The 24th annual 7F Users Group meeting at the Sheraton Denver Downtown, May 11-15, 2015, attended by nearly 300 owner/operators, featured a dozen and a half user presentations—more than any other user organization. There were individual sessions for compressors, combustion section, turbine section, safety practices, performance and controls, auxiliaries, generators, and 7F top 10 issues. Summarized here are three compressor presentations of top interest to attendees:
-
Compressor upgrades, case-by-case basis.
-
Compressor upgrades, fleet-wide solution.
-
S3 vane liberation.
Compressor upgrades, case-by-case basis
Compressors typically open the program because of the high level of interest users have in that section of the gas turbine. One of the first speakers presented on planning for compressor upgrades using a financial risk-based model for project justification and timing. It was in sharp contrast to the old way of thinking in the regulated power business when an upgrade might be approved because it made “good sense” and its capital expense could be included in the rate base.
The goal of the study was to protect against the risk of a forced outage attributable to the following 7FA compressor failure mechanisms identified by the OEM:
The speaker said the ultimate mitigation solution was to install Package 4 or Package 5 upgrades on all compressors in the company’s fleet of about a dozen 7FA.03 engines in simple- and combined-cycle service; nearly a 50/50 split between flared and unflared units. But the financial impact of this strategy was considered prohibitive. Other important facts: Most of the gas turbines in this user’s fleet are now hours-based, having transitioned from starts-based operation; units are on the edge of the area known as the “ring of fire”; and borescope inspections are conducted semi-annually or annually.
Conducting a rigorous analysis proved difficult, the user said, because very little available industry data provided actual failure probability and long-term efforts to get fleet failure data from the OEM had been unsuccessful. However, planning by analyzing consequence only can produce misleading results because it does not include a comprehensive analysis of operational and financial risk exposure. Rather it focuses on the worst-case scenario, which may or may not be a realistic expectation, and the consequence is represented as a constant exposure over time.
Planning by analyzing risk exposure is a more meaningful analytical approach, the speaker said, when failure data are available for the engine model under study. Recall that risk is the probability of the event happening multiplied by the consequence of the event. This method tempers worst-case scenarios, the speaker said, and amplifies realistic scenarios with probability weighting. It also is a “living process” to characterize future exposure changes over time.
Next, the presenter considered quantifying risk in financial terms. Payback period, often used for simple analyses—such as might help decide when to conduct an offline compressor water wash—is not recommended here because it does not account for failure probability and the risk analysis is incomplete. Project net present value versus time accounts for risk but is very time-consuming.
By contrast, “dollars at risk” is a simple calculation: Probability of failure multiplied by the consequential cost of failure. But as noted earlier, determining the probability of failure takes some effort and requires in-depth engine knowledge—absent a publicly available industry database. Simply put, the risk threshold is the cost of the upgrade required to reduce the probability of failure to an acceptable level.
The speaker went on to explain how he and his colleagues conducted their dollars-at-risk analyses for both flared and unflared compressors, mentioning that flared units have a significantly higher probability of failure than unflared compressors—except where aft stator issues are concerned. He presented charts that might be of use to other users (users only); they can be found in the forums section of the 7F Users Group website.
Perhaps the most interesting takeaway from the risk-analysis effort was that the company’s game plan for compressor maintenance/upgrades to maximize the productive lifetime of assets at optimal cost changed considerably after expending the effort to develop probability data. More specifically:
-
Some upgrades were expedited.
-
Some upgrades were strategically planned based on risk exposure, to normalize investment cash flows.
The speaker recommended that before committing big dollars to a one-size-fits-all solution, such as a Package 4 or 5 upgrade, consider evaluating the risks of specific failure mechanisms on each of your engines, with the expectation your findings might identify one or more solutions that better balance risk and cost, and minimize budget impact.
An attendee commented that the highest risk to your engine may not be what you expect; all cases must be evaluated to ensure a proper decision.
Finally, since no analysis is 100% correct, the presenter and his colleagues developed an interim risk mitigation plan to keep a more watchful eye on equipment and to help prioritize tasks during outages. The action plan included the following:
-
Increase the frequency of borescope inspections based on service hours until upgrades are performed to better manage risk.
-
More closely monitor compressor efficiency and increase the frequency of offline water washes to restore efficiency and reduce corrosion risk in the forward stator stages.
-
Audit equipment and plant practices to ensure compliance with TILs 1323-3R1 and 1603 (see link above) regarding online water washes.
-
Develop fleet recommendations for plant-specific application filter technology.
Compressor upgrades, fleet-wide solution
The next speaker, representing a company with more than 20 7FAs, ranging in age from about two to more than 20 years, explained why he and his colleagues supported a holistic approach to mitigating compressor risks. He began by saying a significant number of machines have or had legacy compressor issues—including R0, S0-S4, aft stator rocking, clocking, etc. Although not mentioned, ownership of at least some engines had changed over the years, so different O&M philosophies likely would have been a wild card in conducting an engine-by-engine analysis.
Package 4 and 5 solutions were implemented—including enhanced R0s, S0-S5 solution, bigfoot mod, etc. Experience gained included the following:
-
Onsite tip grinding was necessary, in at least some cases, to assure proper clearances. Suggestion to attendees considering the same upgrade approach was to allow time in the schedule to do this.
-
Bigfoot machining onsite had no negatives, but keep in mind that a generous amount of space is required to implement this upgrade.
-
FME control is a major consideration; its important to keep foreign material out of the machines being upgraded.
-
Onsite work went reasonably well. The same crew was used, to the extent possible, for all Package 4s.
-
Compressor exchanges were necessary when compressor issues identified couldn’t be corrected in the shop.
-
Significant wear was found on shrouded S17 during implementation of TIL 1780. Attendees were advised that S17 requires close attention and that spares are had to come by.
-
R17 dovetail cracking (flat slot bottom) was field-repaired “successfully” by the OEM. However, during a rotor de-stack, a Row 17 wheel was scrapped because of cracking in that region. An audience poll revealed that 35% of the users in attendance had experienced 17th stage cracking and wheels were scrapped on one-quarter of those machines.
-
Field repairs were attempted on two 7FA rotors, one with 600 starts, the other 1300 starts. Blending on one unit during a hot-gas-path inspection failed to meet expectations and the OEM recommended that the 17th stage disc be retired. The other unit had been repaired successfully in the shop during a previous outage, making a follow-up onsite fix impractical.
-
A couple of years ago, this owner had an opportunity to see the results of successful repairs by the OEM from a few years earlier. There was no evidence of re-cracking in the two units inspected. But because the repairs were not well controlled they were not repeatable in the field or in the shop.
-
The owner wanted to monitor 17th stage cracks to see if they were growing over time and figured out a way to do this via borescope inspections. Next goal was to develop an engineering blend to help reduce stresses by 25% and buy time when one or more cracks were identified. A representative model was used for this purpose and a procedure was developed that allowed technicians to blend an entire row in two shifts.
-
When doing a Package 5 in 2014, the speaker said that the discs for Rows 12 through 17 had cracks. Cracks in wheels 14 through 17 exceeded GE blend limits and were scrapped. An exchange compressor got the gas turbine back into operation.
A user mentioned that the OEM has looked into increasing repairable crack lengths and putting units with small cracks back into operation with the proviso for more frequent inspections.
S3 vane liberation—a case history
Incident involved a gas-fired 7FA+e with a DLN2.6 combustion system in cogeneration (base-load) service. Other vitals: flared compressor, uncambered IGVs, nearly 61,500 fired hours, 81 fired starts, 22 emergency trips, COD December 2006. Borescope inspection at scheduled combustion inspection in March 2014 revealed that a 2-in. piece of an S3 vane at the 11 o-clock position liberated and caused downstream damage.
Unit had been running normally with no indication of damage. In fact, vibration and thermal performance data from the OEM’s M&D center indicated only minor changes in the engine’s vibration profile since the previous major in September 2012. There had been no meaningful changes in compressor efficiency or gross power output since the major—impressive since 285 vanes and 366 blades were found damaged in the S3 row and further back in the engine. The IGVs and airfoils in Rows S0 to S2 and R0 to R3 were fine.
Important to note was that the recommendations from a wide range of TILs had been made during the major (1638, 1509-R1, 1780, 1132-2R1, etc), along with fluorescent inspection of all rotor blades and stator vanes. Three cracked S0 vanes were identified and replaced. IGV inner bushings also were replaced. The presenter showed many photos of the damage, including several of the S3 crack surface which was smooth at the trailing edge.
The fracture surface was analyzed, with experts believing the smooth surface was caused by a rubbing together of the two sides of the crack as it propagated inward from the trailing edge prior to liberation. The accident destroyed evidence that might have indicated the crack initiation point. Experts believe polishing scratches found on several airfoils were conducive to cracking.
The fracture surface became noticeably rougher at the end of the shiny area with the fracture then resuming its initial direction of propagation. A typical beach mark fatigue pattern was in evidence up to the final overload fracture at the leading edge. Close examination confirmed crack propagation by corrosion-initiated high-cycle fatigue and that the crack propagated during the starting and stopping of the engine. There had been nine fired starts since the major and no emergency stops. Just south of 12,000 operating hours had been accumulated between the two scheduled outages.
Independent analysis by a highly qualified research organization revealed the crack had initiated at a deep scratch on the pressure-side surface of the vane, close to the trailing edge. In the opinion of investigators, these scratches likely occurred during the final stage of vane manufacture. Further, metallurgists believe the crack initiated where the depth of local scratches on the surface was increased by the formation of local corrosion pitting at the bottom of the scratches.
Finally, investigators were led to believe the crack had been present at the major inspection but was not detected during penetrant testing.
