Case Studies
Generatortech, Inc. has extensive experience in analyzing rotor winding shorted turns.
Since 1989, over 4000 Generatortech flux probes have been installed in generators worldwide.
We have a large and growing database of test data on these units. Many generators with shorted turns detected using our equipment have been taken apart during this time period and we have been gratified to have had our shorted turn predictions verifed.
This section provides a look at a few case studies that forcefully make the case for flux probe monitoring of your rotors. If you are responsible for making maintainance decisions for your generator sets, you owe it to yourself to arm yourself with the only tool that can quickly and accurately assess the condition of your rotor's insulation system before coming off-line.
While off-line tests are valuable, they can not detect speed-dependent shorted turns (shorts that only develop when the rotor is at rated speed). Tests in balance-pits, with our equipment, has confirmed the presence of speed-dependent shorted turns in a number of cases. These tests allowed the service companies to repair the speed-dependent shorted turns at their facility before shipping the rotors back to their customers, saving enormous time and expense.
Case 1: Coils 6 and 7 Completely Shorted Out of Two-Pole Rotor
This case presents a two-pole rotor with 7 coils/pole that has two coils completely shorted out of the field winding circuit. The top turns of coils 6 and 7 on one pole initially formed a partial short across a blocking wedge that created enough heat to burn through the retaining ring insulation and allowed the top turns of Coil 6A and Coil 7A to form a short through the retaining ring itself (see photo). The resulting coil-to-coil short bypassed all 9 turns in both Coil 6A and Coil 7A (18 total turns). The large decrease in active turns in the rotor winding increase field current requirements by 17% and the heat production in each active turn increased by at least 38%. The coil-to-coil short created a field ground that was not picked up by a faulty ground detection system. The resulting retaining ring damage required required a replacement to be installed.
Coil-to-Coil Short: A short developed across a blocking wedge which created enough heat to burn through the retaining ring insulation, creating a field ground fault. The coil-to-coil short bypassed all 18 turns used in Coil 6A and Coil 7A.
Lead Slots Overlay Graph - shows Coils 6A and Coil 7A completely shorted out of the rotor winding. Also, note the effect of using magnetic rotor wedges on peak size (Coil 1).
Rotor cross-section - shows all 18 turns in Coils 6A and 7A bypassed by a coil-to-coil short
Case 2: Single Turn Short in a Four-Pole Rotor
This case presents a large nuclear generator with a four-pole rotor that has a single turn short in Coil 4-Pole A. The short was not well detected when operating at high loads when the flux density zero-crossing (FDZC) position was far from Coil 4. At low loads, with the FDZC position close to alignment with Coil 4, the short in Coil 4A was obvious. The single short increased the required field current by 1.0% in order to produce the same rotor magnetic field strength with few active turns in the winding. The extra current increased heat production in each active turn by a little more than 2%. The reduction in active turns in Coil 4A reduced heat production in that coil by 12.5%. The reduction in the magnetic field over Pole A as compared to Pole C created a Once/Revolution vibration that tracked with changes in field current.
Rotor cross-section - showing location of the Coil 4A slots with a single turn short out of 7 total turns.
Lead Slots Overlay Graph - large nuclear 4-pole generator with a single short in Coil 4-Pole A. The short is not well detectable at high loads, but is easily detectable at low loads.
Case 3: Repair of Coil 8A Turn Short
The images below show a large 680 MVA generator with 8 coils/pole and 5 turns in Coil 1 and 6 turns in Coils 2-8. A turn short in Coil 8A was not clearly detectable at full load, but was obvious at low loads. The rotor was rewound and the Coil 8A short was repaired, however, full load data could not confirm the repair. Low load data was needed to see that the rewind was successful.
High Load Data - the Coil 8A turn short was not clearly detectable at high loads.
Low Load Data - clearly shows that the Coil 8A turn short was successfully repaired during the rotor rewind.
