End Winding Discharges Detected with Partial Discharge Monitoring

Partial discharge monitoring has become an essential tool for utilities managing aging generator fleets. This case study from BC Hydro illustrates how continuous online monitoring can detect deteriorating insulation conditions and guide effective maintenance interventions.

The Challenge of Aging Generator Assets

Hydroelectric generators are designed for long service lives, often operating reliably for 40, 50, or even 60 years. However, as these machines age, the stator winding insulation systems can experience various forms of degradation. The challenge for utilities is detecting these problems early enough to perform planned maintenance rather than responding to unexpected failures.

Traditional periodic testing during scheduled outages provides only snapshots of insulation condition. Between these tests—which may occur years apart—conditions can change. Continuous online monitoring addresses this gap by providing ongoing visibility into the health of the stator winding while the generator operates under normal conditions.

Case Background

Company: BC Hydro, British Columbia, Canada

Machine Specifications:

• Type: Air-cooled hydro generator.
• Rating: 239 MVA, 13.8 kV.
• Age: 37 years old.
• Manufacturer: Confidential.

At 37 years of age, this generator represented a mature asset within BC Hydro's fleet. While still providing reliable service, the machine warranted close monitoring to ensure continued safe operation and to enable proactive maintenance planning.

The Monitoring System

BC Hydro equipped this generator with a comprehensive partial discharge monitoring system. The installation consisted of six PDA couplers (capacitive sensors) on each of the three phases—eighteen sensors in total. These sensors were connected to a HydroTrac continuous monitoring instrument.

Capacitive couplers detect the high-frequency signals generated by partial discharge activity within the insulation system. By continuously capturing and analyzing these signals during normal generator operation, the monitoring system can identify changes in discharge patterns that indicate developing problems.

The advantage of continuous monitoring versus periodic testing is the ability to observe trends over time and detect problems as soon as they begin to develop, rather than waiting for the next scheduled outage.

What the Data Revealed

The HydroTrac monitoring system detected partial discharge activity specifically in the end-turn region of the stator winding. End-turns (also called end-windings or overhangs) are the portions of the coils that extend beyond the stator core on both ends of the machine. These regions are particularly vulnerable to certain types of insulation problems because they are exposed to the internal atmosphere rather than being protected within the core slots.

The discharge activity was most pronounced on Phase C of the three-phase system. Analysis of the partial discharge patterns provided important diagnostic information about the nature and severity of the problem.

Understanding the Diagnostic Signature

The partial discharge data from Phase C were displayed in a two-dimensional plot showing the relationship between discharge magnitude and the AC voltage cycle. This type of visualization is a standard diagnostic tool in partial discharge analysis.

As shown in Figure 1, the plot revealed a large hump at high discharge magnitudes. This hump appeared in both the positive and negative portions of the AC cycle, indicating discharge activity occurring throughout the voltage waveform. The magnitude levels and pattern of this activity indicated a condition requiring attention.

_{Figure 1: 2D PD Plot Showing Hump at High Magnitudes – Phase C}

The ability to characterize discharge patterns in this way is one of the key advantages of modern partial discharge monitoring. Rather than simply detecting that discharge is present, the system can provide detailed information about discharge magnitude, frequency, and phase relationship—all of which help diagnose the underlying cause.

Maintenance Response

Armed with specific information about the location and nature of the discharge activity, BC Hydro's maintenance team was able to plan a targeted intervention. The data from the PDA couplers indicated that the discharge was occurring in proximity to specific sensors, allowing personnel to focus their efforts on a defined area of the winding.

The Repair Process

The maintenance work focused on the ten stator bars located closest to the PD couplers that had detected the elevated discharge activity. These bars were subjected to a two-step repair process:

1. Cleaning: The surfaces of the ten identified bars were thoroughly cleaned.
2. Recoating: Following cleaning, the bars were coated with red glyptal epoxy.

Glyptal epoxy is a standard insulating material used in electrical machine applications. The recoating process restored the insulation surface on these bars.

Measurable Results

One of the most valuable aspects of continuous monitoring is the ability to immediately verify whether a repair has been successful. Rather than waiting for the next scheduled test or experiencing continued problems, operators can see in real-time whether the corrective action has addressed the root cause.

In this case, the results were clear and dramatic.

Post-Repair Performance

Following the completion of the maintenance work, HydroTrac continued to monitor and trend the partial discharge levels on all three phases. The trending data for Phase C showed a significant reduction in PD activity (Figure 2).

_{Figure 2: Trend of PD activity – Phase C}

The trend plot spans from April 2005 through January 2006 and clearly illustrates the impact of the repair. Before the maintenance intervention in October 2005, Phase C exhibited PD levels exceeding 2,500 mV. After the cleaning and recoating of the ten bars, discharge levels dropped below 500 mV—a reduction of more than 80%.

Importantly, the monitoring data showed that this improvement was sustained over time. The PD levels remained low and stable through January 2006 and beyond, confirming that the repair had successfully addressed the underlying problem rather than providing only temporary relief.

The Value Proposition

This case study demonstrates several important benefits that continuous partial discharge monitoring delivers for generator fleet management:

Early Problem Detection

The monitoring system identified an insulation problem while the generator was still operating normally. There was no indication of failure, no alarm, no trip—just gradually developing discharge activity that the continuous monitoring detected in its early stages. This early warning provided time to plan and schedule maintenance during a convenient outage rather than responding to an emergency.

Diagnostic Precision

The monitoring data didn't just indicate that "something is wrong somewhere." It provided specific information about where the problem was located (Phase C, end-turn region, near specific sensors) and what type of discharge pattern was occurring. This level of diagnostic detail enabled targeted repairs rather than extensive investigation or wholesale replacement.

Efficient Use of Maintenance Resources

Because the monitoring data pinpointed the problem area, maintenance personnel could focus their efforts on ten specific bars rather than inspecting the entire winding. This efficiency translates directly to reduced outage time and lower maintenance costs.

Objective Repair Verification

The post-repair trending data provided quantitative, objective evidence that the maintenance work had been successful. Rather than relying on visual inspection or subjective assessment, the same monitoring system that detected the problem confirmed that it had been resolved. This verification gives operators confidence in returning the machine to service.

Long-Term Asset Management

Beyond the immediate case, continuous monitoring provides valuable information for long-term asset management decisions. The data helps utilities understand which machines are experiencing problems, what types of problems are occurring, and how quickly conditions are changing. This information supports better planning for future maintenance and eventual replacement.

Broader Context: Partial Discharge Monitoring Technology

The technology demonstrated in this case has become increasingly common in generator condition monitoring programs. Understanding how it works helps explain why it's so effective.

What Is Partial Discharge?

Partial discharge is a localized electrical discharge that only partially bridges the insulation between conductors. Unlike a complete breakdown (which would cause immediate failure), partial discharges are small events that occur repeatedly within voids, along surfaces, or at other weak points in the insulation system.

While individual discharge events may not immediately damage the insulation, over time they can contribute to degradation. More importantly, the presence of partial discharge serves as an indicator that something is wrong with the insulation system—there are voids, contamination, deterioration, or other defects that shouldn't be there.

Why Monitor Online?

Partial discharge can be measured either offline (with the generator shut down and test voltage applied) or online (during normal operation). Each approach has advantages, but online monitoring offers several unique benefits:

• Real operating conditions: The generator experiences actual operating voltage, temperature, and mechanical forces
• Continuous trending: Changes can be detected immediately rather than waiting months or years between tests
• No interruption: Testing occurs without taking the generator out of service
• Operating history: The system builds a complete history of discharge behavior under various operating conditions

Application to Hydro Generators

Hydro generators present both opportunities and challenges for partial discharge monitoring. On one hand, these machines typically operate at relatively moderate voltages (compared to large turbine generators), which can make discharge detection easier. On the other hand, the machines are often quite old, may experience varying operating conditions, and can develop a variety of insulation problems.

End-winding problems, like those detected in this case, are particularly common in hydro generators because:

• The end-windings are exposed to the internal atmosphere
• Airflow patterns can deposit contaminants on insulation surfaces
• Temperature and humidity variations can affect surface conditions
• The machines may operate for decades, allowing gradual buildup of contamination

Lessons for Generator Fleet Management

This BC Hydro case provides several takeaways for utilities managing similar assets:

• Invest in Monitoring Infrastructure: The relatively modest investment in PDA couplers and monitoring instrumentation paid dividends by detecting a problem early and enabling targeted repairs.
• Act on the Data: Having monitoring technology is only valuable if the data is reviewed, analyzed, and acted upon. BC Hydro's team clearly had processes in place to review trending data and respond appropriately.
• Document and Share Results: By documenting this case (through publication at IRMC 2006), BC Hydro contributed to industry knowledge and helped other utilities learn from their experience.
• Consider Fleet-Wide Deployment: The success of this machine suggests value in deploying similar monitoring on other generators in the fleet, particularly aging units.

Conclusion

For BC Hydro, continuous partial discharge monitoring on this 37-year-old hydro generator successfully detected end-winding discharge activity that required attention. The monitoring data enabled targeted maintenance—cleaning and recoating ten stator bars—that resulted in a significant and sustained reduction in discharge levels.

The case illustrates the practical value of online condition monitoring for rotating machines. By providing continuous visibility into insulation condition, early warning of developing problems, diagnostic information to guide repairs, and objective verification of maintenance effectiveness, this technology helps utilities maximize the reliability and lifespan of critical generation assets.

As generator fleets continue to age and the intervals between major overhauls extend, condition-based maintenance approaches supported by continuous monitoring will become increasingly important for maintaining fleet reliability while managing costs.

Reference:

S. Li and J. M. Y. Chow, "Partial Discharge Measurements on Hydro Generator Stator Windings Case Studies," IEEE Electrical Insulation Magazine, vol. 23, no. 3, pp. 5-15, May-June 2007, doi: 10.1109/MEI.2007.369456