This project will focus on developing a full-scale demonstrator retro-filled with a suitable SF6 alternative with condition-based monitoring systems incorporated to perform non-invasively, in-situ monitoring during the long-term energisation. The project will develop an optical test setup which addresses the missing link between long-term safe operation of equipment and traditional academic material testing and characterisation. The project will focus on the applicability of such techniques and critically assess their suitability to give asset managers the information required for retro-filling an SF6 alternative on the transmission network.
Benefits
The benefit of retro-filling assets, where this is possible, is to achieve environmental ambitions to reduce SF6 inventory and therefore the risk of very high emissions of gas with the highest known GWP. This may be achieved through asset replacement but retro-filling represents a more economic and environmentally sound method of doing so as the investment cost is lower and the requirement for raw materials is lower.
Learnings
Outcomes
Partial discharge monitoring is a common tool for assessing equipment condition. Partial discharge measurements for a gas that could be used for retro-fill purposes and compared with SF6 in scenarios with an introduced defect suggest that the same detection equipment could be used without modification whether the detection method is based on IEC 60270 apparent charge or UHF monitoring. Both PRPD and general PD trends are comparable for the two gases.
Understanding the effects of surface roughness on the impulse ratio for SF6 alternatives is important in selecting the gas mixture to be used for retro-fill purposes since it can lead to flashovers. This project is identifying the risk of surface roughness to breakdown as a result of retro-filling with an SF6 alternative.
Recommendations for further work
Further work investigating the impact of crystals formed in C4F7N gas mixtures has started in NIA2_NGET0046 - CrystalClear - Lifecycle Analysis of SF6 Alternative Technologies and Crystal Formation Impacts.
Lessons Learnt
It has been shown that crystal by-products can be produced in conditions that approximate those found in service although with high moisture levels in the chamber. It has been found very difficult to generate enough crystals to enable evaluation of their significance in electric fields. Mass spectrometry was found to be suitable in identifying the crystalline compounds, particularly using soft ionisation techniques. No by-products were generated under dry conditions, unrealistic levels of moisture (for service conditions) were required to produce compounds of note.
The partial discharge characteristics of a 20% mixture of C4F7N in carbon dioxide in a GIS demonstrator with needle defects of different lengths was compared with SF6 in the same equipment. The phase resolved partial discharge (PRPD) patterns are comparable using either of the two detection methods used (UHF and coupled capacitor), thus either detection method can be deemed suitable.
A significant difference was noted in the time to breakdown of the different gases with the same defect type. This suggests the critical defect length is gas dependent and this may vary for different gas mixtures.
Looking again at the 20% gas mixture and SF6, this time under breakdown conditions, observations have been made comparing the two dielectrics under pressure, with particular attention to surface roughness. The following was noted:
- Under AC breakdown conditions, the breakdown voltage increases with pressure. There is more divergence between the values for each gas as the pressure increases. This is an effect of surface roughness.
- The uniformity of the electric field around the electrodes has a significant effect on the breakdown mechanism of the gases.
- The impulse ratio (ratio of peak LI and AC withstand voltage), which is an engineering value that helps with design margin consideration for gas insulated systems, differed for the two gases. The C4F7N mixture was below that of SF6 which maintains the limit at all pressures tested. With increasing surface roughness, the ratio trend with pressure is inconsistent. The variation is largely due to the LI withstand voltage for the C4F7N mixture.
Further testing of gas mixtures has shown that mixtures of C4F7N with nitrogen have higher LI breakdown voltages than equivalent mixtures with carbon dioxide, this for both positive and negative polarities.
The test setup was adjusted to use a longer insulator during breakdown experiments to reduce the risk of flashover and insulator movement. This improved the repeatability of the breakdown testing.
2024/2025 update
When SF6 breaks down under electrical testing, PD clusters may be seen just prior to break down, this is not observed with the retro-fill gas and most likely relates to the longer energisation time in SF6. This energisation time also appears to cause sulphite deposition on the needle tip placed to create the defect. This was not seen with the retro-fill gas, but the needle tip was noted to erode during repeated tests.
Increasing the needle tip radius was expected to provide a more uniform electrical field and thereby create a longer time to breakdown. In practice the largest tip radius (100 micron) reached breakdown first with the smallest (20 micron) taking the longest.
The decomposition pathways of SF6 were modelled using computer aided chemistry techniques and validated by comparison with literature results for experimental data using packed-bed plasma reactors. Having validated the modelling approach the techniques was then applied to C4F7N. The proposed byproducts included C3F5N, C3F8, CF4, C2F6, C4F10, C6F14, CO, N2, C2F3N, COF2 and C2N2, some of which have been reported in previous work. The simulation of by-products was found to become more difficult with the addition of water in the starting scenario. Some rate constants for some reactions are unavailable adding to the uncertainty in decomposition pathways.
Dissemination
A dissemination event was held at the University of Manchester in December 2022 to which representatives of Transmission Operators and Distribution Network Operators were invited. It was attended by representatives from Eirgrid, ENWL, ESB, NIE, SPEN, SSEN and UKPN. The progress and aims of this project were presented along with other projects on SF6 and SF6 alternatives.
A second event was held in September 2024 to which all UK licensees were invited, at the Graphene Institute in Manchester and others were invited to present. It was attended by representatives of Dilo, ENA, ENWL, ESB, GE Verona, Hitachi Energy, Institute of Science and Technology (Austria), Monitra, NIE, NPG, RTE, Siemens Energy, SINTEF Energy Research, SPEN, SSEN-Transmission, SYSTRA, UKPN, Wika. Presentations were also given by GE Verona, Hitachi Energy, RTE, Siemens Energy and SINTEF Energy Research.
A paper “Production, Analysis and Identification of Crystal By-products in C3F7CN Mixtures” was presented at the CIGRE Symposium in Cairns, Australia in September 2023.
A paper “Reducing the Global Warming Potential in Gas Insulated Lines and Busbars by replacing SF6” was presented by Gordon Wilson at the International Conference of Doble Clients in Boston in March 2025 which included work from this project and its T1 predecessor.
A paper “Characterization of C3F7CN Gas Mixtures and their Trace Decomposition Products Detection using GC and Photoionization Mass Spectrometry for Electrical Insulation Applications” has been submitted to Analytica Chimica Acta.