Uprating hits a limit
Beginning in the late 1990s, the changing economics of power generation dictated that PGE run Boardman as a baseload unit. With a shift in operating mode, the utility began to make investments in the plant to maximize its output. PGE first retrofit the low-pressure turbines, increasing the generator's output to 626 MVA. Then, in the spring of 2004, the utility retrofitted Boardman's high-pressure turbine, upping the output to 676 MVA. At that point, PGE realized that it made no sense to put more money into uprates before permanently solving the four known generator problems. The solution would be costly, but the payback would be significant—long-term, reliable operation of the plant at a much higher output.
The entire generator overhaul project (with the exception of an upgrade of the excitation control system) was awarded to Alstom Power Inc. (Richmond, Va.). Following are descriptions of the work that Alstom did.
Rewinding the generator rotor. As part of a complete rewind, Alstom installed new class F insulation and 18-18 retaining rings and repaired cracking in the rotor's tooth tops with a short-ring modification. Westinghouse recommended the use of the class F insulation based on its studies of the effect of uprating on rotor operating temperatures.
The rotor was rewound at Alstom's Richmond service center. After removing the retaining rings, Alstom performed MT (magnetic particle testing) at the rotor's tooth tops to check for cracks. The inspection found linear indications of cracking outboard of the snap ring groove at 51 of 72 locations at the drive end (DE), and at 26 of 72 locations at the non-drive end (NDE). Also found were a few indications of cracks inboard of the snap ring groove.
Engineering evaluation of the inspection determined that Alstom's short-ring modification would remove all of the indications and prevent any future cracking problems. The short-ring mod entailed removal of material susceptible to cracking from each side of the tooth top (Figure 2). The few indications of cracks inboard of the snap ring groove were very shallow and were removed by increasing the radius under the tooth top.

2. Cracked tooth repair. Machining of the tooth tops removed potential crack propagation sites by removing the material susceptible to cracking at the tooth top. The photos shown were taken before (top) and after (bottom) the machining was completed. Courtesy: Alstom Power
The rewind added new, class F Nomex slot liners, NEMA G-11 creepage blocks, and retaining rings. The existing blocking was determined to already be class F insulated and was reused. New 18-18 retaining rings then were machined and installed. Balancing was done with the new slip ring shaft installed to ensure proper balancing of the complete rotating assembly (Figure 3).

3. Rotor prep. The rotor was balanced before the generator was reassembled Courtesy: Alstom Power
Installing a new static excitation system. The system included a new slip ring shaft, new brush rigging, and a new excitation control system. Alstom supplied the first two pieces and a new steady bearing as well.
One of the project's requirements was to retain the existing exciter housing. Doing so required modifying it to allow proper air flow for cooling the slip rings and brush rigging, adding inlet and outlet air louvers with filters, and mounting a circulating fan on the slip ring shaft. The new brush rigging features Alstom brush holders that allow the brushes to be inspected or replaced while the generator remains in service.
Retightening the stator core. The original plan for this part of the project was to torque the hex nuts on the ends of the stator's through bolts and building bolts. But Alstom technicians found that friction between the bolts, hex nuts, and washers made it difficult to precisely control the amount of tension applied. Alstom also was concerned that twisting the bolts might damage their insulation. So the decision was made to replace the hex nuts with hydraulic nuts (Figure 4). Hydraulic nuts were chosen for two reasons: They are easily accessible for future retightening, and the tension on each bolt can be precisely set by controlling the pressure applied to its nut.

4. Different nuts. Hydraulic nuts replaced hex nuts to enable precise tensioning of the stator's through bolts and building bolts. Courtesy: Alstom Power
To determine the proper tension for the bolts, Alstom began by setting a nominal value for the interlamination contact pressure, based on company standards. The total force applied to the stator laminations was calculated by multiplying this pressure by the stator core's cross-sectional area. This total load then was divided by the number of through bolts and building bolts to determine the nominal bolt tension.
The existing through-bolt hex nuts at the stator's NDE were replaced with hydraulic nuts in groups of four to ensure that the core did not lose compression. After mounting all of the hydraulic nuts, the through bolts were tensioned in three steps, starting at 80% of the final value. A waiting period between the steps allowed the core to settle, ensuring maintenance of final bolt tension levels.
Following replacement of the through-bolt hex nuts with hydraulic nuts, the building-bolt hex nuts at the stator's DE and NDE were replaced using a similar procedure, but with different tension specs. Because these building bolts are fixed to the frame at the center of the stator, each half of the building bolt tightens half of the core length. After the hydraulic nuts were installed on the bolts at the NDE and DE, they were tightened to 100% of the final value. Final electrical testing of the through bolts, and EL-CID (Electromagnetic Core Imperfection Detection) testing of the stator core revealed no problems.
Designing a new endwinding support system. The new system was intended to minimize endwinding vibrations resulting from plant power uprates and increased MVAR loading. In 1997, PGE studied the problem and concluded that vibration levels would increase to unacceptable levels if the Boardman plant's output was uprated further.
To solve this problem, PGE asked Alstom to design a new endwinding support system that could do the following:
- Reuse existing windings.
- Uncouple endwinding vibrations from the core.
- Enable retightening.
- Reduce measured end-turn vibration to less than 5 mils, peak to peak.
- Shift overall basket n = 2 vibration modes outside of the 110-Hz to 140-Hz range.
Alstom designed a free-floating support system (Figure 5) that met all of the requirements. The system's outer support ring incorporates multiple pressing plates equally spaced around the outside of the winding head. These plates press radially on the outside of the winding head through a special, adjustable tightening device. The system's inner ring supports the interior of the winding head. The pressing plates' compression of the windings against the inner support ring creates a very cohesive endwinding basket structure that is resistant to vibration.

5. Alstom's free-floating endwinding support system. Embedded in its outer support ring are multiple pressing plates equally spaced around the outside of the winding head. Courtesy: Alstom Power
The inner and outer support rings are made of glass-epoxy and filament-wound rings. All other system components are made of high-strength glass-epoxy laminates. At the NDE, the outer support ring also carries new phase rings, which replace the old frame-mounted phase rings. Alstom has used this retightenable system design with great success since the 1970s.
Preparation for installation of the new support system began with baseline inspections and tests. An initial bump test established the baseline data for endwinding vibration characteristics. The DE of the stator had a 4-node mode at 109 Hz and a cantilever mode at 127 Hz. The NDE had a 4-node mode at 115 Hz and a cantilever mode at 127 Hz. The 4-node modes are of concern because they are readily excitable by the 120-Hz forcing frequency of the machine. A comparison of these baseline results to the results of a final bump test, after the system was cured, was used to verify its effectiveness.