Learnings

Outcomes

Wireless phase reference (WP3) 

Calculating True PF requires that the hardware precisely triggers the voltage and current acquisition at the same time to accurately determine the phase difference between the voltage and current along with respective magnitudes.  The Bushing Data Monitor (BDM) is the hardware responsible for acquisition of current and voltage in the monitoring device used for this project.   

With a regular BDM setup, the bushing and voltage cards are peripherals of the same device. It is therefore possible to guarantee that the device can start the acquisitions simultaneously.  In this project two methods of using a GPS source were used to timestamp data on the separate devices and ensure that any phased shift that is introduced is minimal.  It has been demonstrated that the devices measuring current and voltage can use separate GPS signals and be connected via WiFi to the BDM and calculate TPF. 

Analysis of bushing monitoring data (WP4) 

Trending characteristics of bushing monitoring data was preserved when relocating the bushing monitors between two sites was preserved. 

Relative PF and loss angle measurements were completed at all three sites successfully.  Variations over time affected both measurements continuously, this is true when observed on an instantaneous basis and when using time-averaged data.  This presents a challenge when setting alarm levels to indicate deteriorating condition.  It is necessary to establish normal behaviour for bushings based on the individual transformers.  Since the variations occur on all three phases at the same time, it is assumed the causal factors are external to the bushings. 

In the case where it was possible to observe both relative and true PF, the data were generally consistent across the period of study.  During the period, a change was observed in the loss angle and true PF that was not seen in the relative PF.  The cause was likely external to the bushing and shows that true and relative PF are likely complementary rather than duplicative. 

Total harmonic content (THD) also observed the change at the site where loss angle and true PF changed due to external factors.  The monitoring has potential to pick up changes in loads on the local network.  THD varied over time, with an average change of 2-3% on all three sites, but the average values were different at three different sites, likely indicating local loading factors – industrial or residential loads.  THD was unaffected by ambient conditions or tapping on the transformer, both of which were as expected. 

The tap-changer was expected to have an effect on the loss angle and PF measurements since a change in tap alters the winding characteristics, albeit to a limited extent.  In practice, there were no discernible variations in phase angle and relative PF or loss angle and true PF correlated with changes in the tap position. This was true whether looking at the ‘instantaneous’ values or data smoothed using a rolling average, over periods from days to more than a year.  A number of explanations were considered, including that the change in tap is too small to be detected because of the transformer impedance, or because of the level of sensitivity in the monitor.  Further work would be required to understand the cause but from a practical perspective, tapping may be ignored in understanding bushing condition through online monitoring. 

No anomalies in the monitored bushings were observed using thermography.  However, the entry-level fixed infrared (IR) system that was installed on one site allowed for some observations.  The load on one of the transformers was relatively constant so it was possible to observe seasonal variations in temperature.  In combination with other ambient conditions, leakage current data and load, it was possible to observe the thermal time constants in the bushings and to identify spikes in temperature as solar gain rather than reflections.  There is likely potential for automation of detection of high resistance connections (“hot joints”) and low oil levels using high-end IR cameras. 

Partial discharge monitoring was studied using the peak to average power (PAPR) method.  None of the bushings studied demonstrated PD trends of concern over the period.  Sporadically observed events were seen on all three phases simultaneously and believed to be external to the bushings, either inside the transformer tank or outside the transformer entirely.  Changes in load resulted in changes in activity on two of the transformers. 

Tap-changer position had no impact on observable PD as seen in the PAPR plots.  Neither windspeed nor air pressure were observed to have an effect on the PAPR plots, in line with expectations.  However, on one site, peaks in PAPR plots correlated with air temperature rising above 20 °C and troughs were noted to be aligned with drops in relative humidity (calculated from air temperature and dew point). 

Following a review of likely bushing failure modes and the timescales under which they operate, in conjunction with the likely indicators of deterioration, it has been demonstrated that it is possible to develop an asset health index for bushings with solid insulation. 

At the outset of the project the technology readiness level was adjudged to be 4 – early development.  The target was to reach TRL 8 – a point at which it at least some of the work is ready to be deployed.  Although much has been learned during the project, the research results are yet to be fully evaluated by internal subject matter experts and at this stage TRL 7 is a more likely position.  However, the outcomes suggest that condition-based replacement is likely possible for resin-impregnated bushings. 

Recommendations for further work 

Further work could allow the wireless voltage reference to be a product for future use in monitoring assets rather than a practical demonstration of a theory. 

Further investigation of the variations in bushing measurements over time could determine the cause and enable compensation of the effects in establish a condition monitoring strategy.  This could be augmented with machine learning techniques to allow automated analysis of values to reduce the risk of false alarms. 

With further analysis and study, it may be possible for harmonic factors to be considered in transformer thermal rating calculations and ageing models (as they increase losses).  In addition, exclusion of harmonic content could be excluded as a cause of spurious trips or a factor in equipment failures at NGET grid supply points.  

Lessons Learnt

This project has focussed on 120HC/300HA Trench bushings.  In the course of this project it has been identified that there are two variations of bushing tap connectors.  Any future work connecting to this type of bushing should note the variant required for connecting equipment. 

Where weather stations are deployed on sites that are required to communicate with other equipment over a significant distance (~ 100 m) then Wi-Fi repeaters have been found to provide a suitable solution. 

Dissemination  

A paper titled “Online Power Factor Measurements of 132 kV Bushings” based on the outcomes of this project has been submitted to IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP), Manchester, UK, September 2025. 

A paper titled “Impact of environmental and loading conditions on Online Power Factor and Partial Discharge Measurements of 132 kV Bushings” based on the outcomes of this project has been submitted to CIGRE GCC conference, Kuwait, November 2025. 

A paper titled “Challenges in Asset Management: Data, Decisions & Uncertainty” including a case study from this project, has been submitted for CIGRE 2026, Paris Session.