Ester-based insulation fluids offer several benefits for the operation of transformers. They have improved fire safety, reduce the risk of moisture build-up in paper insulation, they are biodegradable, and some are from sustainable sources. They also offer the possibility of improved thermal performance, due to their typically lower viscosity. However, such liquids are typically more expensive than mineral oil, and have only been used in high voltage power transformers relatively recently, so operational experience with them is limited. This project will work to improving the understanding of the impact of alternative liquids on transformers through simulations and experimental investigations. Particular attention will be given to the thermal performance of such liquids, and the possibility of improved ratings.
Benefits
This research project is intended to deliver benefits through the potential for transformers filled with alternative fluids to have increased capacity compared with mineral oil filled equivalents. All of the alternative fluids have environmental benefits in that they are biodegradable, but, in addition, the fluids based on vegetable oils are more sustainable than those relying on mineral oil sources. The project will also explore the potential for the retardation of natural ageing of solid insulation through replacement of the fluid during its lifetime.
Learnings
Outcomes
WS1: State of the Art Review and Analysis
In summary, there is a wide body of literature, standards and patents existent for ester oils with a large body of laboratory ageing studies. The main shortcoming of the literature is a lack of data from ester filled transformers which have been aged in the field. A key advantage of ester fluids is their biodegradability and reduced environmental impact compared to mineral oil. Laboratory ageing studies show that using an ester oil in place of mineral oil significantly reduces the ageing rate of the included kraft paper which can extend plant lifetime. Synthetic ester is better for use in plant due to its improved resistance to oxidation, however it is less biodegradable than natural ester oil.
WS2: Modelling of thermal conditions in operational transformers that use alternative liquids
Ester oil transformers demonstrated marginally superior performance, possibly linked to the characteristics of the oil time constant.
Issues with the way that heat run test data are reported for transformers may be overcome through more detailed analysis but taken at face value could lead to errors in the calculation of their rating enhancements.
WS3: Identification of Natural Ester Liquid degradation processes
The results of compatibility testing have been reported for different fluids and found to be acceptable in most cases. Ageing studies of cellulosic insulation in different fluids has confirmed that the bio-based insulating liquid under study performs at least as well as mineral oil. All the aged oils, including the natural ester aged for < 10 days, appear to meet the relevant IEC acceptance limits for service aged transformer oils. The bio-based insulating liquid could potentially improve cooling efficiency in a high voltage transformer whilst both ester oils can mitigate the aging of paper, potentially extending plant lifetime.
The gelling seen in food-grade vegetable oils is not replicated to the same degree in esters based on the same feedstocks, this would appear to be the influence of additives since the same improvement was seen upon addition of an antioxidant to the food-grade oils. Other additives were explored but nothing was found that was able to prevent gelling completely under harsh oxidative conditions. Therefore, the risk of natural esters gelling in a transformer in transit remains.
WS4: Development of thermal network models for ‘Business as Usual’ activities
All loading-guide models apply similarly to transformers with ester oils and mineral oil, but default parameters for mineral oil are not valid for ester oils; transformer-specific parameters are necessary.
The oil time constant is critical as it is the only time constant significantly affecting the 6-hour rating and accurate estimates are essential for reliable loading-guide models.
Obtaining a reliable oil time constant is challenging because standard tests often maintain cooling until near steady-state, obscuring accurate measurement. “Speed-up” methods during testing can be used but may invalidate the results. An alternative is to calculate the time constant from the effective thermal capacity of the transformer, which should be documented in test reports and is preferable to extending factory tests which incur significant cost.
WS5: Performance and degradation of alternative insulation fluids
Modelling studies of the OD (forced cooling) mode shows that changes in viscosity of different fluids as they age has a measurable but insignificant effect on rating.
Of the esters that were investigated, one supplier’s esters, both natural and synthetic, did not meet requirements under ageing conditions, emphasising the need to carry out testing on new products to confirm they meet specifications.
All of the fluids under test were found to be well within performance specifications under electrical testing (57,000 discharges).
WS6: Recommendations for asset management, ratings & operations
National Grid’s existing TRALC model for transformer ratings is already fit for purpose for modelling transformers filled with alternative fluids as long as suitable input parameters are available. A new thermal modelling tool is not required to carry out ratings in the same way that they are currently performed. A longer oil time constant would provide a more accurate model and would give better ratings for 6-hour periods than using mineral oil inputs.
A new rating model was provided that can be run in COMSOL. In comparison with existing rating models the following observations were made to demonstrate its advantages:
IEC 60354: A simple model with winding and oil time constants but tends to overestimate emergency ratings.
TRALC2 (original): Adjusted from IEC 60354 to account for tap position and winding loss temperature dependence but exhibits non-physical heat transfer when cooling status changes.
IEC 60076-7: Contains significant issues:
- Incorrect and confusing parameter definitions and explanations (e.g., for overshoot phenomena).
- Structural problems and reliance on unrealistic assumptions, leading to inaccuracies.
IEEE C57.91 Annex G: Closer to a physics-based model but:
- Requires extensive input data (e.g., transformer masses).
- Uses an overly simplified method for redistributing heat between top-oil and bottom-oil temperatures, leading to potential inaccuracies in transient responses.
Recommendations for further work
IEC 60076-7 “Power Transformer – Part 7: Loading guide for mineral-oil-immersed power transformers” could be improved in the way that it handles temperature dynamics in transformers. Test data for transformers subjected to heat run testing were found to contain discrepancies. There may be improvements that could be made through standardisation of reporting.
The TRALC model would benefit from a more dynamic approach to oil and winding time constants. It is proposed that a revised TRALC model would use winding rise over bottom oil temperature as a steady state variable, ensuring heat conservation during cooling transitions and being a more physically accurate representation of the situation.
There is a lack of up-to-date international guidance for synthetic esters in service. A new guide is being prepared by IEC but not currently available. NGET should note the guidance when published to reflect its requirements into business-as-usual maintenance activities.
Lessons Learnt
The testing carried out to date has identified that compatibility of some fluids with nitrile rubber requires further investigation.
Oil and winding time constants were found to be absent from the test certificates for ester filled transformers. A Python script was utilised to extract data from temperature rise plots in test certificates and estimate the values needed. This could be applied to future test certificates for the same purpose. Python was also used to improve image processing to extract data tables from pdf copies of test certificates and significantly speed up the process.
IEC 60076-7 “Power Transformer – Part 7: Loading guide for mineral-oil-immersed power transformers” could be improved in the way that it handles temperature dynamics in transformers. Test data for transformers subjected to heat run testing were found to contain discrepancies. There may be improvements that could be made through standardisation of reporting.
SFC-MS has been shown to be a suitable method for the analysis of natural and synthetic esters, especially for monitoring their ageing. It is more effective than GC-MS because of the high molecular weight of the materials studied.
Dissemination
A paper “A comparative aging study of various commercial biodegradable insulation systems” was presented at the 23rd International Symposium on High Voltage Engineering in August 2023.
Two papers were published at the IEEE 5th International Conference on Dielectrics (ICD) in July 2024. The papers were entitled “A Study of Ageing and Gelling in Natural Ester oils” and “Supercritical Fluid Chromatography-Mass Spectrometry (SFC-MS) Analysis of Natural Esters on Aging”.
Publications in progress at the time of writing:
A paper entitled “Electrical ageing of insulation oils” is expected to be presented at the Conference on Electrical Insulation and Dielectric Phenomena (CEIDP) in Manchester in September 2025
Two journal papers in preparation - “A comparison of the ageing behaviour of 8 ester oils for high voltage transformer applications” and “An Investigation of the chemical degradation of natural ester electrical insulation fluids using supercritical fluid chromatography - mass spectrometry (SFC-MS)” will be submitted during 2025 to suitable journals.
A review of progress on this project was shared at a research dissemination event held in conjunction with the annual CIGRE UK liaison meeting for A2 (Transformers) and D1 (Materials) in Manchester in December 2024.
