Removing suspected noise from CT

Hi Guys,

I have an anomaly where by a circuit which is connected to an Immersun PV Diverter which is in turn connected to the immersion heater, show a small negative watts when the unit is not diverting power. 

I have added the "Allow Positive" to the node and it seems to work such that it only reports + numbers.

However i use the - input process on a different node to remove the diverted figure from the power1+power2 figure to give me a "true" house reading and it is here where the small negative still has an effect.

Is there a way to use the after processing figure in a feed to create the figure i require.

I have added a simple sreen shot of figures. The "House inc IMM" should be the same as "House" As Immersun is 0

Keep up the great work.

Regards

Paul

 

pb66's picture

Re: Removing suspected noise from CT

Hi Paul

If you suspect it is noise or just prefer not to see it you can remove or hide it by adjusting the value by a chosen amount (4w) or threshold (5w) before doing the "allow positive" and then adjusting back again by the same amount effectively making a "allow values over 4w" 

However the probability is the reading is accurate and the 4w is the draw of the immersun unit itself, since you cannot differentiate between the control circuits consumption and the power it is controlling you could also consider using just a plain +4w on house and -4w on diversion as it is always present 24/7.

Post your processing lists if you need any pointers. or if the math doesn't stack up as, it maybe the difference is only 2w but adding rather subtracting or vice versa is doubling the difference (if that makes any sense) it's difficult to tell from the info given

Paul

wheelbarrowhandle's picture

Re: Removing suspected noise from CT

Hi Paul,

Thanks for taking the time to respond. The problem is that, A) the figure isn't a constant. and B) it cant be the immersun box as the figure is negative. When the immersun box is working the figures are positive as i would expect.

Here is the processing lists

emonTx_1/rx/power4 config - CT4 around the live wire going to the immersun box.

1 Allow positive
2 Log to feed Immersun (feedvalue:2352.00)
3 Power to kWh Immersun Kwh (feedvalue:4.93)

Then from EmonPI=i

emonPi/rx/power1_plus_power2 config

1 Log to feed House Immersun (feedvalue:2683.00)
2 Power to kWh House Immersun Kwh (feedvalue:132.91)
3 - input emonTx_1/tx/power4 2352
4 Allow positive
5 Power to kWh House Kwh (feedvalue:3.90)
6 Log to feed House (feedvalue:331.00)

As you can see the CT4 around the live lead is +2352 which is fine. It when it off it goes negative. which is cant be as it would still be drawing some power to power it's self.

Regards

Paul

 

 

pb66's picture

Re: Removing suspected noise from CT

Ahh ok, i didn't grasp the "anomaly" was switching direction as I expected the allow positive to remove it, which it does, but then you are using the unprocessed input again rather than the processed result hence your opening question, which now makes perfect sense since the penny has dropped.

Yes you can! -feed is what you need to use rather than -input then select the "Immersun" feed.

Paul

EDIT - Scrap that! the +,-,x,/ feeds didn't appear to make it into the emonPi version of emoncms. Sorry I hadn't realized that until I just checked.

EDIT 2 - Figured out why they are not there despite being in v8.5, it's because of the new 'node' input rather than the original input, the 'implemented' processes are all repeated in the node module but are not all present. Hopefully the +-x/ feeds will make it into future versions as they are without a doubt far more powerful than the +-x/ inputs and more robust too.

Maybe some calibration might help?

sheppy's picture

Re: Removing suspected noise from CT

I have a similar problem, what I do is use the worse affected input and then add the whole house calculation on top of it. In my case its the solar input and it reads:

Calibration plus (-38)

Calibration multiply (0.92) - I need this to make the calibrated EMONTX agree with the Renesola Monitor

Allow Positive

Log to Feed (Solar_Power)

Plus Input (Grid)

Log to Feed (house_power)

Minus Input (HWC Divert)

Minus Input (Spa Divert)

Calibration Plus (1)

Log to Feed (standby power)

I then use house_power in the solar graphing app and standby power in my dashboard. The last calibration plus 1 is to get around a fairly constant leakage which moves between each of the diverter circuits. This gives me a graph of solar overlaid with grid on the solar app and a dashboard I can quickly check as I leave for things left on.

It will get developed further once I solve the how to get the fridge and PVR PC usage into EMONCMS from my current cost individual appliance monitor question.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.