ITC HOLDINGS CORP. INVESTS IN ELECTRICITY TRANSMISSION INFRASTRUCTURE IMPROVEMENTS as a means to improve electric reliability, reduce grid congestion and lower the overall cost of delivered energy. Prior to ITC Holdings Corp. (Novi, Michigan, U.S.) assuming ownership of the transmission systems in the state of Michigan, a series of catastrophic failures occurred involving transmission system transformers. As a part of ITC's investment in the grid, the company initiated a transformer monitoring project to allow real-time monitoring to mitigate imminent transformer failures.


ITC Holdings selected an intelligent diagnostic device (IDD) designed by Doble Engineering (Boston, Massachusetts, U.S.) to continuously monitor 138-kV class bushings and above. These on-line monitoring devices are used on more than 80 transmission transformers in Michigan.

The IDD, one component of the GE Energy (Atlanta, Georgia, U.S.) FARADAY tMedic transformer monitoring system, continuously monitors bushing leakage current by means of the Doble Expert System Diagnostics. The Web-based IDD contains a microprocessor-based signal-processing module, a control panel for user interface, alert indications, and either an RS-232 serial port or Ethernet for local PC access, as well as options for remote communication. The status of each bushing set is analyzed to detect abnormalities in insulation by calculating capacitance and percent power factor (%PF) via IDD sensors installed in place of each bushing's potential tap cap. The system issues appropriate alerts when a problem is detected. Measurements are relative as opposed to absolute because of system voltage variations. Alerts and system faults are managed either from the Web browser interface or directly on the IDD front panel.


On July 10, 2008, the engineering department received an INFO alert from the IDD monitoring the transformer's bushings on a critical interconnection 345/120-kV, 370-MVA autotransformer. The operator reported a CODE 12 from the readout on the IDD. This indicated detection of a combination of a measurement problem and low %PF deterioration. An INFO alert is the first of a three-tier alarm system. The alarm levels consist of INFO, WARNING and ACTION alerts. The engineering department decided that it would download and review the data the next morning if the level did not increase to WARNING.

The indication of such a high %PF came as a surprise because extensive off-line transformer maintenance and testing had been done earlier in March. Furthermore, the maintenance included replacing the H2 and N bushings because of a high power factor. At that time, the H1 and H3 bushings tested at or near nameplate values and did not require attention. ITC had experienced challenges with the IDD in the form of false alerts in the system, but those have been resolved through firmware updates. However, bushing alerts of any level can never be ignored.


On the morning of July 11, the IDD was accessed and all data was downloaded for review. Initially, the alert description and feature data was reviewed, noting that the EVIDENCE on the Web browser alert description window and the reduced phase angle from 240 degrees correlated. The phase angles should be 120 degrees from the reference H2 (0-degree) bushing.

A history of data from the IDD was downloaded, reviewed locally and sent to Doble for analysis. According to the data, the %PF on the H3 bushing had been increasing steadily since April 28, 2008, the date of the last IDD configuration.

Doble quickly responded to ITC's concern of the possible deterioration of the H3 bushing. Doble noted that the phase angle of the bushing had decreased by one degree over the last two-and-a-half months, resulting in an increase of power factor. A slight increase in H3 capacitance was also noticed, indicating the bushing was starting to degrade.

As a general rule of thumb, one-degree difference from normal equals approximately 1% in power factor. The H1 bushing phase angle was relatively constant.

The current magnitude of the H1, H2 and H3 bushings was also examined. Over the period of record available, there was a slight increase in the magnitude of the H3 bushing current, which translates to an increase of capacitance. H2 showed a higher current magnitude than the other two phases; however, this is normal because the bushing nameplate capacitance is higher.


ITC agreed with Doble's recommendation to take the unit off-line to power factor test the H3 bushing. ITC engineering determined that, because of system load, an outage would be planned when conditions allowed, unless a WARNING alert came in prior to the scheduled outage. A WARNING alert indicates the monthly %PF has reached 2%. In that case, the unit would have been taken off-line immediately. The outage was scheduled for July 21.

From July 11 to July 21, IDD data downloads were reviewed every two days to confirm bushing %PF was stable. During this period, ITC was able to plan the upcoming work, which included arranging for oil-handling equipment, tankers, spare bushings, crane and other miscellaneous equipment to be on site as well as personnel.

The outage for the transformer began on July 21. The off-line tests results (Table 1) indeed indicated an increase in %PF on the H3 bushing — a 1.36%PF value versus a 0.26%PF nameplate value. The absolute or off-line %PF test of the H3 bushing compared to the IDD downloaded data confirmed the increase in %PF. A Tip-Up test, which shows an increase in %PF with an increase in voltage, was also done. For a bushing in good condition, the %PF should not change with respect to voltage. The H3 bushing was deteriorating, and ITC decided to replace it as well as H1. Work to replace the two high-side bushings began immediately, and the transformer was returned to service on July 27, six days later.


Once the transformer had been returned to service, ITC wanted to perform a more-thorough investigation on the degraded bushing to confirm the results of the on-line monitoring device. This was the first occasion that an alert led to actual downtime on a transformer that required maintenance. A decision was made to contact Doble Engineering and Hubbell Power Systems' PCORE Electric to assist in this investigation.

Both the H1 and H3 bushings were sent to PCORE Electric for remanufacturing, where the bushing would be completely torn down and rebuilt. Prior to tearing down the bushings, a series of electrical tests were conducted; tests included C1 and C2 capacitance, power factor, partial-discharge measurements (Table 2) as well as gas-in-oil samples. The abrupt increase of the partial-discharge noise at 143 kV was a certain indication of the near-failure condition of the bushing

The oil samples were sent to Doble for analysis. The results of the oil analysis (Table 3) for the H3 bushing had high levels of hydrogen, indicating partial-discharge activity, and small amounts of acetylene, a result of arcing. Partial-discharge measurements and visual electrical treeing on the paper samples all confirmed the bushing was experiencing partial discharge.

ITC Holdings believes that it was imminent this bushing would have failed in service, causing extensive damage to the transformer and its proximity. Having the IDD continuously monitoring the bushings allows for early detection of a possible abnormality, provides a risk-management tool and facilitates the planning of maintenance to be done on an as-needed basis.

Table 1. Test Data Results for Doble M4000 Test Set

Test conditions Test date: 7/21/2008 Air temperature: 35°C Test time 10:47:03 a.m. Tank temperature 50°C Weather: Partly cloudy Relative humidity: 39%
ID Serial Percent PF Capacitance Test kV mA Watts Percent PF correction Correction factor Capacitance (pF)
Bushing C1
H1 1797916 0.26 406 10.015 1.511 0.036 0.22 0.93 400.78
H2 05-105312 0.27 492 10.018 1.888 0.047 0.25 1.00 500.74
H3 1796658 0.26 401 10.02 1.502 0.219 1.36 0.93 398.47
Bushing C2
H1 1797916 0.235 5317 3.001 20.147 0.429 0.21 1.00 5344.1
H2 05-105312 0.250 6081 3.001 22.877 0.506 0.22 1.00 6068.2
H3 1796658 0.305 4973 3.001 18.768 0.725 0.39 1.00 4978.2

Table 2. Factory C1 Partial-Discharge Tests (note 2 pC to 3 pC is ambient noise)

Testvoltagein kV Partial-discharge noise in picoColumbs (pC)
H3, S/N 1796658 H1, S/N 1797916
10 2 to 3 2 to 3
50 2 to 3 2 to 3
100 2 to 3 2 to 3
143 60 to 70 2 to 3
150 70 to 90 2 to 3
200 150 to 200 2 to 3
220 - 2 to 3
330 - 2 to 3

Table 3. Oil Quality Results

Dissolved gas analysis results
Gas Chemical formula Measured levels (PPM) IEC 61464 limits (PPM)
Hydrogen H2 6020 140
Oxygen O2 12600
Nitrogen N2 52000
Methane CH4 1280 40
Carbon monoxide CO 229 1000
Ethane C2H6 570 70
Carbon dioxide CO2 783 3400
Ethylene C2H4 3.1 30
Acetylene C2H2 0.9 2

Richard F. Wancour ( has employed at ITC Holdings Corp. since January 2007 and currently works in the asset management department as a senior engineer. Prior to his present responsibility, he was employed by another utility as a substation design engineer for more than six years. Wancour holds BSEE and MSEE degrees from Wayne State University and is a licensed professional engineer in the state of Michigan.

Stephen Molter ( has been senior engineering technician with ITC Holdings since 2004. Prior to joining ITC in 2004, Molter worked at a southern utility in substations for 27 years, and as a consultant to industrials and various substation maintenance service groups for six years.