Hi,
part 1) As power readings are received, how are these converted to KWh? To integrate the KWh value, where does the TIME value come in? Is the “fixed interval” time used? Or does it take the delta T from the previous value entered?
part 2) The documentation says feed “fixed interval” times should be set to equal or greater than the time expected for data values to be received ( so as to not use up too much memory). So, what do these times do in the calculations if they can be set to values GREATER than the expected send intervals? If its too much larger - say data is sent every ± 10 secs, and th “fixed interval” is set to 60secs, how will it affect the results?
Hi Paul. I’ve only just stumbled across your post which poses a good question - one that had occurred to me recently in the context of whether the emon can give an accurate kWh for rapidly fluctuating loads - like the Mk2PVrouter supporting a heat pump’s hot water production - when emon is only recording occasional snapshots from the c.t.
Perhaps the kWh algorithm works on the raw input before the input is periodically logged to a feed?
Cheers, David.
This depends on the software in the emonTx or in the front end of the emonPi. If you have the “CM” (Continuous Monitoring) version, it accurately reports the average power over the whole period since the last report, and accumulated kWh to the end of the sampling period.
Your fears will only apply to the Discrete Sample versions prior to August 2020 (emonTx) and July 2021 (emonPi).
To partially avoid this issue, I decided to take readings every 15 seconds… Things can happen in thoese 15 seconds, but the chances are lower than if taking readings every 60 seconds, for example…
I suspect when I use my sous-vide cooker, which pulses every X seconds and does so for a fraction of a second, throws by readings off!!
It depends on how long your “fraction of a second” is. Depending on the number of inputs active (i.e. configured), emonLibCM measures (quoting its documentation) “Better than 1900 sample sets per second – using 4 channels of an emonTx V3.4 @ 50 Hz.”
Doing the same with emonLibDB (which has a lot more work to do, so it’s not quite so fast) “Better than 1480 sample sets per second – using 3 + 12 single-phase channels of an emonTx4 @ 50 Hz, which equates to 29 or 24 samples per mains cycle at 50 Hz or 60 Hz respectively.”
So your “fraction of a second” has to be a mighty small fraction to confuse it.