extract full period of waveform signals to calculate active power accurately

[Jun_l]

Hi, I have done a vi that allows me to measure voltage U and current I signals and calculate the dissipation factor tan delta at the end using calculations of power:

1) Apparent power S = Urms * Irms

2) Active Power P = mean value of the instantaenous power u(t)*i(t)

3) Reactive Power Q = sqrt (S^2-P^2)

4) Phi = arccos P/S

I have a subvi Power that does the calculations and compare the results to those using Power.vi from Labview.

Before the calculations, the signals are conditioned - as seen in the picture Signal_Cond., i extracted the u(t)*i(t) signals, multiplied with 10 periods, and calculated the active power from it.

My values of voltage and current are correct, however - my phase Phi fluctuates a lot using my own subvi. This fluctuation is smaller using labview's Power.vi, and there is a difference of 0,5-1° in both values. Also, I am measuring the current through a 50 Ohm resistor, shouldnt the phase angle be smaller than 4°?

I tried to condition the U-I Signals for the labview's Power.vi, but it says that my signals don't have the same dt? So im guessing that my 'signal conditioning' using dynamic wires and extracting 10 periods of it then changing it back to arrays isn't a very good solution. Does anyone have any other options on how to extract a n-period of waveform signals so as to ensure accuracy of my Active Power and Phase angle phi?

I would appreciate some help!!!

Jun

[Henrik_Volkers]

First  eliminate the express vis and the dynamic data wires 🙂

Now create some analytical signals with known parameters to test your vi.

The tricky part comes when there are switched power supplys or consumers ... with low power ...usually not in sync with the line frequency...

ZES had some papers publisched on how they handle some of the problems ...

[Jun_l]

I tried the programm at first using signals directly from a signal generator, the values were what I expected. They only seem to not work in my test setup...

I have written my programm in a way that I can let it run while i measure voltage and current as a function of frequency, and only save the values I want when I press the save button. 

I wanted to only extract 10 periods of 'clean' signals of every frequency, which are then used to calculate my values- but I couldn't find another way to do it without using the dynamic wire and express vis. Perhaps it is why my values fluctuate so much? Do you or anyone here have any idea on how to 'cleanly' subtract/sample a n-periods of signals? My sample rate is 200 kHz and num of sample 100k, frequency from 50 Hz to 2 kHz. Im using a NI USB 6466 so both channels should be able to sample simultaneously.

ps: thanks for the tip about ZEM!

Jun

[Henrik_Volkers · Accepted solution]

here is a simple version as a subvi and snippet , you see: not red dots 😉

The 'evel' dynamic data will almost always show red dots, indicating that some(?) data is converted to a matching datatype .. maybe the data you expected 😄

If you use a current transformer, keep in mind that these can create a phase shift. (according to en wiki up to 6° for simple ones)

I saw you use the german version of LV. I always strongly recomment to install the english version ... 

[Jun_l]

This worked perfectly, thanks! Can't believe I have never came across to this vi.

My phase difference and power measurement still aren't accurate, but I'm guessing this is due to my test setup, I have a phase different somewhere coming from my components but still can't figure it out...

I actually do prefer the english version and I've tried changing the language but apparently I can't simply switch it at settings but have to reinstall it? Since I got the software from my institute I decided to leave it just like that.

Again, thanks for your help Henrik, appreciate it!

Jun