Metering algorithm configuration tool derivative current sensor

[florianhumblot]

I am trying to use the three-phase power meter design with current sensors. Everything seems to be working correctly except when mains frequency is not perfect 50Hz.

We can see this using the meterlib algorithm configuration tools too, setting frequency to 51Hz:

The algorithm simulator has no errors within the 0.10 error limit. However, if we change to derivative current sensors:

Now there is a -1.9% error on all the current measurements.

I've tried modifying almost every single parameter and combination but I could not manage to stabilize the current sensor outputs.

Am I missing something?

The software is from the AN4265

The settings are all left default in the above screenshots but I have tried using the knob-control to modify all the parameters.

[diego_charles]

Hi @florianhumblot 

Hoping that you are doing excellent!

We also replicated the errors above 0.1% when setting the frequency on 51 Hz in the signal generator. Since we have a nominal frequency of 50 Hz, in the General setting area of the library, generating signals with that fixed frequency 51 Hz, will cause that error percentage.

Could you detail further the 51 Hz? For example, if you are trying to adapt the library for 50 Hz AC mains with some degree of error/variations on its period, or if you are contemplating a scenario where you could get a fixed frequency of 51 Hz.

Many thanks your patience.

Diego

[florianhumblot]

Hi @diego_charles,

Thanks for getting back to me!

The errors reported by the algorithm simulator seem to be between 0.9% and 1.1% per 0.5Hz deviation from the nominal frequency.

In Continental Europe the frequency setpoint is 50 Hz. It shall not go below 47.5 and above 51.5 Hz. However, as far as I understand metering requirements, the energy meters should remain accurate within that range.

Getting a total kWh metering error of 2% is practically impossible within that range if the current measurement error gets close to 5% if the frequency drops too low, or 1,5% if the frequency is too high.

So reiterating, this would really be to account for frequency variation/errors in local substation/the grid and not because we envision any special setup at a fixed 51Hz frequency.

Regards,

Florian Humblot

[diego_charles]

Hi @florianhumblot  

Many thanks for reaching us with that issue, and for your patience. 

I have received feedback from the specialist AE team:

This error relates to a simulator issue in config tool where source current signal is not computed as a derivative and hence after integration simulator provides wrong number.

In real application, library provides similar performance for Rogowski coils as for shunt resistors or CTs - protocol with Rogowski coil sensor attached (refer to yellow area with 49 and 51Hz frequencies).

We can update simulator but it will take some time.

In the meantime you can do actual  calibration tests for the frequency variations you expect. 

Sincerely, 

Diego