Attention to detail early in the HRSG lifecycle sets the tone for long-term reliable service

The first five years of HRSG operation are critical for building the foundation necessary to assure a reliable and satisfactory service life, said engineers from HRST Inc participating in an editorial roundtable with CCJ staff.

In the early years, they continued, control and organization of supplier documentation will benefit current and future decision-making. Plus, timely inspections and analysis can help ward off potentially damaging conditions or identify items to allow contract closure with suppliers.

As time goes on, analysis also can determine the effects of any current or impending changes to operations that were not planned when original specifications were prepared. Finally, there are commissioning and early operating practices and conditions that can harm the HRSG. Many examples are in evidence industry-wide, the experts said. Several are identified below.

First year

Focus plant personnel in Year One on gathering, organizing, and protecting documentation. Proper and efficient management of maintenance and operation require comprehensive and complete documentation, Team HRST said.

It’s important to confirm that all suppliers have submitted complete design packages. Developing a reconciled list of documents with the latest revisions and dates will prove invaluable as time goes on and personnel change. Obtaining copies of the ASME Manufacturer’s Data Report forms can verify component design pressures and materials of construction, and aid in future repairs.

ASME B31.1 requires every facility to have an inspection program for covered piping systems (CPS). This primarily constitutes pipe 4 in. diam and larger between the HRSG and steam turbine or user. The program calls for a comprehensive list of documents. B31.1, Article 141, “Operation and Maintenance,” provides guidance.

Confirm that the documents are complete, current, and well protected. Guard against catastrophic loss and assure ready availability by keeping complete sets of documents in multiple accessible locations.

Inspection, testing, analysis. Your pre-first-fire inspection should focus on design details, use of correct materials, and proper installation, and be supported with ample photography.

Inspectors should go through the gas-side access and crawl spaces, paying close attention to the inlet duct and firing duct. Violent exhaust flow is known to take down components in the inlet duct if they are not installed properly. Note interferences from restrained thermal expansion of tube bundles and duct liner systems.

Inspect each drum, paying close attention to the final (secondary) steam separator that covers the steam outlet to the superheater. The HRST engineers said it limits carryover of water droplets and impurities that can quench or plate out on downstream tube surfaces. Excessive carryover can lead to loss of superheat and elevated risk of tube damage from high temperatures. In severe cases, carryover may quench the tubes, causing fatigue damage at the header joints.

All pipe and pipe supports external to the HRSG must be in the correct position and any hydro-stops removed before fired operation. This is a good time to document the hot and cold marks, before they fade or fall off. If low points in steam pipe cannot be drained, consult the OEM or contractor.

First two years

Near the end of Year One, or just before expiration of warranty periods, a comprehensive inspection and test program is highly recommended by the HRST experts. It will provide baselines for future inspections. Photographic documentation is especially valuable.

Within the gas path, inspectors should look for signs of restrained expansion, interferences, and wear. Some cracks and twists can be expected and are often self-limiting. Others are signs of imminent or future trouble. Improper material selection or the occasional installation of the wrong material will often be revealed as a difference in oxide color or scale formation. If damage is found, this is the time to address the mechanisms conducive to future component failures.

Examples of inspection findings by Team HRST include the following:

Stressed tubes. Warped tubes are not necessarily a liability, but they are signs that the tube has been stressed and has yielded. Additional fatigue may cause failure. Cracking might be apparent in tube joints. Stress is common in the superheaters, reheaters, and economizers. Failures have been known to occur shortly after commissioning.

External tube wear. Tubes are susceptible to oscillation from forced vibration attributed to exhaust flow. A tube-tie support lattice typically is installed at multiple points along the height of the tube bank to limit this oscillation. If the tube ties are spaced too far apart or if they are loose, tube vibration causes fin wear, and eventually wear on the tube wall.

Overheated materials are those that have been subjected to a sufficiently high temperature for a length of time to cause changes in microstructure, or excessive scale formation. Perhaps the materials are mismatched to their location in the HRSG. Overheating typically is confined to the inlet duct and firing duct. Inspectors may find one liner sheet, gas baffle, or tube that looks out of place. If not corrected, early failures can be expected.

Duct burners. Look for signs of flame impingement on the walls or downstream tubes. Long flames or excessive localized heat can be caused by failed burner nozzles, inadequate fuel distribution, inadequate exhaust distribution, or improper controls.

Online inspections. There are several ways to identify areas of concern with the HRSG in operation, including these:

Infrared imaging inspections of the ducts—in particular the roof, doors, and pipe seals of the first three modules (HP evaporator and forward to the gas turbine)—can point to excessive casing temperatures where issues are likely to exist.

Monitor operation of the steam separators to verify their proper operation. Do this by testing for the purity of steam leaving the drums to ensure downstream equipment is protected against carryover.

Examine flames through the burner viewports to verify flame lengths are not excessive.

Review data in the plant historian to identify excessive cycling of valves, valve leakage, excessive desuperheater spray water, and other harmful conditions.

Walk-down high-energy piping systems to locate unwanted movement/displacement.

Second through fifth years

Every year or two after COD (commercial operating date), HRST engineers recommend a standard 10- to 20-hr inspection to look for wear issues throughout the HRSG. As conditions develop, they can be identified and the underlying cause corrected before large-expense outlays are required. After four years’ time, issues that tend to be time-dependent should get more focused attention—including corrosion, expansion-joint failures, pipe seal failures, and fatigue cracking of pressure parts and liner systems.

Analysis for vulnerabilities and changes. Flow-accelerated corrosion (FAC) may be the most detrimental and expensive of tube and pipe failures encountered by HRSG owner/operators. It is highly dependent on water chemistry. Damage from FAC can develop rapidly, with visual wear occurring in less than one year.

HRST engineers told the editors that the HRSG should be evaluated for FAC risk before the third year of operation. A follow-on FAC inspection plan can then begin. However, your mitigation program should begin earlier if there are telltale signs on FAC in the LP or IP drum in Year One.

While HRSGs are now constructed using FAC-resistant materials in high-risk areas; in those cases, the scope or frequency of inspection can be reduced—perhaps—but never neglected.

Operational changes. Several years after COD, the owner may need an operating profile different from that specified for the original design. This could involve a change in cycling frequency, operation at a lower load, etc. A thorough engineering evaluation of the new operating parameters on the HRSG is highly recommended by HRST engineers. Here are some things to think about as you go through this process:

Cycling study. Many specs require that the HRSG be designed for a certain number of starts per year. Each start contributes to wear and tear. A cycling study will look at those components vulnerable to thermal transients, high temperatures, and low flow rates that often accompany startup and shutdown conditions.

Bear in mind that cycles affect the allowable HP steam-drum ramp rate, so if startup time must be reduced or is in question, the cycling study will reveal the balance between cycle frequency and ramp rate.

Low-load operation. Operation of the gas turbine at low loads affects steam and water flowrates. This can lead to unstable flows in economizers and poor flow distribution in superheaters and reheaters, all affecting tube reliability. High exhaust temperatures from some turbines require excessive desuperheater spray that damages downstream pipes and tubes

Gas-turbine modifications primarily affect the HRSG through a change in exhaust mass flow and temperature. This changes the ratio of HP/IP/LP steam generation as well as overall heat recovery. Many components must be examined for suitable operation. Modifications with high exhaust temperatures may result in excessive desuperheater spray demand as well as superheater and reheater tube metal temperatures above the design temperature.

Avoid ongoing operations that damage the HRSG, including the following:

The desuperheater is responsible for a significant share of damage to superheaters, reheaters, and steam pipes. In some cases, very few fatigue events are needed to initiate tube failures and cracks in the steam piping. There are two underlying causes:

First, there may not be enough energy in the steam to evaporate the spray water. This causes restrained thermal expansion and potentially high stresses in components quenched by the water droplets. The steam temperature downstream of the spray nozzle should be 50 deg F or more above the saturation temperature.

Second, damage to pipes and tubes often occurs when water flows through the spray nozzle without proper atomization. This is caused by a leaking block valve during periods of no spray water demand or from a broken spray nozzle. Water then can flow upstream or downstream through the steam pipe.

A drain pot with automatic detection and drainage is required on all new HRSGs with a desuperheater. In such cases it is important to verify that the drain pot will collect any leakage (sometimes it does not) and is functioning properly.

Burner operation. Excessive heat can damage downstream tubes and duct-wall liner systems. Burner failures result in high localized heat input while faulty control logic can cause a wall of flame pushing through the tube banks.

Burner system controls must be checked as well. Burner output should be controlled to the limitations of the HRSG, not to the burner design maximum. This is especially important when multiple HRSGs have burners controlled to provide a minimum steam generation rate in a common header. If one unit goes offline or if some of the burner elements are isolated, how is the additional heat distributed among remaining units? It is important to verify that controls distribute heat as intended. Burner management system permissives and fuel-skid set points must prevent excessive fuel flow to any one HRSG.

Piping and header drain control. It is critical to verify that low points in the steam pipe and tube circuits are properly drained and that there is no water leakage through steam conditioning systems. If water hammer occurs, it is important to inspect the damage and the position of the pipe. Water hammer could create a low point that did not exist previously. The reheat system can cause extremely expensive pressure-part damage if not maintained and operated properly (including sizing, location, and control).

After shutdown of the HRSG, steam will condense and flow to the lower headers of the HP superheater and the reheater. If this water is not removed prior to startup, steam can push water droplets up tubes or water can restrict steam flow. These conditions cause high stress resulting in tube failures.

HP-drum pressure ramp rate. HP steam drums in cycling service are susceptible to fatigue cracking of nozzles, particularly the downcomers. Exceeding the maximum ramp rate will increase the risk of such cracking.

Posted in 501 F&G Users Group | Comments Off on Attention to detail early in the HRSG lifecycle sets the tone for long-term reliable service

501F exhaust-system repairs never ending? Replace to eliminate the problem

Exhaust systems are a presentation and/or discussion topic at most gas-turbine user-group meetings, particularly those supporting F-class and the more advanced frames. Cracking of components, broken struts, material distress, and leaks at joints are almost sure to occur in virtually any system handling a 1000 lb/sec or more of turbine exhaust at 1000F or higher and moving at near mach speed—especially so in starts-based engines with original equipment.

There was a period a few years back when the accepted “solution” was to have personnel standing by to make repairs during annual—or semiannual—outages. In meeting after meeting, discussion focused on weld material, weld configuration, welder qualifications, expansion joints, and related topics. It came to be that users able to restrict repairs to every other outage were recognized for their acumen by peers.

The effects of hot gasses leaking from cracks into areas where they could cause accelerated wear and tear and/or personnel hazards provided owner/operators the incentive to press for root cause analyses and resolve the problems through better design and materials selection. There was a lot of finger-pointing to be sure, mostly directed at the gas-turbine OEMs. However, that might not have been entirely fair given the limited industry experience available to guide the design of early F-class units.

The OEMs of record—Siemens Energy Inc and Mitsubishi Hitachi Power Systems Americas Inc—upgraded the designs of their exhaust cylinders and manifolds for the 501F series of engines to eliminate the issues experienced. These have been retrofitted at several plants, as discussed in some CCJ articles published over the last several years. They can be accessed by using the search function on the magazine’s home page at

More recently, Ansaldo Energia Group’s PSM, headquartered in Jupiter, Fla, has expanded its 501F aftermarket product line to include exhaust systems. The first complete PSM exhaust systems were retrofitted in NV Energy’s Silverhawk Generating Station (Unit B) and Walter M Higgins Generating Station (Unit 2) prior to the 2017 summer run.

Both plants, which began commercial operation in 2004, are 2 × 1 combined cycles powered by Siemens 501FD2 gas turbines. The utility plans to replace the exhaust systems on Silverhawk Unit A and on Higgins Unit 1 during their next major outages—provided upcoming inspections/evaluations confirm expected performance.

Background. The 501F exhaust system has two principal components: an exhaust cylinder, connected to the turbine case on its inlet side and to the exhaust manifold on its outlet side (drawing). The cylinder may be of single- or two-piece construction. PSM’s “drop-in” replacement is a horizontally split two-piece cylinder, the only option available for pre-501FD3 machines when they were purchased. This configuration typically is preferred by users because they can maintain their current maintenance practices in future outages.

Depending on component dimensions, a spacer piece may be required downstream of the exhaust manifold to fit up with the heat-recovery steam generator. An expansion joint is located at the round-to-square transition between the manifold or spacer and the HRSG. In the case of a simple-cycle unit, there would be no HRSG and the manifold or spacer would connect via an expansion joint to the transition piece directing exhaust gas to the stack.

Scott Amos, a gas-turbine subject matter expert in NV Energy’s central engineering group, and Fatima Bouzidi, maintenance manager at Silverhawk, walked the editors through the 501F exhaust-system challenges their company faced—including the following:

    • Failure of static seals, allowing hot exhaust to contact bearing support struts, causing overheating and cracking which allowed the aft bearing and rotor to “drop.” The rotor drop reduced critical compressor and turbine radial clearances to the point where metal-to-metal contact between the casing and rotating components became a concern. Experience of other users is that a crack can propagate quickly, extend through the entire strut, and force the unit into an outage with extensive damage.

    • Thus timely strut repair and clearance correction are recommended to prevent the possibility of a wreck. Repairs and realignment are expensive undertakings, to be sure. Remember, too, that access to the aft bearing for inspection and the taking of measurements on legacy equipment is not a simple matter. In NV Energy’s experience, it takes about a day and a half to cool an exhaust system with compromised seals before you can drop down through the manway and crawl forward to the bearing.

    • No end to the cracking/repair cycle experienced with the exhaust cylinder. This was due, in part, to failure of baffle plates, allowing recirculation of exhaust gas to the dead-air space. PSM’s “fix” prevents recirculation and the casing creep issues attributed to it.

Exhaust gas also can overheat expansion joints, allowing them to leak and create safety and environmental issues. Repairs are time-consuming and expensive, and never “final.”

NV Energy’s engineering team had been following industry experience on 501F exhaust systems for years, while continually evaluating the condition of its Silverhawk and Higgins units, before deciding on a course of action for its assets. Amos said after receiving multiple alerts from the OEM and listening to owner/operator colleagues at forums such as the 501F Users Group (next meeting Feb 17-22, 2019 in Paradise Valley, Ariz), it was obvious that repairs are temporary and the only long-term fix is replacement of the exhaust cylinder.

Bearing drops on the four 501Fs at Higgins and Silverhawk were monitored by PSM. The relatively constant drop in rotor centerline position of Higgins 2 over time averaged about 10 mils annually. Silverhawk B experienced a more severe drop: about 15 mils per year on average.

With the future easy to predict, NV Energy engineers conducted detailed design reviews of replacement exhaust systems offered by the two OEMs and PSM. The company focused on reliability improvements and reduced maintenance promised by the offerings. Its goals were the following:

    • Eliminate exhaust-bearing drop.

    • Improve the low-cycle fatigue life of the cylinder and its struts.

    • Improve system durability in all modes of operation, with a focus on cycling.

    • Reduce the thermal constraint on the strut shield.

    • Improve the exhaust-cylinder/HRSG connection and protect the HRSG inlet expansion joint.

    • Provide better access to areas that may require repairs.

    • Assure compatibility with the existing structure, piping, and other connections.

Major design changes by PSM (from the original exhaust system) to achieve the goals bulleted above were these:

    • Change exhaust-bearing support struts from Type-410 stainless steel to a nickel-based super alloy.

    • Redesign exhaust cylinder and manifold support structure, and improve aerodynamics.

    • Eliminate baffle seals by integrating a new manifold front flange into all PSM exhaust installations.

    • Reduce thermal stress by making changes to material geometry.

    • Change diffuser and strut shield to Type-347 stainless steel from Hastelloy®.

    • Redesign exhaust-cylinder and -manifold gas-path seals.

Posted in 501 F&G Users Group | Comments Off on 501F exhaust-system repairs never ending? Replace to eliminate the problem

Factor European experience into O&M practices at US combined cycles, Part I

European Technology Development Ltd’s (ETD Consulting), International Conference on Power Plant Operation and Flexibility, held in London in July 2018, covered recent developments in plant materials, operation, inspection, maintenance, and costs related to both baseload and cycling operation for different types of plants—including combined cycles. Participants included plant owners/operators, consultants, researchers, manufacturers, and inspection-agency professionals.

Conference organizer Dr Ahmed Shibli, managing director of ETD Consulting, pointed to the following global trends:

    • Europe is going through the same market changes as North America. Cycling is global.

    • Some countries have more cycling and low-load experience than North America—such as those with geothermal and hydroelectric plants that cause varied operation of fuel-fired assets. This experience base has wide-ranging value.

    • Europe and Japan are taking active roles in high-temperature materials development and testing.

Shibli’s primary purpose in arranging the meeting was to help identify the basic causes of equipment problems in both baseload and cycling units, and to clarify modifications, inspections, and procedures that will minimize costs.

Who wins when plant components fail? A question asked by one of the first speakers, David Allen of Impact PowerTech Ltd (UK), “Who wins when plant components fail?” set the tone for the conference and was the perfect segue for his presentation on the upgrading of materials and welds. Allen’s next question was equally pointed: “Why do we so often stick with last-century technology?”

The speaker began with a philosophical look at planned obsolescence. He cited interesting examples from the automobile, white goods, and communications industries. Participants were captivated, and on alert for more.

Allen did not claim current evidence of planned obsolescence within the power industry, and in fact stated that a manufacturer with a unique design is careful to maintain its good reputation.

But he did note some dangers similar to those expressed by Eskom (South Africa) at last year’s Air-Cooled Condenser Users Group meeting in Las Vegas. To wit, we all experience the ongoing pressures for low bids, rapid investment payback, and increasingly stringent trading conditions.

The potential menace is commonly applied standards that fail to ensure the most reliable long-term component operation. In other words, once a supplier meets the standards or codes in the specification, further refinements or improvements can become less important (and more expensive) to the supplier. The same is true on qualifying bids for component supply. Providers’ reputations are at less risk as long as they meet the qualifications.

Perhaps because “innovation brings risk,” or because “regular repair and maintenance activities provide jobs,” we could be letting ourselves down, suggested Allen. And in his words, “We are still building new plants with 40-year-old materials (P91, Alloy 617) and ignoring the newer, potentially better alternatives.”

And how has market-driven flexible operation changed things? “Cycling makes it worse!”

Details have become clear, stressed Allen:

    • Flexible operation only makes service exposure more onerous.

    • Thermomechanical transients cause additional cyclic loading, which is conducive to the following: 1. Mechanical and thermal fatigue cracking, 2. Creep-fatigue cracking (with creep ductility exhaustion attributed to repeated transient creep strain), and 3. Creep cracking (with creep life reduction caused by increased loading).

    • Poor temperature control can severely shorten creep life.

    • Creep issues do not go away; they get much worse.

    • Fatigue brings additional challenges.

    • Sticking with last-century technology increases risk!

So how do owners/operators minimize the potential costs? Allen outlined some thoughts:

  1. 1. Design out “at-risk features” (thickness and materials mismatches, closely spaced header stubs, dissimilar-metal welds, etc).

  1. 2. Improve temperature control during startups and transients.

  2. 3. Make components thinner—perhaps.

    • Thinner components minimize thermal/mechanical mismatch loading and may therefore perform better when fatigue is a problem.

    • But thinner components will experience higher pressure stresses and may therefore perform worse when creep is a problem.

Allen reviewed typical owner/operator options. Some choose to continue operation until plant end-of-life, mindful primarily of safety. Many take the financial hit and accept high inspection, repair, and replacement costs. The better strategy is to install upgraded retrofit components with stronger materials and welds at the same thickness (better for creep) or thinner (better for fatigue).

He then presented details and examples of current P92 materials with good-quality heat treatment, labeling this “a materials upgrade solution that is ready now.” Following in-depth discussions of various manufacturing methods, heats, ductility, and tensile strength, he asked participants to “Stop fearing P92!” 

And he gave specifics from a recent review, showing a “strong correlation between heat treatment and ductility.” He concluded that “normalizing is about twice as important as tempering. Under-normalizing is the main problem. Under-tempering makes the problem worse.” And “notably, normalizing time is more important than temperature.”

    • The four most creep-brittle casts, with long-term average area reduction (Ra) in creep test values in the range of 3% to 12%, all had normalizing times in the range of 0.2 to 0.6 hours.

    • The seven next most creep-brittle casts, all with Ra values in the range of 18% to 24%, had normalizing times of 0.2 to 1.0 hours.

    • No cast with a normalizing equivalent to 2 hours at 1050C or 1 hour at 1070C had a long-term Ra value lower than 28%.

Summary: This indicates that simple controls on allowable heat treatment can resolve the Grade 92 creep brittleness concern. Further work is ongoing to assess very long-term ductility out to 100,000 hours, and beyond.

Allen next looked at “near-future MarBN, a novel high-alloy steel for powerplants.” This material, being studied primarily in Japan and the UK, is martensite plus boron and nitrogen. The process involves “careful microalloying with boron and limited nitrogen for high creep strength.” Allen listed temperature capability of P92 as about 20 deg C better than P91 and “expects temperature capability of MarBN to be at least 25 deg C better than P92.”

But the microalloying process is extremely sensitive.

Following discussions on welding processes, he offered the following conclusions:

  1. 1. “Today we can replace P91 with P92.

  2. 2. Tomorrow we could use MarBN for even greater security.”

Posted in 501 F&G Users Group | Comments Off on Factor European experience into O&M practices at US combined cycles, Part I

7HA users wrestle with emergent issues at inaugural meeting

A CCJ ONsite Special Report

Welcome the 7HA Users Group into the pantheon of power-industry organizations seeking peer-to-peer interaction to solve problems expeditiously and provide a collective check on official OEM positions regarding fleet- and plant-level issues.

Representatives from well over a dozen facilities and five countries, responsible for 30+ machines, attended the inaugural 7HA Users Conference, in Fort Worth, Tex, Sept 12-13, 2018. The table illustrates machine operating stats as reported by the participating owner/operators. Obviously, it is early days for the technology with respect to commercial operations.

The scene was reminiscent of the early 1990s when the F-class technology emerged in commercial settings, dozens of units were sold worldwide, engines began operating before any appreciable operating experience had been gained with the fleet leaders, and before long, units were being air-lifted from around the world to have serious deficiencies addressed.

You can read about that experience in “Flaw in Design of Turbines Results in Massive Recall from Utilities All Over,” by William M Carley of the Wall Street Journal, May 6, 1996.

We don’t preface this article this way to create ill will among users and OEMs, but to remind the community that the evolution of high-energy, highly engineered, cutting-edge power systems is rarely a smooth process. To get to today’s landscape of F-class machines humming around the world doing what is expected of them, for example, the industry had to get through a tumultuous early to mid-1990s period which deeply affected all five major large-frame OEMs at the time. Now there are three.

What follows is an exclusive industry report from the 7HA Users meeting. All technical issues are described generically out of respect for the extreme sensitivity owner/operators face managing the OEM relationship. GE was there in full force, though, with the meeting structured as first day devoted exclusively to owner/operators, and the second day to the OEM response, followed by a plant tour. CCJ was invited to attend only the part of the OEM session that had nothing to do with addressing issues.

The material, to the extent possible, is organized in priority of fleet issues and/or issues affecting multiple sites. Then a punch list is provided—issues experienced by one or two sites. Learning through open and transparent discussions among the folks who live and die with the equipment is the fastest route to commercial success with the technology.

Keep in mind that the 7H machine has established new industry records for power output, efficiency, emissions, and turndown flexibility. These achievements were reported in an earlier article. 

First-stage bucket failure

As if on cue to accentuate the importance of such conferences, the first speaker noted that one of its machines had, just days earlier, experienced a first-stage turbine (S1B) bucket failure after less than 10,000 operating hours. All four units (at two sites) being operated by this company were then shut down as a precaution.

Days after the conference, Reuters reported on September 20 that four GE HA-class machines in the US were shut down because of an “oxidation issue” and the company expected that all 51 machines shipped to date would be affected. The article also noted the fix, according to an OEM spokesperson, would require “minor adjustments.” Users at the meeting said GE had “acknowledged that this is a fleet-wide issue.

Others have reported that the oxidation affects a key alloy in the blade and shortens its life. GE, already dealing with existential corporate and Wall Street level complications, was dealt yet another blow by this event.

The bucket issue also has thrown into some turmoil the delivery schedule for machines to US customers, according many present at the conference. One, with a commercial operating date (COD) a few years out, said that the buckets destined for their machine, what were to be the first “Gen II” S1Bs, were now being diverted as replacements in the failed units. This action has postponed COD.

Users wondered aloud whether this “fix” would be the right one, since there is no commercial operating experience with Gen II hardware. The next opportunity to “look at them” in the first machine incorporating them, said the site representative, was not expected until a scheduled outage many months out.

Another operating site rep said the OEM was 9-10 months behind on spare parts and that they were “desperate for spare stage 1 blades,” and needed to “get through the fall on someone else’s blades.” A third site was expecting its S1B replacements in 2019. A fourth noted that they were getting a mix of Gen I and Gen II hardware. Finally, a site with several thousand hours on its HA.01 machines reported that the dampening pins for its first-stage buckets already had been replaced twice.

AFS failure

Some HA combustors feature axial fuel staging (AFS), of which a key component failed and caused damage at one site. The GT was operating at 298 MW when the failure occurred. It not only caused internal damage to the combustion system, but also breached a combustor can into the turbine compartment, triggering the fire protection system and tripping the unit.

There are 48 of these components on each machine with AFS. At the site experiencing the failure, 40 of these components were replaced. At least two other plants have had AFS components replaced.

As later elaborated on by a site representative, the DLN 2.6+ combustion system on such units consists of a fuel flow path referred to as “axial fuel staging” (AFS).  Fuel enters the GT compartment and passes into a ring manifold and is delivered to each combustion chamber via “pig tails.” Fuel then is equally distributed to four ½-in. tubes attached on the outside of the “Unibody” (but internal to the combustor), and injected into the hot gas path through four nozzles.

The idea behind this flow path is to introduce fuel gas downstream of the flame zone and just upstream of the turbine first-stage nozzle. The fuel gas auto-ignites, increasing the energy available for the power turbine but with no increase in NOx emissions. This also allows for extended turndown while maintaining emissions compliance.

At the site with the AFS tube failure (during normal operation), a flashback occurred causing complete disintegration of two fuel nozzles, damage to the Unibody assembly within the combustor, and some splatter impinging on the turbine first-stage nozzles and buckets within the quadrant of the failed combustor.

There was also a breach of the combustor wall. Hot gases entered the turbine compartment and triggered the fire-protection-system heat detectors, which also automatically initiated protective action of immediate unit trip.  Data review revealed that approximately 45 minutes before unit trip there was an instant spike in NOx, exhaust spreads, and low-band combustion dynamics.

The root cause of the AFS tube failure is unknown at this time pending completion of the analysis effort by the OEM.


One site with around 3700 operating hours on its HA.02 machines reported that they had been plagued with cold-start transient vibration and start-to-start vibration issues associated with bearing No. 3, which the OEM had recently acknowledged were fleet-wide issues. The presenter said that half a dozen other machines were experiencing similar problems.

The vibration levels increase over time, and are associated with thermal changes in the unit. Low-load operation makes it worse. The vibration is supposed to be below 5 mils, but at times has been indicated as high as 12 mils, and routinely is around 7. Perhaps the most aggravating thing is there is no real pattern to the deviations, other than they tend to be “stable” at baseload.

Numerous foundation checks by the OEM didn’t solve the problem, nor did attempting to correct the balance with strategically placed weights (“shots”). In fact, “placing the shots in the wrong places led to numerous fruitless iterations.” An hypothesis of coupling misalignment also has not proven out.

Excessive vibration leads to secondary issues, such as oil leakage at the deflector plate from one of the generator rotor bearings, terminal strips coming loose, and failures of exhaust-thermocouple attachments. “We are replacing many exhaust thermocouples,” one said. Another site experiencing similar vibration problems is testing a prototype thermocouple that may be more robust.

One site reported that they had changed out the seals on the leaking generator bearing, added a row of labyrinths, and made some other adjustments which appear to have corrected their problem.

Complex controls

Several attendees commented on the complexity of the H-class control system and the shortage of OEM control-system engineers familiar with the H machines. “Controls are really complicated,” one said, “whenever we do a logic change, GE doesn’t have enough controls engineers to support us.” Another worried that “the digital intelligence is not 100% developed” for these systems. A site experiencing only one trip blamed it on a fieldbus error in the control system and subsequently requested that the GT controls be hard-wired.

A third presenter complained that the original GT/G controls had no redundant vibration configuration and cautioned attendees to “be aware of the primary-frequency response logic for performance tests.” This site also had incorrect digital valve position (DVP) firmware settings resulting from miscommunication between the OEM and the sub-vendor. The presenter urged his colleagues to “be familiar with foundation Fieldbus and Profibus for controls as applicable to the Mark VIe GT control platform.

Punch-list items

Those in the queue for their machines, and those considering ordering an H-class unit, will want to consider the following punch list of items discussed at the meeting. These issues have not risen to fleet level issues, or at least not yet, and some may be peculiar to one or two sites, placed in the general category of “teething” issues common to all new facilities, and may have shared responsibility with the EPC contractor.

    • Digital valve positioners. The same site referenced above with about 3000 operating hours reported failures of digital valve positioners, an “instantaneous event” that leads to a GT trip, while another site reported that temperature-control issues with the DVP led to three trips.

    • Water intrusion. Fan/blowers in the GT housing reduce heat levels but also create a differential pressure which sucks water in when it rains. A site in an arid climate didn’t experience this until the first time the area got a good soaking. These blowers are redundant and both need to be in service because they are critical to maintaining an internal temperature so that electrical and digital electronics don’t “burn up.” One plant lost a bearing in the motor of one of the fans and had to replace the entire motor.

    • Air-filter cleaning system. This system has 400 small valves. One plant experiencing leaking in a few of these valves decided to replace all of them. The communication-system bus which “talks” to all the individual valves failed.

    • Online washing. Several plants do this everyday and had to make modifications to “make sure the pump logic was properly integrated.” One user cautioned that the drains are sized for “zero margin,” and had to replace a few bad valves and fix loose flanges.

    • Water carryover. More than one plant has experienced water carryover from the evap cooler into the gas turbine.

    • Generator. One plant had a flashover event with the collector brushes. The original number of brushes was based on the nameplate rating, but the OEM recommended that they be reduced from 24 to 12. One user cautions that generator end windings have “required lots of money and work.” The OEM calls it “normal wear and tear,” according to one user, and that you have to be careful negotiating with the OEM because they will accuse you of “improper maintenance.”

    • ST/G bearing leak. A site with a single-shaft combined cycle reported that the No. 1 bearing in the low-pressure steam turbine/generator had sprung a “major leak” and the plant had to “limp through the summer.”

    • Miscellaneous heaters. The same site just above also reported “lots of workarounds” with the startup natural-gas auxiliary heat exchanger, the natural-gas performance heat exchanger, and the air inlet heating system (upstream of the inlet air filters), as well as “lots of issues with the air-cooled condenser.”

    • Cap effusion plate. A site overseas reported minor cracks were detected in the combustor-cap effusion plate following a borescope inspection after around 8500 operating hours.

    • Igniters. One plant experienced packing leakage and overheated igniter wiring on two machines, discovered during the first inspection. The presenter noted that an igniter had liberated at another site he was familiar with.

    • Lift-oil pumps. One site experienced failures with the pumps, found a breaker off of one lift-oil pump and another that wouldn’t start, and noted that “lube-oil skids are tightly designed, very compact.”

Posted in Miscellaneous | Comments Off on 7HA users wrestle with emergent issues at inaugural meeting

Latest GTs pose special challenges for NOx, CO catalyst system design

The state-of-the-art gas turbines (GTs), such as the H-class and J-class machines, are designed to maximize fuel-to-electricity efficiency, achieving and even exceeding 60% in combined-cycle mode. Already, there are an impressive number of these machines in the field; one supplier listed 80+ H-class GTs operating, in commissioning, being installed, or on order.

This achievement is accomplished through increased firing temperatures, upgraded combustion staging, and advanced metallurgy. The tradeoff to higher temperature, of course, is increased thermal-NOx formation. Dan Ott, president, Environex Inc told the editors that the burden of turbine-exit NOx emissions ends up on the SCR system, at least to achieve the same stack emissions as specified for F-class sites.

Additionally, he said, these machines are being promoted for even faster starts, more frequent cycling, and operation at loads down to 20% compared to the 50%-to-baseload range historically required of earlier designs. Lower-load operation coupled with higher exhaust NOx levels present a demanding set of design challenges for the post-combustion NOx and CO catalyst systems.

Currently, gas-fired turbines often are required to achieve 2- to 2.5-ppmvdc stack NOx, 2- to 5-ppm ammonia slip, and CO limits of 1 to 6 ppm. These limits became a de facto standard, at least in the US, after they were demonstrated based on F-class and aeroderivative engines with typical turbine-exit NOx between 9 and 20 ppmvdc. 

The new H-class machines exhibit 25 to 30 ppmvdc NOx emissions (Fig 1) with some excursions above 35 ppmvdc. This is by no means a trivial increase, Ott stressed. The SCR system designs are going from 75% to 85% NOx removal for F-class units to designs of 92% to 94% removal efficiency for the new advanced turbines. Higher conversion rates mean more catalyst, more frequent catalyst replacement, far more elaborate ammonia injection grids, and more frequent tuning.

Equally important to the SCR design is the impact from the ammonia-slip limit. A Frame 7F turbine with dry-low-NOx technology and a turbine-exit NOx level of 9 ppmvdc can rather easily meet a 2-ppm NOx/2-ppm ammonia-slip limit with 78% NOx removal and allowance for 22% excess ammonia (2-ppm stack/9-ppm inlet = 22%). When the turbine-exit NOx increases to 30 ppm, a 2-ppm NOx/2-ppm ammonia-slip limit requires 93% NOx removal with only 7% excess ammonia.

The success of the SCR design is critically dependent on the amount of excess ammonia that can be injected. Systems should be designed for greater excess ammonia at higher NOx removal requirements, Ott said, but the regulations currently do not allow this.

Fig 2 compares designs with different NOx-removal and excess-ammonia allowances, taking into account the complexity and operability of the designs. Note that most of the newest, higher-efficiency designs are in the “highest risk” zone where the risk of failure on commissioning is high and maintenance (catalyst replacement, catalyst cleaning, and ammonia tuning) is frequent and costly.

All turbines, including the newest designs, are now required operate at low and variable loads to respond to dynamic grid demands. This is in due in part to the growing impact of renewable power. Output down to 20% load is not uncommon.

Figs 3 and 4 show real-time operating data for NOx and CO emissions versus load for an advanced-class turbine. Most low-NOx/low-CO combustor technology is designed to function within guaranteed limits for NOx and CO above 50% load. Below that, NOx and CO increase rapidly because of suboptimal fuel/air mixing. These spikes in emissions at low load require the SCR and CO catalyst systems to achieve emissions reductions as high as 98%, far exceeding the capability of most designs. 

Ott’s conclusions: “We have reached the limit of current SCR and CO catalyst technology. We must increase awareness of these issues and look for broader solutions, including modified hardware and regulatory relief, to allow these new turbines to perform as required in an increasingly dynamic and unpredictable energy market.”

Posted in Miscellaneous | Comments Off on Latest GTs pose special challenges for NOx, CO catalyst system design

Meeting Announcement: 2019 Alstom Owners Group

Posted in Miscellaneous | Comments Off on Meeting Announcement: 2019 Alstom Owners Group