I am looking for a way to filter out bad energy readings. None of the built in filters seem to do what is required… If the next reading is more than 0.1 from the last read then replace it with the previous reading as this reading is bad. The outlier filters almost does this if you set the window size to 1 and the radius to 0.1 problem is that it pulls its previous values form the original sensor. So it will never filter out the bad reading. Talking to dgomes on github they stated that this can be done with 2 cascading Outlier filters… I have no idea how this is possible as they both work the same. Anyone have a clue how to do this using cascading filters?
I could also make a temple sensor that just does all the math for me and use that sensor at the filter but no clue how to get the last filtered value from the template sensor…
Does anyone else have any ideas? I know others have said these spikes are from sensor availability issues but in this case they are not from what I can tell. When looking at the states database for the sensor there are NO unavailability records unless I am looking at the wrong thing? and normally if something goes unavailable it goes from value to 0 then back to value meaning the jump in reading should be at minimum twice the original value when this happens. for me its only jumping by like 10% or 25%.
Please feel free to educate me if I am wrong, Willing to see what I might not be understanding!