By David Gandy,
and John Scheibel,
EPRI identified induction heating as the most promising on-site postweld heat treatment technique for repairing cracked turbine rotors and disks.
Repair and refurbishment are increasingly common practices as installed steam and gas turbines age and as power producers seek to maximize unit lifetime, availability and profitability. Performance and reliability have generally been excellent for equipment serviced by both OEM and independent repair shops, producing savings associated with avoided component replacement ranging from tens of thousands to millions of dollars for individual applications.
The demands of the competitive marketplace never cease, however. For low-pressure (LP) steam turbine rotors, rim attachment cracking-caused primarily by stress corrosion cracking (SCC)-represents a growing concern in both nuclear and fossil plants, but current options for resolving significant cracking problems do not meet today's ever-tightening outage schedules. For advanced gas turbines, the stresses associated with high firing temperatures-frequently exacerbated by peaking operation-are compressing overhaul intervals for hot-section components and necessitating significant maintenance sooner than was anticipated, while technology limitations and a restricted vendor base can present risks to continued profitable operation of these valuable assets.
Advanced repair methodologies and practical guidance documents available from or now being developed by EPRI provide comprehensive solutions for less expensive, more effective refurbishment of both steam and gas turbines.
Steam Turbine Cracking
Repairing cracked turbine rotors and disks in fossil and nuclear plants poses a number of challenges:
- Mechanical and metallurgical properties of the repaired area must equal or exceed those of the base disk material and ensure continued operation for the remaining life of the turbine.
- Weld repair and postweld heat treatment (PWHT) must not compromise the shrink fit or distort the disk/rotor, damage bolting, or damage blading that is still installed in the rotor.
- The repaired area must be machined to accept new blading with tolerances at least as good as the OEM's specification.
For more than 15 years, submerged arc welding (SAW) and gas-tungsten arc welding (GTAW) have been used for repairing large rotors and disks at U.S. power plants. However, neither process offers sufficiently rapid deposition rates for successful on-site repair within today's outage windows (e.g., welding and machining must be accomplished within approximately 3 weeks for disk repair during a normal nuclear plant outage). As a consequence, a damaged rotor must be shipped to and from a repair vendor's shop. Off-site repair can consume 3 to 4 months, imposing substantial replacement power costs and lost profit opportunities. In addition, the potential exists for additional damage during equipment removal and transport. Alternatively, component replacement can shorten an outage, but at a cost of $9-12 million for an LP turbine.
On-site rotor repair technology is being developed by EPRI in partnership with TurboCare, Euroweld Ltd., and Mechanical & Materials Engineering. The methodology is expected to considerably reduce cost and outage duration for repair of shrunk- and bolted-on disks. It will also be directly applicable to monoblock designs, where damage in the integral disk region currently necessitates total replacement of the rotor assembly.
Initial EPRI work focused on PWHT, the most formidable technical challenge to on-site repair. Heat treatment is necessary to ensure that a repaired area's mechanical properties equal or exceed those of pre-existing disk material while minimizing the residual stress that can strongly influence SCC. During off-site repairs, resistance heating is commonly employed: a rotor is suspended from a crane and wrapped with "electric blankets." This practice is impractical for on-site repairs.
EPRI identified induction heating as the most promising on-site PWHT approach on the basis of evaluations of alternative heat treatment methodologies and extensive finite element modeling. Induction heating enables in situ rotation of a rotor through a horizontal field that offers precise heat localization and superior temperature control. A full-size mockup demonstration of the induction heating system was performed in February 1999, with temperatures monitored and controlled at some 40 locations. PWHT temperatures were accurately maintained within a narrow band on the spinning rotor assembly, producing the specific temperature gradient between the rim and shaft/seal area defined by EPRI as necessary for minimizing disturbance to shrink-fit connections.
Accelerating welding deposition represented the second challenge to on-site repair. Based on an evaluation of several technologies offering the potential for significantly higher deposition rates, strip clad welding was identified as the most promising candidate for producing high-quality welds. This process, developed for internal cladding of piping and pressure vessels, uses relatively wide strips of filler material, rather than the conventional wire filler. In an initial demonstration of strip clad welding, a 4-inch-thick, 4-inch-wide weld buildup was successfully applied to a segment of an ASTM A470, Class 7 base metal disk using conventional SAW power supplies and special fluxes and process variations. Deposition rates several times those of GTAW and SAW processes were achieved, along with close control of bead shape, thickness, dilution and heat input.
A second demonstration was performed in April 2000 using a 30-mm-wide strip of SCC-resistant filler material on a 25-inch-diameter, 3.5-inch-thick ASTM A470, Class 7 disk shrunk-fit to an 18-inch-diameter shaft. Over a 48-hour period, welding, interim heat treatment and cool-down were performed on a continuous basis to produce a 4-inch-tall, 5.75-inch-wide weld buildup. A typical disk weld buildup may consist of 450-700 lb of filler metal; the strip clad process deposited a total of 486 lb in 32 layers at a rate ranging from 26 -28 lb/hr-roughly 13 times that of cold-wire GTAW and three times that of SAW. Subsequent tests indicated superior mechanical and metallurgical properties. Exceptional tensile and toughness properties were demonstrated for the weld itself, and cross-weld properties (including base metal, heat-affected zone, and weld) were quite good. In addition, the stress profile was very encouraging, with stresses significantly below the yield stress of the base material.
Recently, Turbocare has demonstrated on-site capabilities for machining straddle-mount and axial entry attachments to restore as-manufactured (or better) blade/disk fit tolerances and minimize fretting and fatigue damage. An innovative dual-electrode approach is now being explored by EPRI for further accelerating the deposition rates offered by strip clad welding; this approach also offers the potential for more precise control of the composition of a weldment such that, for example, properties in areas at its top and bottom closely match those of differing blade and disk materials.
The EPRI/Turbocare welding, PWHT, and machining solutions will soon be available for commercial use. Filler materials can be tailored to meet specific application requirements, such as nuclear operating conditions that require high strength and good corrosion performance, fossil conditions in which corrosion may be less of a concern, or geothermal or sour gas conditions in which the environment may be extremely corrosive.
The advanced welding methodology (U.S. Patent 6,145,194) will enable repair of shrunk-on and bolted-on disks and monoblock rotors within the plant fence line during tight outage windows, such as the 25- to 30-day outages recently achieved for major turbine overhauls. A repair itself-including the machining away of damaged blade attachments, setup and strip clad welding, interim heat treatment, and final PWHT-could be performed in 8 to 10 days, leaving ample time for final machining and other well-coordinated overhaul tasks.
Improving Cracking Resistance
Mechanical shot peening techniques are being increasingly applied to remediate damage and improve SCC resistance for aging and new disks in steam turbines. For damaged disks, grinding is employed to remove cracks, then shot peening, a cold-working process, is performed. The component surface is bombarded with small metal spheres, or shot, that act as tiny peening hammers, imparting small dimples. Overlapping dimples create a compressively stressed zone.
Despite its increased use, the effectiveness of shot peening remains controversial. At an EPRI workshop, some OEMs reported that this treatment can delay subsequent crack initiation, while others indicated that shot peening may actually accelerate further crack growth. Concerns focus on whether the cold-worked surface is more susceptible to pitting, and on whether the susceptibility to SCC increases once a pit penetrates the relatively shallow compressive stress layer.
Laser shock peening (LSP) represents an emerging alternative to shot peening. It involves the use of laser-generated shock waves to create the compressively stressed zone in component surfaces. Most data on LSP and shot peening are held proprietary by OEMs and repair vendors.
In EPRI work being conducted by Structural Integrity Associates, laboratory testing and analytical modeling are under way to resolve the controversy over shot peening and to provide credible data on the relative efficacy of alternative peening approaches. Test specimens from 3.5NiCrMoV LP disks were subjected to shot peening at 100 percent or 200 percent coverage or to LSP, and then residual compressive stress gradients were defined. The residual stress field in LSP test specimens is about 4 to 5 times deeper than that in shot-peened specimens (Figure 1). Peened and unpeened specimens are being exposed to an aggressive test environment to determine the time to initiation of pitting and/or cracking. To date, early signs of pitting have been detected only in the unpeened specimens; the beneficial effects of peening are expected to extend the time to SCC initiation by at least 6 months even under the aggressive test conditions.
Later this year, EPRI will deliver practical guidance on the use of LSP and mechanical shot peening for SCC remediation and prevention on the basis of credible data regarding the relative susceptibility of peened and unpeened LP disk materials to SCC initiation, effects of disk yield strength and shot peening coverage level on SCC initiation, and influence of cold working on pitting susceptibility.
Welding Technology for Gas Turbine Refurbishment
Complex, high-strength alloys and superalloys are used for constructing hot-section components in advanced gas turbines to meet the exacting specifications required for high-performance, high-temperature operation. Though fundamentally different in nature, the stresses imposed by both peaking and baseload operation take their toll on these materials, placing a premium on approaches for cost-effective component refurbishment and life extension.
Current welding technology offers limited capability for extending the operational life of superalloy blading. Repairs are restricted to the upper tip region of the airfoil for IN738 and GTD111 buckets; damage in more highly stressed areas can require expensive and time-consuming component replacement. Costs for first-row blades, for example, can approach $30,000 per directionally solidified or single-crystal cast blade and $3 million per row.
In 1998, EPRI initiated a comprehensive program to apply laser welding for expanding repair capabilities over the entire airfoil for cost-effective refurbishment of damaged superalloy blading. For IN738 blading, weld buildups offering high-temperature tensile and yield strength properties approaching that of the base metal have been produced using a higher-strength filler material and a proprietary preconditioning, laser welding, and postconditioning process. Laser welding methods and improved filler materials for GTD111 buckets are currently under development, and later this year the technology will be commercially available from a gas turbine repair vendor. The technology will be suitable for repair of equiaxial (IN738 and GTD111) superalloy blading and more costly directionally solidified (GTD111) blading; additional advances will be required to expand repair capabilities over the entire airfoil for higher-end single-crystal superalloy components.
Repair Specifications for Gas Turbines
Strip clad welding in progress. The filler strip is preheated before being introduced to the weld flux, allowing for a much larger weld puddle and, thus, more rapid deposition than conventional welding processes.
For aging large gas turbines, cost-effective refurbishment of hot-section components becomes increasingly critical to continued profitable operation. Similarly, components in the more advanced F-class machines are in many cases requiring repair or replacement earlier than anticipated. Independent repair shops may offer technical skills and physical capabilities comparable to those of OEM facilities, as well as more competitive pricing. EPRI tools provide guidance for minimizing risks during the vendor selection process to ensure high-quality, cost-effective repair.
Comprehensive repair and coating guidelines for hot-section components in GE MS7001 and Westinghouse W501 machines were published in December 2000. These updated guidelines, developed based on 1992 guidelines that have been used extensively within the industry, reflect extensive input from repair vendors and turbine owners/operators. They incorporate state-of-the-art procedures for inspection, damage assessment, and refurbishment of blades, nozzles, combustion liners, combustor transitions, and coatings to guide run/repair/replace decision-making, generate purchase orders, or support competitive bidding processes.
In addition, complete repair specifications are available for hot-section components in GE 6B gas turbines, as well as GE's more advanced 7F/7FA machines. The specifications cover minimum requirements for weld repair, heat treatment, and other tasks that must be performed for a refurbished component to exhibit service life approaching that of a new component. They specify technical, processing, and quality requirements and include forms for dimensional inspection, damage assessment, and other procedures.
Specifications for repair of combustion liners and transition pieces and first- and second-stage nozzles were issued in 2000; specifications for third-stage nozzles and first-, second-, and third-stage buckets will be published in 2001. Power producers applying these specifications can ensure that the bids submitted by different vendors reflect an identical scope of work, allowing price to be weighed against other important criteria such as technical skills, physical capabilities, past performance, and delivery schedule.
David Gandy is manager of materials and fossil applications at the EPRI Repair & Replacement Applications Center.
Vis Viswanathan is technical fellow and senior manager of materials application technology for EPRI.
John Scheibel is target manager of combustion turbine O&M for EPRI.
Christopher R. Powicki of Water Energy & Ecology Information Services provided editorial assistance.