Generators & Electrical Components: Tools for Monitoring Generators

On-line monitoring of generator condition can help ensure the unit operates safely and can minimize maintenance outages. This article provides examples of two available monitoring tools: a magnetic flux probe for rotor windings and a rotor pole temperature monitor.

By Mladen Sasic, Steve R. Campbell, Randy Wallman, and David Wong

Monitoring the condition of a hydro generator is a vital step toward obtaining information on degrading material or incipient failures. This monitoring can be accomplished in a number of ways. This article presents information on the characteristics and application of two types of tools available for on-line monitoring of generators: a magnetic flux probe for rotor windings and a rotor pole temperature monitor.

Click to Enlarge
The ThermaWatch Rotor was installed through the stator hole ventilation of Unit 3 at the 50.5-MW Mercier plant to measure rotor pole temperatures during unit operation.

Magnetic flux probe

Although the most reliable and common way to detect shorted turns and incipient ground faults on rotor salient pole windings is to perform a “pole drop” test, this test has three disadvantages:

— It can only be performed with the generator shut down;

— It is time-consuming to perform, especially on a large rotor with many dozens of poles; and

— Centrifugal forces are not occurring because the unit is shut down, and thus some shorts may not be present that will be present at normal rotating speeds.

As hydro project owners look toward minimizing the work performed during unit shutdowns and move toward predictive maintenance, there is a need for an on-line tool that can replace the pole drop test. Magnetic flux monitoring in the air gap between the rotor and stator has been used for years to detect shorted turns on the rotors of high-speed steam-turbine generators.1 This monitoring involves measuring the magnetic flux in the air gap to determine if field winding shorts have occurred in the poles. The magnetic flux is detected by means of a flux probe. Each rotor pole sweeping by the probe induces a voltage that is proportional to the change in flux from the passing pole. The voltage is then measured by a rotor flux analyzer.

Shorted turns will cause a change in the flux profile within a pole at a given load. As each pole passes, there will be a peak in the induced signal. Each peak of the waveform represents the peak flux across one rotor pole. Any turn short in a pole reduces the effective ampere-turns of that pole and thus the signal from the flux probe. The recorded waveform data can then be analyzed to locate the poles containing the fault.

To apply this technology to salient pole windings of hydro generators, in 2002 the Electric Power Research Institute (EPRI) funded a research project.2,3 EPRI is an independent non-profit company performing research, development, and design in the electricity sector. This research resulted in development of the HydroFlux Monitor, which was the previous generation of the magnetic flux probe.

Characteristics of the tool

A new type of probe was designed for use in a hydro generator.3 The probe is comprised of a number of circuit layers, printed on a flexible base material. The flexible probe is installed on a stator tooth and has a profile of just 0.125 inch for use on hydro generators with small air gaps.

Click to Enlarge
The RFA II portable magnetic flux probe monitor is installed in a hydro generator to detect shorted turns and incipient ground faults on rotor salient pole windings.

The portable magnetic flux probe monitor is called RFA II. The instrument is equipped with inputs for different types of flux probes, so it can be applied on machines with probes already installed. In addition, three different synchronization methods are possible: dedicated from a shaft-mounted marker, internally to alternating current (AC) power input, or externally to any other AC signal. And v II is capable of fast data acquisition at a high sampling rate.

Different communication protocols (such as USB, LAN) can be used for connection to a personal computer. Semi-permanent data acquisition without a PC also is possible. Two stand-alone data acquisition modes are available. The first mode, which is time based, can be used to collect data in user-specified time intervals, as short as one measurement every five seconds. This method is useful to collect data automatically during fast load changes. The second method is designed to save only the data with significantly different flux patterns, avoiding storage of unneeded data.

All measurements are stored in RFA II internal memory and can be downloaded to an Access database.

Application of the tool

Figure 1 shows data collected from a flux probe installed in a hydro generator. The generator was operated at no-load conditions, resulting in a symmetrical shape for both the raw signal (blue line) and integrated values of raw signal (red line). Because the signal shape changes at different unit loads, raw data should be integrated for further processing.

Click to Enlarge

In addition, comparison of one pole to another can be used to indicate shorted turns, as any change in the flux profile within a pole at a steady load must be due to shorted turns. Such comparisons are usually shown as polar plots (see Figure 2). Minor differences in measured flux level between poles and ex-centricity of the rotor will result in a less-than-perfect circle, as seen in Figure 2.

Click to Enlarge

In most of the field tests, correlation of results from on-line flux monitoring and off-line pole-drop tests was not possible. One reason is lack of positive pole number identification in off-line and on-line tests. The other reason is that the two tests are significantly different in approach and in test conditions. The pole drop is more sensitive to first turn shorts but less sensitive to other turn shorts. The on-line flux measurement is sensitive to all shorts proportionately. In some tests, both methods identified one pole with shorted turns. Because there was no rotor synchronization, we cannot be completely sure that these tests indicate the same pole.

Rotor pole temperature monitor

As hydroelectric facilities age, owners must make the best use of existing equipment. To do this, plant owners are uprating their units, which often means increased runner efficiency and corresponding increased electrical stress on the rotor structure. This can increase overall rotor temperature and lead to premature aging of the machine. Measuring rotor pole temperature is an important step toward ensuring proper operation of the machine under new loading regimens.

The Institute of Electrical and Electronics Engineers Inc. (IEEE) provides guidelines for both rotor winding temperature limits and method of temperature detection.4 IEEE states the total temperature of the hottest spot on a rotor winding should not exceed 130 degrees Celsius (C) for Class 130 B for insulation systems; 80° C rise at 40° C coolant). IEEE also states that the preferred temperature measurement method is the resistance method, which involves running a unit at rated load and comparing the resistance of a winding at a known and an unknown temperature. These temperatures can be verified on-line and correlated with other machine parameters.

Other possible applications include indirect detection of shorted turns, which may cause temperature variation in the absence of a flux monitor. Off-line tests exist, such as the voltage drop test.5 However, this test is limited, as shorted turns are sometimes only detected when centrifugal forces are acting on the rotor, meaning the unit must be operating.

To overcome the limitations of off-line or indirect monitoring, VibroSystM developed its ThermaWatch Rotor temperature monitor. This tool allows direct temperature readings at specific locations around critical components of each rotor pole.

Characteristics of the tool

The most important components to be measured are the windings, pole joints (failure), and rotor pole face. However, two main obstacles to on-line temperature monitoring have been limited access to rotor poles and significant noise variables acting upon measuring devices. Design of the probe for the monitor addressed these two obstacles. The probe was designed to mount through the stator ventilation ducts, facing the rotor poles, flush with the inside wall of the stator. And noise variables (such as varying stator temperature and air gap temperature) are filtered out of the output reading.

If the pole face is to be monitored, the probe would be mounted 90 degrees in relation to the center of the rotor pole face. If pole joints are to be monitored, the probe would be mounted 90 degrees from the pole joint. The probe can be adjusted, based on stator thickness, so that the tip always mounts flush with the stator wall.

Three probes are ideal on large machines. The top and bottom probes monitor the temperature of inter-pole regions of the top or bottom pole joints, depending on the rotor design. This is in addition to the pole face and/or amortisseur bars when the pole face is in front of the probe tip. The probe mounted midway up the pole is to read pole face and/or amortisseur values as a pole passes the probe tip. In addition, the probe's sampling rate allows it to pick up inter-pole temperature at the mid-point of the pole height.

The probe was designed to mount radially (with respect to the rotor), rather than axially from above/below the field. Axial installation could provide information on rotor winding temperature, but there are several limitations. First, rotor pole face temperature would be neglected. Second, depending on the rotor design, ventilation systems (including fan blades and air deflectors) make it difficult to position the probe axially. Third, positioning the probe from above incurs a risk of probe or bracket pieces dislodging and falling onto the moving rotor. Fourth, accumulation of dirt, dust, and oil are factors when positioning the probe from below, as the probe tip is facing upwards.

Application of the tool

In 2003, personnel at the 50.5-MW Mercier plant on the Gatineau River in Quebec, Canada, had completed work to upgrade runner efficiency for five units. During commissioning of these units, generator temperature exceeded the expected normal values. Temperature stickers were applied to rotor poles, and VibroSystM's ThermaWatch Rotor was used in September 2007 to provide additional information. Results showed acceptable temperature values, confirming that pole overheating was not the issue. Hydro-Quebec subsequently found the cause of overheating to be insufficient ventilation.

Unit 3 at Mercier contained four air gap sensors, installed at 0, 90, 180, and 270 degrees from upstream. Because of time and access limitations, personnel installed only a single ThermaWatch Rotor at 243 degrees through the stator wall, near the top of the stator. The probe tip was aligned flush with the stator bore wall, facing the rotor.

Air gap Sensor 3 (at 180 degrees) was chosen for comparison. The two parameters — air gap at 180 degrees and rotor pole face and winding temperatures at 243 degrees — would be compared for data interpretation.

This unit was controlled by ZOOM diagnostic software from VibroSystM. ZOOM software accounted for the 63-degree offset between the two probes, and it was possible to correlate the values with the dynamic air gap as measured by Sensor 3. Data was collected and stored in sampling mode over a three-hour time frame at nominal load (10 MW). Ambient temperature inside the machine was 75° C at full load, hot operation. ThermaWatch Rotor readings indicated the hottest spots detected did not reach 130° C.

Sampling (time-based) and pole (pole-reference-based) measurements are both used to display temperature in the software. Sampling graphs allow for display of signal shape, which, in this case, represents the actual rotor pole shape. When looking at sampling graphs displaying air gap and temperature between 4:43 p.m. and 6:15 p.m., two phenomena become apparent.

First, average temperature of the rotor pole faces and inter-pole regions are cooler earlier in the day. At a static operating load, this is easily explained by relatively linear heating of the pole components as the machine reaches nominal temperature conditions. Second, the signals representing temperature readings earlier in the day are relatively flat. As the machine heats up, the signals take on the shape of the rotor poles, and the delta of the pole face relative to interpole regions increases. For example, interpole region “pole 2 - pole 3” compared to pole face 2 has a delta of -2.71° C at 4:43 p.m. Comparing the same regions again at 6:15 p.m. under the same operating conditions, region “pole 2 - pole 3” has a delta of -14.69° C (see Figure 3).

Click to Enlarge

The ThermaWatch Rotor pole temperature signal takes on a shape similar to the air gap signal and therefore similar to an actual pole. Moreover, both signals move in unison (there is no offset). These data suggest that interpole temperature is higher than pole face temperature. This can be explained by the heat dissipation off the field coils and lack of sufficient ventilation in the unit housing.

Pole measurements show similar results. Measurements were compared over a 150-minute time frame beginning at 3:33 p.m. As with time-based graphs, individual pole face temperatures became more defined as the machine heated up. Poles 22 to 24 consistently show slightly higher temperatures when compared at different times from 3:33 p.m. to 5:26 p.m., as well as when compared from turn to turn. However, the differences are negligible at 1 degree. Pole 23 shows the most increase in temperature from initial to final pole measurement (31° C), and it remains the hottest pole at 5:59 p.m. at 90.7° C. This maximum temperature is somewhat high but within acceptable range.


1Stone, Greg C., Edward A. Boulter, Ian Culbert, and Hussein Dhirani, *Electrical Insulation for Rotating Machines: Evaluation, Aging, Testing, and Repair*, Institute of Electrical and Electronics Engineers Inc. and John Wiley & Sons, Hoboken, N.J., 2004.

2Stein, Jan, “Field Testing of Continuous Flux Monitoring for Hydrogenerators: Feasibility Study” (1011282), EPRI, Palo Alto, Calif., and New York Power Authority, White Plains, N.Y., 2004.

3US Patent No. 6466009, issued Oct. 15, 2002.

4”IEEE Std C50.12 - 2005 IEEE Standard for Salient-Pole 50 Hz and 60 Hz Synchronous Generators and Generator/Motors for Hydraulic Turbine Applications Rated 5 MVA and Above,” Institute of Electrical and Electronics Engineers Inc., New York, 2006.

5Lyles, John F., “Generator Maintenance Course,” G.E. Armstrong Enterprises, Inc., Pickering, Ontario, 1994.

Mladen Sasic, P.Eng., is engineering director for Iris Power. Steve Campbell is vice president of R&D and a cofounder of Iris. Iris is the licensee for the magnetic flux probe. Randy Wallman is a consultant for VibroSystM Inc., servicing western Canada, Alaska, and the Pacific Northwest in the U.S. David Wong is commercial representative for the southwest region of the U.S. for VibroSystM.

This article has been evaluated and edited in accordance with reviews conducted by two or more professionals who have relevant expertise. These peer reviewers judge manuscripts for technical accuracy, usefulness, and overall importance within the hydroelectric industry.

More Hydro Review World Current Issue Articles
More Hydro Review World Archives Issue Articles

To access this Article, go to: