Robust presentations and group discussions at the Charlotte (NC) City Center Marriott last week, complemented by tours of three manufacturing and repair facilities, combined to make the 2012 Annual Meeting of the 501D5-D5A Users a valuable learning experience. A significant portion of the technical exchange between the OEM and owner/operators, and among the two-score users in closed sessions, focused on rotor-related issues—vibration in particular.
The group, meeting for the 16th year, is chaired by Gabe Fleck of Associated Electric Co-op Inc, Springfield, Mo. He is supported by Vice Chairman Barry Mayhew, maintenance manager, Cardinal Power, Ontario, Canada, and Director Lonne Grote, lead O&M technician at Rocky Road Power Plant, East Dundee, Ill.
Although the D5 and D5A fleets are not large, Siemens Energy Inc provides them the same level of technical support it gives to its F-class fleet, judging by the quality of the OEM’s presentations and the first-hand knowledge of the engineering leaders making those presentations. The first D5 was installed 30 years ago. Today there are 89 units in operation, with the fleet leader having more than 235,000 reported hours. There are 61 D5As in service, the first achieving commercial status in 1996. The fleet leader has more than 110,000 operating hours. Interestingly, given the relatively low price of gas, the D5s are not running as much as they were a year ago; the D5As are slightly behind last year’s run times.
OEM presentations. Technical topics addressed by the Siemens service engineering team included the following:
• Turbine rubbing distress attributed to a loss of turbine axial clearance has been observed on some R2, R3, and R4 blade platforms. Keep in mind that the turbine cylinder grows faster than the rotor on startup. Rubbing may occur because the vanes, which are attached to the case, move downstream somewhat while the rotor blades hold a fixed position. A root-cause investigation has suggested cutback modifications to increase axial clearances between stationary vanes and rotating blades and these mods have been implemented in a few engines. Field experience is being tracked.
• Tip shroud wear has been observed on some R4 D5A turbine blades, particularly those made of U520 material and having significant operating hours. Blade material was changed to IN738 about seven years ago to improve manufacturability. Hard-facing of shroud contact pads may mitigate the problem. Field tests are in progress. Users were urged not to mix U520 and IN738 blades in a given row.
• The importance of rotor belly bands in preventing the loss of cooling air from between adjacent turbine wheels was stressed. Keep in mind that reduced cooling air flow to rotor components accelerates component degradation because of higher metal temperatures. If wear/cracking of the original two-piece belly bands is experienced, suggested solution is removal and replacement with a proven three-piece design which can be installed in the field.
• Cracks have been observed in the exhaust baffle seals on some units. One concern is the possible recirculation of hot exhaust gas, which could cause overheating of the turbine casing and/or thermal distortion of the exhaust cylinder and some structural elements. Replacement of the seal in-kind requires removal of the exhaust manifold, a time-consuming job. A new seal has been developed, based on F-class experience, which does not require removal of the existing baffle plate for installation. Field testing of the prototype design is in progress.
• A vibration improvement program initiated by Siemens to assess issues experienced by some engines in the 501D5-D5A fleet was said to have the OEM’s highest priority. The effort was initiated following the user group’s 2011 meeting and is being guided by a committee that includes both Siemens engineering managers and user representatives. Independent consultants are assisting Siemens in the evaluation and specialist teams were formed to conduct a root-cause investigation, perform engine testing, model rotor dynamics, evaluate data collected, and make operational recommendations. The comprehensive interim report presented by a half-dozen subject-matter experts at the Charlotte meeting identified promising corrective measures and next steps.
Users unable to attend the conference can access the OEM’s presentations via the Customer Extranet Portal. If you are an owner or operator of a D5 or D5A and not registered to access material on the CEP, and need help navigating the registration process, contact your plant’s Siemens representative. The information you’ll find on the portal is vital to reliable and efficient operation of your engine.
User presentations and candid discussion are the lifeblood of user-group meetings. Several owner/operators presented at the Charlotte meeting. Below are a few notes on four of the presentations; details to follow in the third quarter issue of the CCJ. Users registered on the group’s website can access these presentations as soon as they are posted (in perhaps a couple of weeks). If you’re an owner or operator of a D5 or D5A, the registration process is simple; contact Gabe Fleck email@example.com if you experience problems.
• Exhaust-bearing vibration spikes caused by coking of lube oil in the bearing. This was a particularly interesting presentation, one perhaps of value to owner/operators of gas turbines other than the D5. The affected engine, a D5 in base-load service (only two or three starts annually), experienced plugging of its bearing cap exhaust line with coked lube oil. This led to coking in the bearing along the shaft which caused vibration interpreted as a rub. Plant personnel could not clean out the flex line from the bearing cap to the vacuum manifold line on a regular basis because the unit was running all the time. But when they did, using a homemade Roto-Rooter type of device, vibration levels returned to normal.
Identifying the root cause of the vibration issue and correcting it took a great deal of effort over a period of years. The complete story includes (1) replacement of two vibration probes originally located downstream from the bearing cap with four probes on the bearing cap to warn of plugging earlier, (2) replacement of the original flexible vacuum line off the bearing cap with stainless-steel pipe having a smooth internal surface, and (3) addition of a dedicated exhauster to serve the bearing, thereby enabling regular cleaning of the vacuum line to avoid coke formation.
Important to note is that when the vacuum line coked up, oil squeezed by the labyrinth packing and ran along the shaft until it dripped onto the insulation below. Fires resulted—more than a dozen over the years. The speaker said the dry chemical extinguishing medium and oil created a real mess. This scenario apparently is in the past now with the bearing running normally and within spec regarding lube-oil temperature. Next step for plant personnel is a thorough check of the exhaust frame and associated seals for proper alignment and condition during the next outage. Thought is that the higher-than-desired temperature in the bearing tunnel probably is caused by leakage of exhaust gas into that space.
• Closing keys broke off R1 compressor stage and went downstream, causing FOD. Vibration levels offered no indication of the event or resulting damage. The broken keys, attributed to high hours on turning gear, were found by OEM personnel during the investigation of another matter. A 27-day round-the-clock was needed to make the necessary repairs.
• A loud bang signaled a R4 turbine-blade failure that caused six adjacent blades to snap off (at about 10% of their respective heights from root) and bring the rotor to a complete stop from full speed within two minutes. Resulting vibration caused the bearing oil line to rupture. The engine’s R4 turbine blades were of the latest generation, installed in 2006. The peaker had 400 starts and only about 2000 hours of service on the blades when the failure occurred.
The owner has not yet completed its accident analysis, but believes the failure might have been caused by a combination of shroud wear (see second bullet above in the section on OEM presentations) and a casting anomaly. Thinking is that blade vibration attributed to shroud wear from high turning-gear hours magnified the anomaly, causing the blade to fail. Corrective action taken includes implementation of intermittent turning-gear operation, hard-facing of shroud contact pads, and the addition of root springs.
Considerable discussion on the need for more rigorous nondestructive examination (NDE) of castings ensued. Some users expressed dissatisfaction with the way rotors are being bladed today and suggested that all blades in a given row should come from one vendor and one casting run. Indication was that most blade rows mix and match among casting suppliers and heats. Another thought was to digital x-ray all castings before proceeding further with manufacturing.
• With lunchtime fast approaching Chairman Fleck squeezed in one more user presentation, that by a user reporting on the performance of OEM turbine vanes upgraded by a third party. No time for details or discussion at this point. The takeaway: Third-party mods—specifically, the addition of cooling holes—offered better performance than vanes with the original cooling configuration.
User-only discussion sessions can cover a great deal of territory at a well-managed meeting. Here are a few notes from Charlotte:
• Strut shields. One user was planning to have new shields fabricated. The originals suffered so much cracking and fatigue over the years, he said, further attempts at weld repair were pointless. High exhaust tunnel temperatures were identified with the D5 but not the D5A. A colleague suggested that tunnel temperatures be maintained below 400F to avoid cracking.
Strut-shield integrity is of concern, the first operator said, because if exhaust gas gets by the shield and attacks the strut, damage can cause a shift in the exhaust casing which conceivably could migrate and encroach on R4 turbine blades. Discussion followed on how best to replace struts. Consensus thinking was “one at a time” and to follow procedures developed by F-class users. Replacing all struts was consider a two- or three-day job.
• Weld repair of exhaust-cylinder cracks: Use Inconel 82 wire. Wire matching the material characteristics of the exhaust cylinder material doesn’t work reliably.
• Varnish mitigation always is a topic associated with seemingly endless discussion. At the 501D5-D5A meeting that certainly would have been true if the floor leader didn’t limit the exchange. Experience with various types of oils, “proper” testing methods, etc, all were noted. Does anyone every really solve a varnish problem? If you have, please take a few minutes to respond in the notes section below.
• Mashing of fins in rotor air coolers was another discussion topic. Replacing bundles of copper/nickel tubes with ones of stainless steel was a solution for one user.
• Vibration on engine start received significant air time. Most users participating in the discussion experience vibration levels higher on starts than during normal operation. Numbers above 8 mils were experienced by one user on starts; a couple of attendees thought that too high. Vibration during normal operation was below 3 mils, which others considered normal. An attendee mentioned that the resonant frequency of bearing pedestals was 3500 rpm and this could be associated with the high startup vibration. Yet another user said he has two of the same engines side by side and they have dramatically different vibration profiles. Example: One suffers high-vibration episodes when ambient temperature exceeds 100F. Conceivably, most or all vibration issues will be resolved as the OEM’s vibration improvement program moves forward (see last bullet in section on OEM presentations).
• Proper greasing of trunnions was stressed: Grease them often, grease them well was the advice.
• The subject of proper specifications was brought to the floor. In some cases, the dissatisfaction of contractor performance can be traced to specifications lacking appropriate detail. When this occurs, expectations are not met. Consensus of the group was that specifications should be developed by personnel on the deck plates with expertise in the equipment requiring repair, inspection, upgrade, etc.