Strange Data Anomalies

edited January 5 in Bright App

I am an Octopus Agile customer who also uses the Bright App and a GlowMarkt IHD. As some will know Octopus / TMA have a systemic problem uploading timely and reliable smart meter data, unlike the Bright API which seems to be solid. Anyway because Octopus drop maybe 3-4 days of meter readings each month that require an escalation to their help desk for them to do a manual pull, I've put a layer of data fusion into my HA system:

  • My IHD does the ~10s updates to my local MQTT instance and I do ½hr cut detection and deltas to do realtime HH estimates.
  • Once a day I use the Bright API to do an incremental download of new historic HH data.
  • Once a day, I use the Octopus API to do an incremental download of new HH data, but only commit whole days and then retry any dropped days so this catches up eventually.

The HH estimate is usually out by the odd 0.001 kWh because of edge jitter, but it is good enough to feed into HA plots. The Bright and Octopus APIs access (albeit via different routes) the same Smart Meter HH registers so these agree (if I don't have dropped Octopus data).

Except that for a window 8-16 Dec 2023 when the Bright HH readings were exactly half of those from Octopus (and my IHD estimated numbers). Luckily the Octopus data was consistent with my actual usage estimates, so no billing issues, but just a bizarre and pretty much one-off Bright artifact where it reported half the true value.

The Bright App also reported the same ½ scale values for the same HH slots.

I am happy to provide my details and data if anyone from Glowmarkt wants these.

Sign In or Register to comment.