Metering of facility main service displaying unusual readings during testing of back-up generator

[Caker]

We have a piece of energy metering installed at an industrial client facility. The readings from the meter look to be normal except whenever the facility test the back-up generator at the facility, during which we see unusual readings from the metering, namely that the real and reactive power begin to oscillate, with a cycle taking between 13 and 19 seconds to complete.
We are not sure what the exact cause of this issue is. We have suspicions that the cause could be that the voltage reference is on the utility side of the transfer switch between the utility and generator system, and the Current transformers are located on the load side of the transfer switch; but do not have documentation of the installation to be certain about this.
Are there any suggestions as to what is going on with the system, or alternative potential causes?

---

Below are pictures of charts for 100ms data collected during one of these events. (Cannot figure out how to embed images, so added google drive Links. If I get feedback on how to embed images I will update accordingly).

Overview: https://drive.google.com/file/d/1G1eMwWstdLQL9uGHQDJnEu6TEtJ37W9g/view?usp=sharing
Frequency (Slight increase during event): https://drive.google.com/file/d/1-RvD-Cz0wmxTbIxtS_-4ryWzij3Z0vzE/view?usp=sharing
LL/LN Voltage (Slight decrease during event): https://drive.google.com/file/d/1YTVe6LLINLgaDhS75guO4qoc0yrUlEWQ/view?usp=sharing
Current: https://drive.google.com/file/d/1JBKlJ6740ZeKsLoQO7hq_YvzBrnqbLcC/view?usp=sharing
Apparent Power: https://drive.google.com/file/d/19ZTHE1XaUa5AQjeeMkF9jlhaKyLzUsRg/view?usp=sharing
Demand: https://drive.google.com/file/d/1MmWqh4p4VJEMoWHZvCTLu7ukz3z_CKav/view?usp=sharing
Power Factor: https://drive.google.com/file/d/1P-jr-wSTKzjaY3Bw7e_pl34p-59ukHPN/view?usp=sharing

 

[ron]

With the big guess that the metering is correct, maybe it is the standby generator malfunctioning, or the load varying when it restarts on a relatively soft source compared to the utility.

The load generally decides the type of consumption, whether it be kW, KVA or KVAR.

 

[OldBroadcastTech]

Just a S W A G.but perhaps a slight frequency difference between utility and generator ?

 

[zbang]

I'd go for that and it should be easy enough to check.

"Sure, I'll take 'hum bars' for $300, please."

 

[synchro]

Just a S W A G.but perhaps a slight frequency difference between utility and generator ?

Give that man a cigar.
Apparently the frequency difference is between 1/13 and 1/19 Hz.

 

[GoldDigger]

I would not call it a S W A G at all, but a purely simple diagnostic: Whenever the CTs and the PTs in a metering system are connected to different systems which are not only at the same frequency but also phase locked, that power metering cannot work.
I suppose you could call it a S W A G that the metering was connected that way, but on the face of it that fits the symptoms so precisely that it would be my first thing to check.
The utility metering has no problem since both CTs and PTs are upstream of the ATS. And I can see that it might seem convenient for the private metering system installer to put the PTs on the"wrong" side of the ATS. If the currrent transformers were not downstream of the ATS, they would not measure any power consumption when the site is on generator power.