rf_46 circuit label uncertainty

[dataknut]

Inspection of the raw data shows that rf_46 has a lot of different circuit labels:

• some have the form xx_Imag$nnn - these may be 'imaginary power'
• one is voltage (it says so!)
• some have the form xx1$nnn or xx2$nnn - these may or may not be power, imaginary power or voltage

• some are 'normal'

  rf_46 was not re-used and the similarity of the label-sets seems to indicate typos.

[dataknut]

fixed: fixCircuitLabels_rf_46() fixes the labels to just the first label-set.

[dataknut]

It looks like things are more complex than we thought. Power, voltage and imaginary (reactive) power may have been monitored for this dwelling. Thus:

• labels with '1' = probably power with 1 voltage reading (Incomer Voltage$4405)
• labels with '2' or '_Imag' = probably imaginary power with 1 voltage reading (Incomer Voltage$4405)
• labels with none of these forms - probably power

We have yet to fix this in the data so we advise removing dwelling rf_46 from the data prior to analysis.

[dataknut]

nope:

(https://github.com/CfSOtago/GREENGridData/blob/master/plots/rf_46_circuitLabels.png)

• _Imag 4399 -> 4404 = 4399 -> 4494 so these are probably imaginary power
• voltage = voltage (as labelled)
• but 1 & 2 ? We don't know.

This means that the results of fixCircuitLabels_rf_46() are almost certainly incorrect and the rf_46 data in the UKDA reshare archive is not reliable.

For now guidance remains the same: do not use rf_46.