Owner/operators of 501G gas turbines (GTs) manufactured by Siemens Power Generation meet in person twice annually to share experiences with one another and with the OEM’s engineers. However, members of the relatively small group—there are only 20 SGT6-6000G (new engine designation) machines in service—are in contact daily via the organization’s interactive website and by phone.
Interestingly, the group’s attendance metrics are in sharp contrast to what you find at other GT user meetings. At most model-specific conferences, there’s one attendee for every three, four, or five engines in service; at the G there’s more than one attendee for each operating unit. For example, 34 users participated in the annual meeting in San Diego, January 29-February 1.
In 2007, for the first time, the 501G and F users held their annual meetings concurrently. The exposition was a joint event, as were all social functions. Plus, there were sessions with participants from both organizations because the subject matter was of common interest. Each group also conducted its traditional members-only sessions to review frame-specific subject matter. This included the Siemens Day interactive session with the OEM’s experts. The G users met with Siemens January 30, the F users the following day.
This article focuses on the 501G Siemens Day presentations; the 501F report elsewhere in this issue covers the non-OEM presentations of interest to both groups as well as the Siemens Day activities for F users.
Chairman of the G steering committee is New England-based Mark D Winne, plant manager, Millennium Power Partners LP; other members are Lakeland (Fla) Electric’s Timothy L Bachand, PE, engineering manager (energy supply), and Texans Charles Davis, director/plant manager, Ennis Tractebel Power Co LP (a unit of Suez Energy Generation NA Inc), and Jim Beckett, director of asset management, Calpine Corp.
Evidence of how important the G users are to Siemens was obvious from the meaningful participation in the meeting of its key service-organization executives. Rick Mould, executive VP, Operating Plant Service Div, opened the session and explained how the new regional arrangement of Siemens’ global power-generation service business was going to benefit the G fleet. Mould heads up US/ Canada region and reports directly to Craig Weeks, president, Global Operating Plant Service Div. His fluid, no-notes presentation expressed the company’s commitment to resolving outstanding issues with the frame as quickly as possible.
Ron Bauer, director of long-term programs, Operating Plant Service Div, and one of Mould’s top lieutenants, also was in attendance. He had been director of quality management in the old service organization and has intimate knowledge of Siemens’ efforts to resolve customer issues with the G machine and what was, and was not, accomplished. Perhaps more importantly, Bauer has worked collaboratively with the group for several years and knows first-hand what results the users expect.
First formal presentation, on fleet operating experience, was made by Matthew J LaGrow, the W501G service engineering frame owner. At the end of last year, he said, the fleet was approaching 300,000 equivalent base hours (EBH); the lead unit in hours had passed 28,000 and the starts leader had accumulated more than 1500.
In terms of operating profile, intermediate duty is most common (65% of the units); another 30% are in peaking service. Average fleet service factor typically is at about 30% in the beginning of the year, ramping up to near 80% at the summer peak, and then falling back to around 30% by year-end.
The meat of LaGrow’s first presentation concerned trends in reliability, availability, and starting reliability. Interestingly, the G fleet did not begin to accumulate significant operating experience until about 2003, the year that Siemens began gathering plant data for fleet performance analysis. The first G, installed by the City of Lakeland, fired-up in April 1999 but COD wasn’t until March 2001—one month before the second machine began commercial operation at Millennium. Details on fleet performance are available to G users on Siemens’ Customer Extranet Portal (CEP).
Technical issues update This portion of the program was handled by David B Grant, an engineer in the operations support group. He began by explaining the formal eightstep process Siemens has in place for resolving problems. It begins after a problem is identified and proceeds this way: A dedicated team is formed to define the problem, determine the impact, identify the root cause, develop and verify a solution, create an implementation plan and implement the solution, and monitor operational results until the solution is considered ready for full commercial release.
A new design for the inlet-manifold splitter plate and struts was presented first. Grant explained the improved splitter plate by way of a drawing overlay showing the cutback new design superimposed on the original.
Two types of cracks have been observed on some pipe-type struts— usually those in the first and last positions on either side of the manifold (the longest struts). One type of crack observed runs through the weld, the other across the strut. Flow-induced vibration during start transients was identified as the root cause; new hardware designed by the OEM to change the part’s natural frequency was recommended by Grant. He added that strut installation also has been changed to eliminate the weld zone where cracking has been observed on some units.
The new strut design has been implemented successfully on four units in the SGT6-5000F fleet. Siemens experts suggested that users verify the integrity of their struts and welds at combustor inspections or other opportunity. Visit the CEP for a copy of Grant’s slides and other supporting documentation.
Compressor diaphragm hookfit wear was an important part of Grant’s presentation. Damage caused by a seal holder that had migrated upstream and contacted the adjacent rotor disk was first reported in early 2003 on an F-class engine. Excessive wear on the diaphragm hook-fits and the seal holders was assessed to be the underlying cause. Later that year, similar wear (but no damage) also was visible on a G-class diaphragm.
By the end of 2004, a wear model had been developed for the F frame as part of the resolution effort aimed at identifying a means for keeping seal holders from contacting the rotor. A gap analysis indicated that the G compressor had twice the axial clearance of the F and offered no reason for immediate response.
However, in the fall 2006, when significant hours had started to accumulate on some G machines, inspections of several units revealed wear on the diaphragm hook-fits and the seal holders that had greatly exceeded the OEM’s expectation.
Grant described the sequence of events this way: Hook-fit wear allows the diaphragm to move forward at its inner diameter thereby reducing the clearance between the seal holder and the rotor disc. If the seal holder contacts the rotor disc, the friction produces heat that further distorts the seal box and could lead to hardening of the rotor disc material.
A dedicated resolution project team was established for the G frame and data were collected and analyzed. One of the first conclusions: The issue under investigation was hours-based, not starts-based.
Inspection methods were improved to more accurately measure the distance between the rotor and diaphragm (use of threaded wear pins, for example); inspection frequency was increased where considered necessary. More reliable and timely inspection data have enabled engineers to better estimate remaining service intervals. Plus, the fleet wear model was updated as inspection data became available to improve unit-specific wear estimates.
Near-term goal of the project team included analyzing, and if necessary, recommending actions targeting the 48,000-hr inspection interval. In some cases, this could require building- up hook fits (most often, in the first six rows of the compressor) with Type-410 stainless steel to tighten the fits in the casing. On rows where wear is excessive and operating hours relatively low, the recommendation generally is to replace diaphragms. Note that some casing wear also may occur at rows 1-6; if that happens, it also must be repaired.
A redesigned diaphragm for the G fleet should be ready for beta test in fall 2009. Figuring three years in beta test, it probably will be 2012 before the new diaphragms are available commercially—assuming, of course, all goals are achieved.
Starting reliability, said Grant switching topics, is impacted by the ignition system, combustion environment (fuel/air ratio), auxiliary equipment (starting motors, fuel-gas valves, etc), and other factors. He also acknowledged that plant operating personnel influence starting reliability through their O&M practices, and this is reflected in the fleet statistics available on the CEP.
Recent igniter improvements implemented by Siemens have had a positive impact on starting reliability. For example, a 25-kV, single-piece igniter with a mechanical pressure seal is recommended today to help avoid semiconductor degradation, which can adversely impact spark quality.
Component refinements now in the final stages of development are aimed at reducing igniter sticking in the guide tube to ensure timely retraction after ignition. Less exposure time at high temperatures may help extend igniter life and reliability. First operation in a commercial engine is expected in the fall. A diagnostic tool to verify sparking at the igniter tip also is in development.
Turbine blade-ring oxidation and cracking has been found on some units in Row 1 (primarily on the aft face), R2 (primarily on the upstream face), and/or R3 (primarily on the upstream face of the original rings made from 2¼-chrome steel). Grant said the degradation varied from unit to unit and row to row.
Primary cause of the degradation is hot gas ingestion. Here’s how that happens: Bow wake static pressure from the downstream vane pushes gas from the flow path into the blade ring cavity. CFD (computational fluid dynamics) modeling demonstrated the mechanism, metallurgy confirmed the exposure of blade rings to high temperature.
Mitigation targets a reduction in the gap between blade rings with additional features aimed at reducing the volume between blade rings to minimize hot gas circulation. One way to do this is to apply a hightemperature coating. Other enhancements include use of modified R2 vane locking hardware to reduce local volume and of modifications to the isolation-ring aft face to reduce the gap at the vane. Such enhancements are targeted for field testing in summer 2007.
Two other enhancements in development: (1) Install a positive seal between ring-segment isolation rings and downstream vanes. (2) Fabricate blade rings from alloys that offer increased corrosion resistance at higher temperatures.
The exhaust manifold retrofit to address a cracking issue on some units, continued Grant, had been installed on half the operating fleet before the end of last year. Cumulative experience at that time was about 25,000 hours, with the fleet leader compiling about a quarter of the total hours and 250 actual starts. No issues were noted and temperatures in the bearing tunnel were cooler than baseline numbers.
Work by Siemens engineers goes way beyond the manifold as a standalone component. The resolution team continues to invest resources in support of G-frame combined-cycle owners who have reported the following findings at the inlets to their heatrecovery steam generators (HRSGs): material distress on distribution-grid supports, weld failures on high-temperature boiler tubes, and inlet-duct vibration. Siemens has entered into a collaborative agreement with some G users and an HRSG supplier to study the exhaust conditions of the new manifold and the effect—if any—that it has on the boiler.
The two phases of this threephase program are complete. Phase I included characterization of exhaust flow, collection of data on exhaust sound power levels, conducting a traverse test for inlet duct flow, and monitoring key inlet data for operating HRSGs.
Phase II involved both CFD modeling by Siemens and physical flow modeling by the HRSG supplier; also Siemens provided the HRSG manufacturer exhaust characterization data for unit-specific evaluation purposes. The results of Phases I and II are under review by the collaborative team. Should the first two phases of the program suggest that modifications are required to the HRSG, Phase III will be initiated. It would include identifying sitespecific modifications/enhancements to the HRSG.
For owner/operators evaluating the feasibility and benefits of retrofitting new-design exhaust manifolds on its units, among other considerations, keep these suggestions in mind:
- Inspect the HRSG for mechanical condition prior to installation of the new manifold.
- Plan any necessary boiler work so that it doesn’t interfere with the GT work.
Corrosion and debris have been reported by some users. On the G, debris build-up behind R1 turbineblade seal pins may be a cause of one blade liberation incident. High-cycle fatigue was cited as the cause for both crack initiation and propagation. Also, accumulation of foreign matter in the root area can make turbine and compressor blades difficult to remove.
The buildup of very fine particles behind the platform seal pin decreases clearances and changes the blade frequency with respect to machine frequency. Testing for particulates in the cooling air at an F-class machine revealed that the majority of corrosion debris was released from exposed low-alloy-steel surfaces in the flow stream during the first 20 to 30 minutes of startup. Frequent starts and high vibration may, therefore, accelerate build-up.
A cut-back platform design has been released to the G fleet for implementation. The same design has been used successfully on a couple of F-class engines.
Tom Schuchart, manager of global generator service engineering, updated attendees on important issues impacting the three types of generators installed at sites with G units: the AeroPac and TLRI air-cooled, and the modular hydrogen-cooled machines. The G engines, themselves, have modular hydrogen inner-cooled generators; steam turbines in combined-cycle applications may have a modular gas-cooled, AeroPac, or TLRI generator. The report on the AeroPac was given on the F-frame program (G users invited) because that is a popular machine for those users.
Schuchart’s slide on the causes of generator outages was particularly interesting because two items were responsible for 90% of the G fleet’s total generator outage hours in 2006: generator vibration, 66%; rotor collector, 24%. Generator voltage regulator and generator casing were at 4% each; seal-oil supply system, 2%; and I&C, 0%.
The issue of cracking on pole crossover braze joints on some TLRI generators most often associated with V engines was discussed. Users were referred to technical advisories and other customer communications for details. Schuchart reported that a root cause analysis (RCA) points to low-cycle fatigue.
Response was to modify the crossover connector to eliminate the squarecorner braze and to incorporate a flexibility feature. A repair process developed to replace the pole crossover has been qualified in both the shop and in the field; replacement is targeted to take up to seven days, but can be done with the rotor in place.
Schuchart urged owner/operators to inspect under the retaining ring to see if cracking has occurred. This, too, can be done with the rotor in place using a technique developed by Siemens and should be included as a standard inspection procedure.
One of the fleet’s modular hydrogen- cooled generators suffered a stator ground fault in April 2006. That machine was replaced by an alternative supplier and, at the time of the meeting, Siemens was requesting access to the original unit to conduct a root-cause investigation. Schuchart urged users to inspect end windings and parallel rings on units at the next outage of sufficient duration as suggested by a Siemens customer advisory.
Matt LaGrow returned to the podium after lunch to address rotor belly bands, torque converters, and other components requiring special attention. Rotor belly bands, he began, are located between turbine discs and act as seals for the rotor cooling-air circuit. Significant wear has been found on some of the original belly bands during turbine outages; field repair involves machining a slot in the turbine disc and replacing the damaged material.
Design work was ongoing during the first half of this year to develop the next generation of belly-band segments, which would be more robust than the originals. Expectation was that a redesign would be ready this summer. LaGrow also discussed process improvements to facilitate field repairs and a way to detect bellyband failure by modifying disc-cavity flow-controller limits. He recommended inspecting belly bands every 24,000 EBH.
Torque converters have been getting significant “air time” at several user-group meetings. The focus of the short presentation on this subject was alignment. Inspection and overhaul frequency and requirements vary with the vendor. Instruction manuals referenced for this frame suggested an alignment inspection annually at a minimum, or after every 800 attempted starts or 3000 hours of turning-gear operation. A Siemens advisory provides this detail along with additional recommendations.
Users also were urged to check the outboard bearing for wear and to use those results for determining future inspection and repair intervals. LaGrow noted, too, that excessive bearing wear could come from coupling misalignment, thermal growth, excessive loading and/or friction, high number of jogging events, and possibly other causes as well.
New alignment procedures to help improve repeatability of service results were scheduled for release in the spring, a replacement coupling with greater tolerance for misalignment is expected to be available in the fall.
Procedures to repair cracks in the exhaust cylinder were next. LaGrow pointed out that some cracking had been experienced by half the units in the fleet and some findings were similar to cracking found on some F frames—low-cycle-fatigue initiated, HCF propagated. No forced outages or downstream damage attributed exhaust diffuser cracking had been reported before the meeting. Background for making repairs is offered in the OEM’s service bulletins.
LaGrow’s summary of product modifications is a good reference list for any owner/operator. Each mod is identified by name and title and a nominal 50-word précis summarizes its content. You can download this from the CEP at your convenience.
Muelheim-based Dr Norbert Henkel, manager of integrated plant solution packages, offered the G users a respite of sorts. All presentations but his focused on frame-specific information vital to the continuity of plant operations in the challenging competitive power market. Users had to pay close attention. Henkel’s half hour allowed delegates to let their minds wander a bit and reflect on enhancements available to reconfigure their plants for the actual—rather than intended— service duty.
He observed that most of the Fand G-frames were ordered/designed as base-load plants but were operating in the intermediate or cycling mode, where operational flexibility is critical to success. High startup reliability, fast load changes/ramps, rapid startup/shutdown, and frequency control and ancillary services are part of this paradigm.
Today, Henkel continued, plants typically have to start up in less than half the time as-designed. This can be achieved but it involves making such changes as increasing HRSG and steam-turbine ramp rates, warming steam-cycle piping faster, increasing the level of automation, changes to cycle chemistry, etc. These are realistic goals, but a well-designed plan is required for successful implementation. Most users will require a comprehensive cost/benefit analysis to justify the investment required.
Henkel discussed each of the major flexibility improvements in turn to illustrate just what is involved in reconfiguring a plant for fast cycling. Starting with the HRSG, he offered ways to minimize the loss of heat during shutdowns to maximize the opportunity for hot starts rather than warm starts. These included automated vents and drains, stack damper, standby fired boiler for preheating the HRSG and for supplying turbine sealing steam, implementation of a stress and fatigue monitoring system, etc.
Steam-turbine flexibility enhancement involves optimization of stress limits, automatic startup sequencer, elimination of unnecessary hold times, stress and online fatigue monitoring of critical components, etc. The resultant ability to achieve what Henkel called “hot start on the fly” enabled a European 400-MW, singleshaft combined cycle in hot shutdown to ramp to full load in 40 minutes.
While some of the enhancements were not quite ready for G-fleet implementation, it certainly is not too soon to think about what you’d like your plant capable of achieving five years from now.
LaGrow was back for a third time just before the afternoon break. It has to be tough presenting to the G users. Most have been at their plants since COD (or earlier) and have a pretty good idea of what works and what doesn’t, how the replacement components have performed, how the plant really behaves under demanding commercial operating conditions, what issues still must be resolved, etc.
Speakers really have to be “on their toes” and “know their stuff,” respect the attendees for their knowledge and experience, be able to put themselves in the plant manager’s position, and not beat the drum too hard or toot the horn too loud regarding “solutions” that have gained the confidence of at least one customer in attendance.
LaGrow organized his last presentation as a technology roadmap for achieving a world-class GT-based generating plant. He discussed technologies available to extend the lifetimes of critical parts, to maximize use of alternative fuels while holding emissions to the lowest levels possible, and the enhancements available to improve performance and operational flexibility.
One example: He dissected the OEM’s advanced combustion system offering, which has been designed for inspection intervals of 12,000 EOH rather than the usual 8000. By the halfway mark in the 12,000-hr beta test, the system had confirmed its ability to operate on a sustained basis at low levels of NOx emissions down to 50% of rated GT output while keeping CO in single digits. Plus, all ranges of combustion dynamics operated below limits throughout the full load range and no new hardware issues were identified.
Then LaGrow presented details on a trip-factor reduction package, compressor sealing improvements, low-load turndown capability, and turbine tip-clearance optimization. He closed with a review of the company’s global experience in burning liquefied natural gas.
Update on upgrades
Eric Kuchinski, GT modernizations marketing, was the final speaker of the day. He had a robust agenda, but the items that seemed to pique the interest of most attendees in the late afternoon were an alarm to warn of a potential icing condition, an inlet-air heating system, and high-temperature transducers for the combustion dynamics monitoring system.
The icing alarm is designed to notify operators when (1) ambient conditions are conducive to ice formation on inlet filters, inlet structures, bellmouth, and IGV assembly, and (2) ambient conditions and IGV position are conducive to ice formation on the first compressor stage. Operator intervention is required to silence the alarm. This can be accomplished by increasing load to the recommended level, shutting down the unit, or (not recommended) ignoring the alarm at risk.
There were two options for the inlet-air heating system: One protects the inlet system downstream of the silencers, the other protects the entire inlet system including the filters. These and other details on available upgrade packages are on the CEP. ccj