Stateful timeseries#1980
Conversation
* Document in DEBUG.md
* for virtual feeds
* Use HTTP API to update tariff data inputs
* More documentation about using stateful timeseries for costs and standing charges * Use PHP DateInterval specifications for standing charge process
townsen
force-pushed
the
stateful_timeseries
branch
from
98748b8 to
23ada1d
Compare
* Tests on PHP7.4 failing due to use of named parameters
* Keep the test happy even though precedence of the operators is explicitly defined in https://www.php.net/manual/en/language.operators.precedence.php
|
Hello @townsen Thanks for this pull request, appreciate your interest in furthering emoncms development! I'd be really interested in some more detail on how you are using this, are you using this with flat tariffs that only change e.g once a year? Do you use custom dashboards to visualise or are you able to use the emoncms apps? |
|
Hi trystan, 
The results can then be plotted on a graph showing my total energy bill and breaking it down into its 4 constituent parts, like so: 
Another use I have for these processors is to graph the duty cycle of my underfloor heating - which is obviously either on or off and turns on or off at irregular times. Again this goes to a variable timeseries and by using the 'state' process I can producesa graph that shows the state of the UFH over time. The 'step' option on graphing can do a similar thing but if the transition occurred before the current time window then the step graph will not show the correct results, nor will it show the state extending into the future.. There are some issues with the conversion of variable times into fixed times which I discuss in the notes, but in practice they can be ignored. To fix those would require a more complex approach (apportioning a value that changes in the middle of a fixed interval) which I don't have the energy to try and implement right now. I hope this helps. |
2 participants
Making a PHPTimeSeries feed Stateful
The TimeSeries feeds in EmonCMS contain datapoints at irregular time intervals. When used
for graphing the values of those feeds at times within these intervals is usually 'null'
(unless by chance a datapoint exists within the interval). However when such a
TimeSeries represents the state of a system or a tariff cost the value at any point in
time is the most recent datapoint, even if that falls before the interval.
To help solve these two related problems we introduce four new Virtual Feed processes.
See the file STATEFUL_TIMESERIES.md for a full description
With these features you can create feeds that represent point-in-time tariff changes and graph your energy cost over time including standing charges and accurately reflected costs over long periods. There is a small limitation discussed in the document which would require more significant changes to alleviate. This is a start.