Pipe-repair odyssey

Griffith Energy, like many combined cycles built at the end of the 1990s and in the early 2000s, clearly illustrates the commitment, persistence, and capital demanded to ensure the levels of reliability and performance required to succeed in the highly competitive generation business.

The plant, located 14 miles west of Kingman, Ariz, is a nominal 570-MW 2 × 1 combined cycle owned by Houston-based Star West Generation LLC and managed by Mike Hartsig.

Construction start was September 1999, COD January 2002. Resources were strained during this period: Many power producers were lined up in the queue waiting impatiently for equipment; plus, experienced construction/startup personnel were difficult to find. Regarding the first point, gas-only Griffith Energy was assembled with principal equipment from multiple sources. The gas turbines are US-made GE 7F.03s, the supplementary-fired heat-recovery steam generators are triple-pressure boilers from Europe’s NEM, and the condensing steam turbine/generator was manufactured by Asia’s Toshiba.

Griffith was built by BVZ Power Partners, which combined the capabilities of Black & Veatch and Zachry Construction Corp, for the original 50/50 owners PP&L Global and Duke Energy North America. LS Power became the sole owner in May 2006. Star West was established in 2011 and purchased both Griffith and the nearby Arlington Valley Energy Facility from LS Power.

A good starting point for this article is 2010 when Tetra Engineering was hired by plant management to develop inspection plans for P91 high-energy piping (HEP) components, flow-accelerated corrosion (FAC), and attemperators. Tetra’s approach, the company’s Peter S Jackson, PE, told the editors, was to identify the highest-risk locations based on industry experience and conduct inspections in priority order.

The weld-on lateral fitting in the hot-reheat system (trademark Latrolet) used to blend HRH steam from the two HRSGs into a single pipe, was a high priority based on experience at plants designed by Black & Veatch.

Jackson said a team of Tetra engineers, including Robert Hookway, PE, a member of the ANSI B31.1 committee, performed hot-condition walkdowns and baseline visual inspections in summer 2010. One outcome: Hangers needing attention were identified and corrected as needed.

The Tetra team returned in spring 2011 (during the plant’s scheduled outage) to follow-up on the baseline work performed the previous summer. Ultrasonic inspection of critical areas identified non-through-wall cracks on both sides of the 28-in.-diam laterolet (Figs 1, 2). The cracks were approximately 16-in. long and within a few thousandths of being through-wall.

Griffith 1,2

A comprehensive inspection of the crack areas was conducted by Tetra and an independent welding contractor. A complete re-evaluation of the piping system was done using the latest engineering software, replicating the original analysis by B&V. There was no indication of ultra-high stress under any operating condition; no cases run exceeded B31.1 stress limits. However, this assumes all materials and installation work were perfect, which they were not.

After Tetra informed plant management they could not be onsite for the repair, Structural Integrity Associates Inc was engaged to meet with Tetra and plant management to review repair procedures. Plant management chose SIA as the technical advisor (TA) to work with the welding contractor.

Both Tetra and SIA were charged with determining the root cause of the cracks. Lab testing of material removed from the crack areas revealed a combination of stress and creep typically associated with P91. Prior to and following the repair, stringent pre- and post-weld heat treatment procedures were required to meet current standards for working with P91.

After grinding and weld prep, 9018-B9 rod was used to weld the joints, meeting all requirements and expectations. Management worked closely with Tetra to develop a comprehensive program for checking welds in the lateral and associated piping, within the heat- affected zone, annually from that time forward.

The lateral was performing as expected until one day in October 2014 when roving operators noticed water dripping from blanket insulation that had been placed on the laterolet following the 2011 repair. First thought was the blanket insulation might have been saturated with rainwater.  

After inspection of upstream piping and other possible water-ingress points, plant management concentrated on all weld areas of the lateral. Isolating the leaking area in the insulation, management identified welds in the crotch as the likely area.

After consultation with Tetra and corporate management, consideration of operating conditions at the time, and the nature of the dripping, suggested an immediate shutdown was not necessary. Access to the area was restricted to protect personnel.

The plan was to monitor the leak closely by infrared thermography and should the volume rate of drip flow increase, Griffith would be taken out of service immediately for a thorough assessment. Infrared was effective because the blanket insulation was saturated and heat was conducted outward.

In February 2015, with the plant offline for the first time since the leak was noticed (Griffith was running continuously either 1 × 1 or 2 × 1 at the time, not cycling), Tetra returned to the Arizona facility. NDE inspections of the lateral and associated piping verified cracks in the “crotch area” of the laterolet. At least one was through-wall, more likely two. These cracks were not related to the work done in 2011. Phased-array and mag-particle techniques were used to identify and characterize the cracks.

The major cracks were evident to the naked eye. Fig 3 shows a nominal 3-in.-long vertical through-wall crack in the 28-in. lateral deep in the crotch, and a 10-in.-long horizontal crack (OD)—5 in. of which was through-wall—at the bottom toe of the weld above the crotch.

Griffith 3

This section joined the 24-in. Unit 2 HRH pipe to the laterolet. The vertical crack, which exhibited small horizontal fissures coming off the bottom of the weld (removed by grinding), extended 0.25 in. into the lateral and 0.75 in. into the 28-in. pipe; however, the crack was through-wall only in the weld area.

This crack propagated from the inside out, originating from the pipe bore or a buried defect acting as a stress riser within the weld. Close inspection suggested craftsmanship issues, such as the possible use of incorrect welding rods.

Scott Snoddy, the project manager for welding contractor Durus Industrial, collaborated with Jackson and the Tetra team on the current weld repair. This repair would be much more difficult than the 2011 work because a new process was required to avoid the need for argon dams in the pipe to run a root pass. The process was required to avoid further cuts in the piping.

The repair was designed to conform to ANSI B31.1; the weld procedure and welders were qualified to ASME Section IX. A third-party Level III inspector from Phoenix National Laboratories Inc was retained to verify conformance with the procedure and to inspect the work.

For the vertical weld, about 10 in. of material was removed and a gap in the wall of about 3/8 in. was created to access sound material. Preheat to 400F for the excavation was provided by a non-carbonizing flame. Proper preheat was confirmed using temperature sticks. A borescope verified a proper root pass from inside the pipe. Mag-particle testing prior to welding confirmed that all cracks had been removed. Phased-array UT also was conducted at several locations to assure no subsurface cracks existed.

A post-repair review of the laterolet failures and of P91 properties gave plant management no confidence that the fitting would not fail again. Further, metallurgists recommended that P91 not be heat treated more than three times to avoid compromising its structural properties and creating an end-of-life condition.

Plant management was aware of varying opinions about heat treating P91 but choose a conservative route. Another compelling reason to replace the laterolet with a forged tee as suggested by Tetra, B&V, and others was that the type of damage Griffith found at 50,000 hours of service other plants had experienced at 25,000 hours.

Final step. Griffith Energy’s goal in replacing the fabricated laterolet with a forged tee was based on the need for a reliable fix that would be problem-free for the remainder of the plant’s life. While the project might look simple given the simple sketch in Fig 4, it was challenging. Timewise, the retrofit took nine 24-hr days.

Griffith 4

The forged tee was not available off-the-shelf. It had to be custom-made, in Italy. Houston-based US Metals handled the procurement and also supplied the required sections of 24- and 28-in. pipe, plus a 28 × 24 in. reducer to replace the original found to be questionable, while removing insulation for the repair.

Then there were fit-up challenges, because the pipe diameter of some parts being welded together differed slightly. Welding was challenging because purge dams had to be installed and moved from weld to weld as work was completed. All final work had to be done in the air off scaffolding which had to be reinforced to accommodate the weight of the pieces.

The next challenge was accessibility given the location of structural members which required the new and old pieces to be rigged in such a manner has to “snake” through the members. Project coordination, the plant’s responsibility, also had its moments given the number of contractors involved.

The tee and pipe sections required were shipped to Durus’ shop where approximate cuts were made for the pup pieces before moving them to Griffith where final cuts and weld prep were done. The job involved seven welds, one in the shop (the Unit 2 pup-piece-to-tee weld) and the remainder in the field. For the field welds, MT was performed after the root pass, phased array after filler weld was deposited over the root, and phased array again after post-weld heat treatment.

The maximum weight handled by the crane maneuvering components among structural members was about 3800 lb. Scaffolding was designed for 5000 lb. Scaffolders were at the plant on standby for the project to minimize the delay time in making any platform changes required.

After a little over 48 months, the previous repairs, the addition of the forged lateral, Star West and plant management had a long-term fix for a heretofore troublesome component. All in all the project was a success bringing together several contractors, sophisticated welding and rigging procedures, and most important, a safe job—no injuries or near misses.

Posted in HRSGs |

Comments are closed.