Learnings

Outcomes

Extensive published data on the dielectric properties of selected gases (g3 gas, some atmospheric gases and their mixtures) were collected. These highlight the key dielectric properties of the gases and how they compare between themselves and against SF6.

The experimental test rigs were prepared to test the gases for their dielectric withstand and by-products of arcing in the gases. CO2 and N2 were characterised in this work for calibration purposes.

All experimental work with Novec gas and its g3 mixture was completed and reported. Key findings are as follows:

• For a 4% Novec™ 4710/96% CO2 mixture, the equivalence with SF6 at 5.5, 6.5, 7.5 Bar(a) (absolute bar or x105 Pa) is feasible for a minimum operating temperature of −30 °C, −25 °C, and −20 °C, while providing a significant reduction of the environmental impact (≈98% less than SF6).
• The breakdown characteristics of 4% Novec™4710/96% CO2 and CO2 as a function of pressure under non-uniform electric field distribution is strongly nonlinear. At 8.8 bar(a), the positive inception voltage of 4%Novec™4710/96% CO2 is 1.18 times higher relative to CO2 and 1.13 times higher for the breakdown voltage.
• The comparison of flashover voltage for different materials was found to be as follows:

V Flashover PTFE > V Flashover PE−UHMW > V Flashover Epoxy > V Flashover HDPE

• The following by-products were detected by GC/MS following more than 400 Lightning Impulse breakdowns on the 4% Novec™/96%CO2 mixture:
  • Carbon Monoxide (CO)
  • Carbon Tetrafluoride (CF4)
  • Tetrafluoroethylene (C2F4)
  • Hexafluorethane (C2F6)
  • Hexafluoropropylene (C3F6)
  • Octafluoropropane (C3F8)
  • Perfluorobutane (C4F10)
  • Perfluorohexane (C6F14)
  • Trifluoroacetonitrile (CF3CN)
  • Trifluoroacrylonitrile (C2F3CN)
  • Pentafluoropropionitrile (C2F5CN)
  • Oxalonitrile (C2N2)
  • Hydrogen cyanide (HCN)

Many of these were found to increase in concentration with increasing number of breakdowns, CO was found to saturate during the studies.  Oxalonitrile was only detected in the final sample (after 408 breakdowns).  

• In addition to the gases listed above, under PD activity (positive cycle) for 80 hours, additional by-products to those listed above were found: Trifluoropropynenitrile (C3FN) and three others that could not be identified by GC-MS or FTIR.  After these tests the concentration of the Novec™4710 was found to have reduced by 43% and CO reached a concentration of 20% of the overall mixture.
• The toxicity effect of many of the by-products from electrical breakdown and long-term partial discharge studies on the 4% Novec™/96%CO2 is not known but some are known to be very toxic.
• The concentration of Novec™4710 at different heights in a vessel at a pressure of 4.8 Bar(a) was measured as a function of time to evaluate whether stratification can occur.  The stabilisation time for the diffusion of the Novec gas into the carbon dioxide was found to be around 300 hours at room temperature (~17 °C).  After that time, no stratification was observed and would not be expected to occur in service.
• Lifetime estimation based on electrical and thermal considerations were undertaken.  With a tolerance of reduction in Novec™4710 from 4 % to 3.5 % thermal and modest partial discharge (< 1 pC apparent discharge) the lifetime would exceed the expected lifetime of the GIS equipment.  The presence of particles sufficient to consider the gas mixture to be “moderately polluted” would reduce the life of the gas to 11 years or less depending on the nature of the particles.

Recommendations for further work

A similar protocol could be applied to other potential gas mixtures proposed as lower GWP alternatives to SF6.  A project to evaluate mixtures of gases using Novec™5110 is already under way – NIA_NGTO051

Lessons Learnt

The deliverables for this project were revised to allow for the delay in the procurement of gas. It was initially unclear how much gas would be required for the testing. As a large quantity was required, its procurement took longer than initially anticipated.

Published data on the dielectric strength of gases has been collected and analysed. Tests on CO2 and N2 gases under different field distributions and pressures have been performed using specially made test pressure vessels.

• The tests carried out have shown that under negative polarity of lightning impulse, both CO2 and N2 show lower dielectric strength than g3 gas.
• Under positive polarity, the dielectric strength of these gases improves slightly and is closer to that of g3.

These test environments allowed sampling of the gases after many arcing events through the gas to determine the gas by-products that are generated because of the breakdown events. The following was found using the gas analysis apparatus:

• Carbon monoxide was the main by-product of the decomposed CO2.
• Arced N2 shows no new components in the analysed gaseous medium.

A survey of research findings on g3 toxicity has revealed that it is important to use safety equipment and clothing to handle the gas and its mixtures. This is an important aspect for all alternative gases, as it requires careful consideration and possible independent determination/verification of the toxicity levels.

For the g3 gas mixture, the by-products of breakdown and partial discharges were quantified. The toxicity levels of the gas and its by-products have been summarised to enable procedures to be developed or updated and ensure the safety of operating staff. The impact and interaction of the gas with spacer insulating materials has been quantified.

Dissemination

The findings from the project to date have been discussed at the Cardiff Materials Research Network (17th – 18th January 2019) and at CIGRE WG D1.67 (Dielectric performance of new non-SF6 gases and gas mixtures for gas insulated systems) in Berlin (March 2019).

An open-access Journal paper “AC Volume Breakdown and Surface Flashover of a 4% NovecTM 4710/96% CO2 Gas Mixture Compared to CO2 in Highly Nonhomogeneous Fields” has been published in Energies 2020, 13(7), 1710; it is available at https://doi.org/10.3390/en13071710.

A virtual dissemination event providing a summary of the project is planned for 2021