Earth Notes: On Website Technicals (2025-03)

After updating the by-month electricity chart to include 2025, the formatting looked a bit different on my Mac (like old/previous rendering) and on the server (new: years are split into two columns in the key).

I took this opportunity to fix a colour numbering bug, and to use the plot for code form:

plot for [i=2:19] "data/16WWMonthlyElectricityConsumption.dat" \
        using 1:i title col smooth bezier lc i, \
     for [i=2:19] "data/16WWMonthlyElectricityConsumption.dat" \
        using 1:i notitle with points lc i, \
     "" using 1:20 title col smooth bezier lc 1 lw 3, \
     "" using 1:20 notitle with points lc 1

Tinkering continued, eg using linetype int((i-2)/10) to help make the different decades' lines easier to distinguish:

For my Sources/Links sections I use a very consistent format for links that have a particular date, eg press releases or news stories:

    <li>YYYY-MM-DD: ...

Sometimes there is no -DD or -MM-DD. Years are almost all 20XX.

So now I have updated the script to automagically wrap the date part in a time element, to ease machine parsing. Maybe that will help screen readers and search engines.

A side-effect of this tweak was every top-level main HTML page being rebuilt: desktop, mobile/lite and off-line. Somewhere over 100 (~25%) of the desktop pages had not yet been rebuilt this year, interestingly, thus still had the 2024 site copyright date in their footer.

Today (nominally at 01:00) a pasteurisation cycle became due. But grid carbon intensity was not (relatively, cf the last week) low, so all that happened then was the Eddi heater priority being set to 1, ie for the DHW cylinder.

PV diversion (starting 07:48) then brought the DHW tank up to its immersion cutoff temperature (assumed to be ~60°C) and the Eddi automatically started to divert to heater 2, ie the Thermino (13:17).

I had to teach the detection script a new way to detect the DHW tank being hot for this case (the second line of the awk conditions below). For good measure I added a backstop for when both cylinder and Thermino are full (state 5, on the third line).

Here are the awk lines to detect pasteurisation done from the Eddi log by printing the salient log line and exiting:

/"sta":[34]/ && /"hno":1/ && /"div":0/ {print $0;exit}
/"sta":[34]/ && /"hpri":1/ && /"hno":2/ {print $0;exit}
/"sta":5/ {print $0;exit}

A test case has been added for today: the DHW cylinder getting 'full' without the Thermino then also getting full.