Earth Notes: On Website Technicals (2025-05)

This afternoon, alongside the frequency response log for the Eddi, I added a more general log of reasons for Eddi to have been put in STOP mode. This gets entries only when the Eddi would otherwise be diverting or boosting, so reflects energy spilled to grid that might not otherwise be. (This is getting added to archived logs too.)

Much of today's thrilling log, including initial manually-driven entries:

STOPPED 2025-05-30T18:04:42Z grid 7d red carbon intensity
STOPPED 2025-05-30T18:04:45Z grid 7d red carbon intensity
STOPPED 2025-05-30T18:05:01Z grid 7d red carbon intensity
STOPPED 2025-05-30T18:06:02Z grid 7d red carbon intensity
...
STOPPED 2025-05-30T18:32:02Z grid 7d red carbon intensity
STOPPED 2025-05-30T18:33:02Z grid 7d red carbon intensity
STOPPED 2025-05-30T18:34:01Z frequency too low at 49.85Hz
STOPPED 2025-05-30T18:35:02Z frequency too low at 49.89Hz
STOPPED 2025-05-30T18:36:01Z frequency too low at 49.89Hz

The next day I added a grid flow figure in W after the timestamp. Negative means spilling to grid.

In many ways the most interesting interactions between 16WW energy systems and the grid are dynamic, not just fixed 4pm to 7pm avoidance, useful though that is.

The system currently blocks diversion to the heat battery when grid frequency or voltage are low.

I can also block diversion when the grid intensity is 'red' (top quartile).

Ultimately I will use less electricity when I avoid DHW by immersion, since the heat pump will use fewer units to cover the same demand. Though I may import and thus pay a bit more. And the water might not be quite so hot...

This new code fragment does the work:

    if [ -f _gridCarbonIntensityGB.7d.red.flag ]; then
        echo "STOPPED: grid 7d red carbon intensity" 1>&2
        if [ "6" != "$STATUS" ]; then sh script/myenergi/eddiStop-netrc.sh 0; fi
        exit 1
    fi

This once-per-minute script may be getting too heavy.

2025-05-20

This morning the grid-red rule held back the start of diversion to the heat battery by ~1.5h.

I briefly looked at blocking diversion (or allowing to heater 1) if grid voltage is low (or high).

The stats do not look that hopeful from a cursory examination of Eddi coarse 10-minute sample data, but I might need to look again at much finer grain.

% gzip -d < data/eddi/log/202504.log.gz | sed -n -e 's/^.*"vol":\([0-9]*\)[^0-9].*$/\1/p' | sort | head
2332
2334
2338
2340
2341
2343
2343
2344
2345
2347
% gzip -d < data/eddi/log/202504.log.gz | sed -n -e 's/^.*"vol":\([0-9]*\)[^0-9].*$/\1/p' | sort -r | head
2475
2473
2473
2473
2471
2471
2471
2470
2470
2469
% gzip -d < data/eddi/log/202504.log.gz | wc -l
    4312
% gzip -d < data/eddi/log/202504.log.gz | egrep '"vol":25[0-9]' | wc -l
       0
% gzip -d < data/eddi/log/202504.log.gz | egrep '"vol":22[0-9]' | wc -l
       0
% gzip -d < data/eddi/log/202504.log.gz | egrep '"vol":23[0-4]' | wc -l
      20
% gzip -d < data/eddi/log/202504.log.gz | egrep '"vol":23[0-5]' | wc -l
     140
% gzip -d < data/eddi/log/202504.log.gz | egrep '"vol":24[4-9]' | wc -l
     200
% gzip -d < data/eddi/log/202504.log.gz | egrep '"vol":24[5-9]' | wc -l
      83

So maybe (below) 236V STOP diversion and (above) 243V allow diversion that might otherwise be STOPped?

I have inserted this extra 'secondary' task code segment for the lolz:

    # Prevent diversion when grid voltage is low.
    VOLTS="$(echo "$RESPONSE"|sed -n -e 's/^.*"vol":\([0-9]*\)[^0-9].*$/\1/p')"
    if [ "$VOLTS" -lt 2360 ]; then
        echo "STOPPED: mains voltage $VOLTS low; diversion stopped" 1>&2
        if [ "6" != "$STATUS" ]; then sh script/myenergi/eddiStop-netrc.sh 0; fi
        exit 1
    fi

A quick look at frequency extremes and distribution in the same data:

% gzip -d < data/eddi/log/202504.log.gz | sed -n -e 's/^.*"frq":\([0-9.]*\)[^0-9].*$/\1/p' | sort -n | head
49.76
49.83
49.83
49.83
49.83
49.83
49.83
49.83
49.83
49.83
% gzip -d < data/eddi/log/202504.log.gz | sed -n -e 's/^.*"frq":\([0-9.]*\)[^0-9].*$/\1/p' | sort -rn | head
50.21
50.19
50.19
50.18
50.18
50.18
50.17
50.16
50.16
50.16
% gzip -d < data/eddi/log/202504.log.gz | sed -n -e 's/^.*"frq":\([0-9.]*\)[^0-9].*$/\1/p' | sort -n | uniq -c | more
   1 49.76
  19 49.83
  24 49.84
  19 49.85
   4 49.86
  13 49.87
  36 49.88
 281 49.89
 130 49.9
  96 49.91
  50 49.92
  68 49.93
 134 49.94
 313 49.95
 394 49.96
 163 49.97
  79 49.98
  47 49.99
  82 50
 151 50.01
 555 50.02
 219 50.03
 138 50.04
  86 50.05
 118 50.06
 207 50.07
 417 50.08
 147 50.09
  90 50.1
  47 50.11
  39 50.12
  59 50.13
  42 50.14
  32 50.15
   5 50.16
   1 50.17
   3 50.18
   2 50.19
   1 50.21