My app talks to a meter via modbus to collect V, I, and PF per phase every 5 seconds. Now we are calculating power every 5 seconds, and thus, energy for this 5 second period as

Cumulative Energy += Power x 5 / 3600 (where '+=' increment and assign)

My question is, will this method be accurate? How can we improve upon this?

Constraints: Data acquisition is only for per phase V, I, and PF at every 5 seconds.

 

Hi

It depends on the variability of the load changes.

For example a motor that turn on and off in time periods of 10 minutes-on and 5 minutes-off during all day duty, may not be sensibly affected by a 5 second of sample period.

But if your load turns on/off faster than let say 10 seconds, your calculation may be very inaccurate.

For 5 seconds of sampling period, you may consider to be a good energy meter if the minimum "load change event" is 20 times slower.. This is about 100 seconds.

Many instruments calculate the energy with very fast sampling rate ..and If you polled (by modbus for example) the energy data every 5 seconds, all of the above mentioned doesn't matter.

But if you don't have this data available through your instrument you must apply the mentioned criteria.

Regards,
JOHN

 

It depends on how steady your load is.

If you have a fixed load that doesn't vary e.g. resistive heater direct on line it will be very accurate.