A belated update on the project.
Using a RPi-3B running Raspberry Pi OS (Buster) on a 128Gb SSD via USB-SATA adapter and utilising Mosquitto, InfluxDB, Node-RED & Grafana free programs to configured a MQTT broker & data recorder, used Shelly EM hardware as the primary measuring source.
The RPi-3B is ethernet LAN connected to my Win 10 desktop PC.
The Shelly EM is Wi-Fi connected to the RPi-3B MQTT broker.
Over 3 quarterly electricity billing cycles, very happy with the results produced, overall consumption and cost are suprisingly accurate.
Total consumption and cost figures are manually reset on the quarterly meter reading day.
Tariff is entered/changed as required in the Node-RED formula.
Changing tariff within the quarterly measurement periods (very common occurance at present!) unfortunately results in incorrect total consumption/cost period readings for that quarter, probably a way around this other than manually resetting, recording and adding the two periods together, have not figured this out as yet.
Attached, a snapshot of the Grafana display showing a 7 day period.
The RPi-3B CPU temperature is also recorded.
Have also recently configured a hot water tank temperature monitor using a Shelly-1 relay with add-on temperature board & 2 probes, Wi-Fi connected to the MQTT broker, had reason to believe that the domestic HW temperature was inconsistant and thought the tank stat was on the blink.
One temp probe mounted directly beside the tank stat, the other attached to the tank top, HW discharge pipe.
Has actually proved that the HW tank stat is surprisingly accurate and repeatable!
Snapshot attached, again on 7 day display with boiler running morning and afternoon, tank stat set at 60 deg. C.