Zoe Lewis
Gather.town id
MIS04
Poster Title
Looking for space weather impacts on UK power station transformers
Institution
Lancaster University
Abstract (short summary)
It is well documented that space weather may impact electricity infrastructure, and several widespread blackouts have been observed in the past few decades and directly linked to the largest geomagnetic storms (e.g. the Hydro Quebec incident in 1989). However, less is known about the impact of lower-level GICs on the health of transformers in the long term. In this study, the long term impact of geomagnetic activity on 13 power station transformers in the UK was investigated. Dissolved gas measurements from 2010-2015 were used to look for evidence of a link between degradation of the transformer insulation and heightened levels of SYM-H and dB/dt as measured at Eskdalemuir magnetometer station. Firstly, case studies were examined of the most significant storms in this time period using dissolved gas analysis (DGA) methods, specifically the Low Energy Degradation Triangle (LEDT). The case studies were then augmented with a statistical survey, including Superposed Epoch Analysis (SEA) of multiple storm events. No evidence of a strong space weather impact can be found during this time period, likely owing to the relatively quiet nature of the Sun during this epoch and the modernity of the transformers studied.
Plain text (extended) Summary
An investigation into space weather impacts on UK power station transformers
Z. M. Lewis, J.A. Wild (Lancaster University Physics Department) and M. Allcock (EDF energy).
Abstract: It is well documented that space weather can impact electricity infrastructure, and several incidents have been observed in recent decades and directly linked to large geomagnetic storms (e.g. the Hydro Quebec incident in 1989). However, less is understood about the impact of lower-level Geomagnetically Induced Currents (GICs) on the health of transformers in the long term. Direct measurements of GIC at power stations are still quite rare, but all transformers sense and record levels of dissolved gases within the insulation, as a way of monitoring transformer health. In this study, the long term impact of geomagnetic activity on 13 power station transformers in the UK was investigated. Dissolved gas measurements between 2010-2015 were used to look for evidence of a link between degradation of the transformer insulation and heightened levels of global SYM-H and the rate of change of the horizontal magnetic field (dB/dt) as measured at the Eskdalemuir magnetometer station. First, case studies of the most significant storms in this time period were examined using dissolved gas analysis (DGA) methods, specifically the Low Energy Degradation Triangle (LEDT). These case studies were then augmented with a statistical survey, including Superposed Epoch Analysis (SEA) of multiple storm events. No evidence of a strong space weather impact can be found during this time period, likely owing to the relatively quiet nature of the Sun during this epoch and the modernity of the transformers studied.
The transformers have been anonymised (A-M), and the six gas concentrations shown are the ‘key’ gases used in DGA: methane, ethylene, ethane, hydrogen, acetylene and carbon monoxide.
Figure 1 shows typical DGA gas concentrations for one of the transformers in the data set (E). There is a general upwards trend on a multi-year scale and gaps indicative of maintenance or oil change.
Figure 2 shows SYM-H and dB/dt for the St Patrick’s day storm in March 2015 and Figure 3 shows the raw gas data for this storm in transformer J. The lines for each gas concentration appear very stable on this time scale.
The LEDT method was outlined in Moodley and Gaunt, 2017 and differs from other DGA methods as it aims to predict a fault before it occurs.
The LEDT combines three gas concentrations on a triangular plot: methane, hydrogen and carbon monoxide. One vertex corresponds to ‘normal’ operation of the transformer and any movement away from this normal region corresponds to early indication of a fault.
Figure 4 presents the LEDT (plotted for 72 hours before and 120 hours after the minimum value of SYM-H, for the St Patrick’s day storm). The LEDT shows the transformer to be operating away from the normal region but the location on the triangle is stable throughout the storm period.
Figures 6 and 7 show another example of an LEDT for transformer D for another storm (October 2011). The LEDT and R index show a small movement away from the normal region during the recovery phase of the storm. This is possibly a direct effect of GIC within the transformer, but the change is very small (+0.006 in the R index).
A superposed epoch analysis (SEA) was performed. Figures 9, 10, 11 and 12 show the results for different gas concentrations centered on the start of the storm main phase. These results show that no general upwards trend exists in the transformer following the onset of the storm.
Z. M. Lewis, J.A. Wild (Lancaster University Physics Department) and M. Allcock (EDF energy).
Abstract: It is well documented that space weather can impact electricity infrastructure, and several incidents have been observed in recent decades and directly linked to large geomagnetic storms (e.g. the Hydro Quebec incident in 1989). However, less is understood about the impact of lower-level Geomagnetically Induced Currents (GICs) on the health of transformers in the long term. Direct measurements of GIC at power stations are still quite rare, but all transformers sense and record levels of dissolved gases within the insulation, as a way of monitoring transformer health. In this study, the long term impact of geomagnetic activity on 13 power station transformers in the UK was investigated. Dissolved gas measurements between 2010-2015 were used to look for evidence of a link between degradation of the transformer insulation and heightened levels of global SYM-H and the rate of change of the horizontal magnetic field (dB/dt) as measured at the Eskdalemuir magnetometer station. First, case studies of the most significant storms in this time period were examined using dissolved gas analysis (DGA) methods, specifically the Low Energy Degradation Triangle (LEDT). These case studies were then augmented with a statistical survey, including Superposed Epoch Analysis (SEA) of multiple storm events. No evidence of a strong space weather impact can be found during this time period, likely owing to the relatively quiet nature of the Sun during this epoch and the modernity of the transformers studied.
The transformers have been anonymised (A-M), and the six gas concentrations shown are the ‘key’ gases used in DGA: methane, ethylene, ethane, hydrogen, acetylene and carbon monoxide.
Figure 1 shows typical DGA gas concentrations for one of the transformers in the data set (E). There is a general upwards trend on a multi-year scale and gaps indicative of maintenance or oil change.
Figure 2 shows SYM-H and dB/dt for the St Patrick’s day storm in March 2015 and Figure 3 shows the raw gas data for this storm in transformer J. The lines for each gas concentration appear very stable on this time scale.
The LEDT method was outlined in Moodley and Gaunt, 2017 and differs from other DGA methods as it aims to predict a fault before it occurs.
The LEDT combines three gas concentrations on a triangular plot: methane, hydrogen and carbon monoxide. One vertex corresponds to ‘normal’ operation of the transformer and any movement away from this normal region corresponds to early indication of a fault.
Figure 4 presents the LEDT (plotted for 72 hours before and 120 hours after the minimum value of SYM-H, for the St Patrick’s day storm). The LEDT shows the transformer to be operating away from the normal region but the location on the triangle is stable throughout the storm period.
Figures 6 and 7 show another example of an LEDT for transformer D for another storm (October 2011). The LEDT and R index show a small movement away from the normal region during the recovery phase of the storm. This is possibly a direct effect of GIC within the transformer, but the change is very small (+0.006 in the R index).
A superposed epoch analysis (SEA) was performed. Figures 9, 10, 11 and 12 show the results for different gas concentrations centered on the start of the storm main phase. These results show that no general upwards trend exists in the transformer following the onset of the storm.
URL
z.m.lewis@lancaster.ac.uk
Poster file