clean the system to ensure steam-quality specifications are achieved as soon as possible during run-up?”

About a dozen attendees responded to this question. Most replies were short—for example, good procedures and planning are vital to success, get a suitable sample early in the startup, have a chemist in attendance to support the operations team (not suggested by a chemist, but fully supported by all the attending chemist consultants), etc. A chemist polled the group, “How many plants represented have chemists?” By show of hands, four times as many plants with chemists than without. That’s in sharp contrast to US experience, where many generating companies have one or two chemists to service an entire fleet.

A user said a general procedure that worked well at his plant was to complete all work and put the plant in a dry shutdown condition with desiccant, dry air, etc. Then fill the boiler with high-quality water and arrange for extra demin to allow heavy blowdown during the ensuing start. Result: Perfect steam quality in less than 24 hours.

The question was answered best, perhaps, by a chemist in the room who offered the following procedure right off the top of his head:

1. Make sure chemical dosing system is properly lined up, tested, and ready to operate; verify that chemicals are fresh.

2. Fill the condensate system and circulate with chemical dosing on for about an hour. Then dump the hotwell and fill/dose/circulate again to purge outage debris. Repeat until the system is clean visually, and chemically via lab testing.

3. Fill HRSG and dump; flush with ammoniated water if plant configuration allows this.

4. Refill the HRSG to startup condition with elevated ammonia concentration and standard phosphate/caustic dosing levels. Have ammonia about 10% to 15% higher than normal for a cold start because you’ll lose ammonia via the steam vents faster than you normally would.

5. Get all chemical analyzers running on demin water and have spare sets of sample-line filters easily accessible. The filters installed will block as you bring the plant into service and you’ll want to change them out quickly.

6. Upon firing the gas turbines, go to “full-open” on continuous blowdown.

7. Blow down sample lines hard as soon as there is sufficient pressure to clear the lines. Rack filters likely will block up; be ready to change them a couple of times.

8. Bypass steam to the condenser until steam purity targets for the turbine are achieved. You never really know how long this will take. Don’t rush the process and approve the turbine start before acceptable steam purity is achieved. Steamers may look indestructible, but they’re very sensitive to what they ingest.

9. Watch the evaporators for drops in pH. Be prepared to manually dose with phosphate/caustic to keep pH where required.

10. Maintain high blowdown until the system settles down and purrs. This may take a few days.

Notebook filled, the user with the four meaningful questions above yielded the mic to the representative from a cogen plant serving the dairy industry. He had a very practical question: Are there any other owner/operators of cogen plants serving the dairy industry present and, if so, do you have the potential for organic contamination of condensate returns by milk products? A few hands shot up. Have you had any success in detecting condensate contamination online—such as by using TOC (total organic carbon) analyzers?

One person in the room acknowledged milk ingress into the HRSG, which left a white deposit on internal surfaces of the water/steam circuit. Another said his plant had issues with contaminated caustics (organics suspected) and tested for TOC. Two other users mentioned investigating instrumentation to monitor condensate returns. Both had TOC on their lists; one also was considering monitoring turbidity, sodium, and conductivity. Not much there to help the questioner except some sympathy.

A representative of a water-treatment services provider said his company had experience in this area and noted that virtually all cogen plants serving dairies were susceptible to condensate-return contamination by milk products. To minimize the impact of contamination, his company has an indirect TOC project underway that promises a very fast response time via an optical-based process. Commercial release is planned by summer’s end. You’ll undoubtedly hear more about the early operating results of this instrument if you attend the upcoming meeting in Brisbane.

Case histories

Morning tea over, the group was treated to short case histories/experiences by several participants which generated considerable follow-on discussion. Regarding “tea,” Chairman Dooley promises top-quality coffee will be available to all Americans (and others) who cannot adapt to local customs before the 2012 meeting begins.

Thermal fatigue of reheater tubes. The facts: A reheater tube in a 12-yr-old HRSG failed during shutdown; steam poured out the stack a few hours after the gas turbine was removed from service. Inspection revealed weld cracking at upper tube-to-header joints in Reheater 1 and nowhere else. Lab work fingered thermal fatigue as the failure mechanism. Lab also found defects/stress-relief cracking in tubes from the time of manufacture.

Root cause analysis (RCA) considered the following:

• Thermal load.
• Possibility of condensate moving through the system at startup and shutdown, thereby creating a severe thermal transient.
• Malfunction of the bypass control valve—sticking, for example. Reheater 2 was equipped with thermocouples that indicated large thermal fluctuations in those bundles. Although there were no t/cs in Reheater 1, thermal fluctuations were assumed given the Reheater 2 data.

Conclusion drawn was that thermal fluctuations combined with pre-existing cracks led to fatigue. The presenter’s question to the group: Why only failures at tube connections to the top header? No cracks were found at the tube/lower header joint about 30 ft below.

An attendee noted from the slides shown that failures occurred in the middle of the header rather than at the ends. Why? The speaker guessed that the bottom header drain location might have influenced which tubes were affected and the thermal loads on individual tubes.

“I often see signs of condensate entering the bottom reheater header,” an experienced consultant offered. Sometimes the problem can be traced to the bypass, sometimes it’s poor drainage. Another possibility is rotation of the top header, which can generate the stresses that lead to failure. I see from the slides,” the consultant continued, “that this is an upflow design; therefore, the drains were much more likely to be a major contributing factor.”

The representative from a HRSG supplier said his company was seeing much interest lately by customers in ordering harps with t/cs installed for monitoring purposes. He added that as many as one-fifth of the new units would be so equipped. “An interesting change in the market,” he concluded.

Spikes in conductivity after cation exchange. LP steam CACE spiked with 50-MW load changes during commissioning. Sodium went from 0.5 ppb during normal operation to as high as 50-100 ppb during spikes. CACE and sodium also spiked in condensate. LP drum had four “T” risers into the drum. Carryover testing according to procedures developed by the International Association for the Properties of Water and Steam (IAPWS) revealed these results: HP drum, 0.09%, IP drum, 0.02%, LP drum, 1.76%.

Physical inspections confirmed carryover from the LP drum, with phosphate deposition. A warranty claim was submitted and accepted; “T” risers were modified and carryover dropped to 0.01%. Conclusion: Purely mechanical carryover; no contribution from cycle chemistry or drum level. Comments/observations from participants included the following:

• Chemist. Carryover tests are very important. Testing program should include a run with a raised drum level as well. This is particularly important for the IP drum based on worldwide experience.
• User. Did you consider carryover damage to the LP turbine? Presenter said it was considered and a proper inspection would be conducted during the next outage.
• Boiler consultant. Also need to inspect the superheater; don’t just stop at the turbine. This comment was countered by a chemist who suggested looking at the reheater, not the superheater because of solubility impact.
• User. How many others have had this experience? Is it common for the OEM of record? A chemist said carryover was not specific to the OEM; the speaker reminded that carryover in this case only was a problem during significant changes in load. An attendee representing the OEM said not everyone was systematically measuring carryover.

Transition-duct redesign. Nominal ½-in. liner with studs applied using a stud gun. Damage noted in March 2000: Sections of liner plate had peeled away; backing angle and stud failure relatively common. Significant insulation loss caused overheating of the casing. Had to shut down twice annually for three days each time to make repairs.

First major repair, in 2001 was to rebuild the eastern wall of the duct in-kind. Over the next several years the original western wall failed in sections and partial rebuilds were done as necessary. Liner failures continued on the eastern wall in areas of greatest turbulence. In 2009 the decision was made to redesign and rebuild the eastern and western duct walls during the 2010 major outage.

The new design featured smaller sheet sizes, use of four nominal 1/8-in. overlapping sheets instead of one nominal ½-in. sheet, sheets fixed at one point only, more studs per unit of area. The old liner system was removed. Inspection a year later revealed no issues; plan is to inspect every six months during regular maintenance outages. Next up is a floor replacement, likely in fall 2013 during a hot-gas-path inspection. Roof replacement is planned for 2016.

A user wanted to know if anyone else had similar issues with Type-409 stainless steel. The designer of the new liner system said performance of Types 309 and 409 normally is acceptable. However, he added, pin/stud location is critical to prevent hang-ups and ripping caused by restricted expansion. Other problems include vibration and associated fatigue. Corners can be an issue as well; must get them correct.

HRSG cleaning. High backpressure was in evidence right after a three-month major outage in 2010, a user told the group. Output was restricted to 350 MW, a drop of nearly 50 MW compared to the plant’s capability before the outage. Important to note that during the outage, no steps were taken to protect the HRSG against material wastage even in an area characterized by high humidity.

Visual inspection revealed tube panels fouled with acid and ammonium salts that had accumulated during operation, plus corrosion products caused by poor offline storage conditions. The stack damper also was found out of alignment. It was repaired and produced an immediate improvement in backpressure. Baffles also were removed from module 5, reducing the delta P a bit more.

CO2 cleaning was pursued using the latest techniques to assure maximum effectiveness—such as the use of tube spreaders for “deep” cleaning. Only about two-thirds of the tubes could be reached. The owner was concerned about CO2 in a confined space and provided extra ventilation air by way of air compressors. There were no issues.

Outage took 12 days and photo records were compiled for all modules cleaned. Pre- and post-cleaning photos show good improvements and some minor fin loss. About 7 tons of material was removed and the unit is now capable of achieving 380 MW. Lessons learned included (1) access is very important (an extra access door was installed in the SCR ductwork) and (2) take precautions during HRSG layup to prevent metal wastage. Regarding the latter, when the HRSG is placed in “storage” today, steam sparging is used to raise metal temperatures and prevent moisture formation.

A question was asked on how steam sparging is done at this plant. The owner said steam connections are already there. They are coupled to the auxiliary boiler and steam is injected into the downcomers with a nitrogen over-blanket.

Cycle chemistry trifecta. Optimum cycle chemistry to deal with FAC and under-deposit corrosion, plus phase-transition-zone failures in steam turbines, was a practical presentation by a noted chemist of value to every owner/operator in attendance. He said system chemistry must be designed to (1) address all possible HRSG tube damage/failure mechanisms, (2) minimize corrosion-product transport, and (3) protect the steam turbine against sulfate and chloride deposits in the LP section and sodium hydroxide deposition on IP turbine blades. The presentation also addressed the fundamental level of instrumentation required to achieve these goals and the key elements of a management program to prevent bad cycle-chemistry situations.

A three-pronged approach is needed to mitigate FAC, the speaker said:

• All-volatile (AVT(O)) or oxygenated (OT) treatment to control single-phase FAC.
• Elevated pH (9.8 recommended) to control two-phase FAC.
• Monitoring of total iron to ensure the “Rule of 2 and 5”—less than 2 ppb total iron in condensate/feedwater and less than 5 ppb in each steam drum.

Single-phase FAC is “switched off’ by the oxidizing environment, but you must assure sufficient “oxidizing power” attendees were told. Best way to determine oxidizing power is to monitor (1) the level of oxygen at the condensate-pump discharge and boiler-feed pump, (2) the color of LP and IP drums for the “ruggedness of their redness (R of R),” and (3) low levels of iron transport.

Two-phase FAC is controlled by the pH of the water phase. Once single-phase FAC is under control, review iron levels. If elevated, then check wall thicknesses, increase pH with ammonia towards 9.8, and add sodium hydroxide or tri-sodium phosphate in accordance with IAWPS guidelines to ensure the pH of the water phase is correct.

Under-deposit corrosion (UDC) presents the greatest risk to HP evaporators because they have the highest heat fluxes. UDC and FAC are linked: FAC corrosion products increase the risk of UDC in the higher pressure parts of the cycle. To control UDC, you need low levels of iron and low levels of deposition in the HP evaporator. Plus, you must sample tubes from the HP evaporator (chemically clean as required) and install the proper instrumentation to control the entry of contaminants into the condensate/feedwater system.

The speaker said that based on his experience, most plants are under instrumented and not taking HP-tube samples, so they’re not controlling UDC risk very well. He added that poor chemistry selection contributes to heavy tube deposits in the HRSG and increases the risk of UDC failures.

A user wanted to know how much dissolved oxygen is needed to achieve “optimal chemistry.” The speaker said there was no one-size-fits-all number. It needs to be capable of holding feedwater iron at a low level while producing the R of R in the drums. If you are below 10 ppb in a combined-cycle plant, he added, then you’re unlikely to have the oxidizing power to provide the level of protection needed.

Is there an indication that you’ve gone too far, another user asked. The chemist said you really can’t have too much oxygen for the LP portion of the cycle, but you don’t want more than 10 ppb of dissolved oxygen in the HP evaporator tube bundles—so this is what has to be controlled. Next question: What’s the best location for HP evaporator tube sampling? Answer: In a horizontal HRSG, extract the sample from as high up the lead HP evaporator tube as possible. This is where you’ll find the worst deposition. If you can’t gain access to the worst tube, then better to have the next tube rather than no tube at all.

High-energy piping, P91 in particular, was the source of much discussion at the meeting with several users reporting such issues as incorrect filler material, welds with defects, difficulties in welding T91 to P22, hardness too high, hardness to soft, replacement of off-spec fittings and pipe sections, construction records not matching what’s installed, incorrect drawings, etc.

Order was brought to the subject of P91 by one of the world’s top consultants on advanced materials for powerplant applications who offered his thoughts on how owner/operators might organize and manage a high-energy piping (HEP) program for their plants. He began with goals. A HEP management program, he said, should have two major goals: personnel safety and high unit reliability. Despite such high-profile goals, it’s disconcerting, the expert continued, that HEP issues are on the increase because of limited maintenance budgets and the loss of technical expertise from the industry.

To reverse the trend, users need to know where to look for problems in HEP and headers; a random approach is not productive. An organized approach to ranking the probability of which HEP components will fail is needed. The ranking should be based upon each component’s potential for damage and incorporate results from inspections and stress analyses.

Once components are ranked in order of risk, inspections can begin. Goals for your inspection program include the following;

• Ability to detect the type of damage most likely to occur as early as possible.
• Provide rapid and repeatable results at the lowest possible cost.

The inspection process becomes substantially more complex, the speaker continued, when the material of construction is Grade 91 or one of the other creep-strength-enhanced ferritic (CSEF) steels. Controlling the microstructure of these unique steels dramatically increased the levels of technical and process control required during steel production and through all phases of implementation.

As the earlier discussion indicated, damage to piping system and other components made from these steels is widespread. Further, the location of damaged material is unpredictable, reflecting poor control of processing steps and poor record-keeping. With such poor quality control, the materials expert said, you cannot from a cost standpoint certify with complete confidence the condition of any component/system after installation. At this time, he continued, there exists no nondestructive inspection tool that can conclusively identify deficient CSEF material cost-effectively. CCJ