7F users collaborate on solutions to fleet-wide compressor issues

7F logoCompressors get considerable air time at 7F Users Group meetings. It seems there’s always something to talk about: If it’s not R0, then it’s probably S0-S4, or perhaps R13-R16, or R17. On Day One of the organization’s 2014 conference in Phoenix (May 19-23), Chris Johnston, PSM’s Director of Airfoils Engineering R&D, spent half an hour explaining reliability improvements his company’s engineers have designed into the replacement parts it offers for this machine. On Day Two, several user case histories and an open discussion period added up to another two hours on compressors. On Day Four, the OEM had a 45-min presentation.

Corrosion pitting. This article concerns the Day Two user presentations and discussion. The first speaker reported on corrosion pitting of four gas-turbine compressors, located at the same site, which began commercial service in 2000. He said excessive pitting was identified in 2001 and attributed to carryover of poor-quality water (containing chlorides) from the evaporative coolers. A reverse-osmosis system was installed about two years after COD to pretreat the raw surface water; revised operating procedures eliminated carryover.

However, the damage done in the first two years was irreversible and, despite the use of quality water for the next 10, an R1 blade in Unit A liberated because of suction-side corrosion at the root of the airfoil. At the time, the unflared compressor had accumulated only about 1100 starts and 5500 operating hours.

In May 2013, the plant owner had the OEM conduct “enhanced” borescope inspections (100% dye penetrant and UT) on Units B, C, and D. A definitive assessment was not possible because the risk factor associated with pitting could not be defined. No cracks were identified in any of the airfoils, just elevated levels of corrosion. Had cracking been found, the affected units would have been removed from service immediately.

The OEM’s recommendation was to re-inspect the units after the run season. The speaker mentioned that GE was selected for the inspection because corrective action would be based on those results and the OEM wouldn’t make decisions based on information from another company’s inspection.

A detailed analysis of corrosion pitting was conducted on Unit B during a scheduled outage. It was carefully planned out, with specific areas of specific blades selected for inspection in the first three compressor stages. Fluorescent penetration inspection and dental molds were used with the goal of identifying the “showstopper” as soon as possible—if one existed. What the inspection team found were flaws identical to those that caused the Unit A failure, but larger. Also found were flaws in non-repairable areas of some airfoils, making the business decision to buy a replacement rotor relatively easy, especially considering time constraints.

For the remaining two units, the following O&M plan was adopted:

      • Inspect the suction side of R1 blade roots for cracking.

      • If no cracks are detected, clear the units to run a specified number of starts prior to re-inspection.

This process would time to schedule unit outages to re-blade the rotors.

Once pitting begins, it’s only a matter of time before the blades will require replacement. Allowing a machine to operate with severe pitting seems analogous to Russian roulette: The decision to run essentially is based on qualitative information. You hope there’s no bullet in the chamber and the next time the machine is started is not the last time.

The speaker strongly advised his colleagues in the audience to maintain inlet filter houses and water quality for evaporative cooling and compressor washing systems in top condition right from commercial start; also, to maintain tight control over droplet size of any water entering the compressor from fogging and washing systems. The best way to avoid pitting and erosion of compressor blades, he said, is not to allow it to happen.

Blending of 17th-stage discs. The next user presentation concerned cracking of the 17th-stage compressor wheel and experience with poor quality rework. The CliffsNotes version of the industry’s experience according to this owner/operator: lots of cracking and blending, lots of replacement wheels, lots of extended outage, cracks on both the forward and aft faces of the wheels, cracking observed after only about 1000 cycles, cracks up to 200 mils deep (OEM blend limit: 100 mils), in-situ blending experience with and without blades, some owners leaving cracks.

The speaker reported on his company’s experience with a new precision-engineered blending process from AccTTech LLC said to reduce critical wheel stresses by more than 20%. Inspection revealed that most of the 60 slots in the disc repaired had cracks; maximum depth was 95 mils. The speaker said AccTTech’s engineered tooling for uniform machining and polishing removed all cracks in two shifts, a fact verified by 100% NDE. Longer disc life is expected because of the lower-stress profile created. Come to the 2015 meeting for a progress report.

Upgrade options. A user conducted a short review of the OEM’s upgrade options for both flared and unflared compressors to help colleagues better understand what’s available to deal with R0 erosion, R17 wheel cracking, stator twisting, and other issues. A user poll earlier in the morning identified Package 3 as the option of greatest interest to attendees.

Included in the presentation was a review of applicable Technical Information Letters (TIL) and the different requirements for original parts and enhanced parts. Example: TIL 1603, which concerns R0 erosion and water ingestion recommendations, requires leading-edge dental molds for pre-enhanced parts, no dental molds for enhanced R0s.

The speaker pointed out that while there were strict limitations on water washing of compressors with standard R0 hardware, the enhanced hardware allows up to 30 min/day of online washing. Currently, he continued, no erosion maintenance is necessary when enhanced R0 hardware is installed; however, long-term “significant” fogging may require erosion mods or chord measurements. The higher erosion tolerance of the enhanced R0s is attributed to retuning of the airfoil, larger fillet, and laser shock peening (LSP).

Fogging-related R0 erosion was covered by another user—a blow-by-blow account of the owner’s experience with various modifications to the OEM’s fogging system (pumps, nozzles, valves, control logic, and instrumentation) over a period of years. Plus the switch from a dedicated pump skid to fogging water extracted from the IP feed pump. One of this company’s plants was among the first to install enhanced R0 blades (late 2009/early 2010) on the promise that the new airfoil was highly erosion tolerant and that no restriction on fogging was expected and erosion would be self-limiting.

Ooops. Inspection after about 600 hours of fogging revealed significant erosion visible to the naked eye; after 1500 fogging hours the OEM became concerned and new sets of enhanced R0 blades were installed in all three engines that had transitioned to the enhanced design. EPRI was asked by the owner to characterize the erosion rate and draw an educated conclusion, based on its previous modeling work, as to how much erosion is too much.

A concern about severe erosion pitting was the possibility that it could lead to crack formation just above the platform on the leading edge. A crack there could propagate from vibration caused by a rotating stall. Laser shock peening was a key step in the OEM’s effort to make the enhanced R0 blade erosion-tolerant. It imparts a compressive stress to a depth much greater than standard peening—somewhere between 50 and 100 mils. If erosion or corrosion penetrates the compressive layer, the speaker added, it’s basically back to the same issues experienced with the standard blade.

The presenter digressed for a moment to be sure everyone in the room understood the terms important to the discussion. He explained the difference between chord loss and pitting depth. The latter is how deep the pitting caves extend into the airfoil from the high point of the leading edge (LE). This is not the same a chord loss—the dimension from the original leading-edge surface to the high points of the current (pitted) LE surface. Thus chord loss plus pitting depth is the parameter of concern when performing a stress analysis.

The bottom line:

      • Data collected suggested a chord loss of from 26 to 36 mils in both the LSP and non-LSP regions of the leading edge after 7000 fogging hours. Approximately 50 to 65 mils of total erosion is considered the critical depth, with failure likely.

      • The erosion rate is not linear, based on EPRI analysis. The curve of chord loss versus fogging hours resembles a characteristic curve of e to the minus x power.

      • Rotating stalls drive the potential for cracking.

      • In fogging applications, the user’s company considers enhanced R0 blades a one-HGP-cycle part.

The speaker closed with the following thoughts:

      • Blending of erosion pitting and the reapplication of LSP might be a worthwhile research project.

      • Consider applying LSP to the entire leading edge, not just the lower portion.

      • If you have the choice between an F-class unit supplied with either a fogging system or evaporative cooler, opt for the latter. The speaker said his company has an evap-cooler-equipped 7FA.03 and it has experienced virtually no erosion.

Posted in 7F Users Group |

Comments are closed.