Modelmakeroz

One thing is to check that AC-in is detected when this happens. But I think you're showing that above; I believe that the large hexagon on its own at the left is AC-in, correct? It's hard to tell from your photo. 
The next thing is to determine whether the AC-in is going out of spec occasionally. Maybe you have a large load (it might not be your house, it might be nearby on the same transformer). Do you have data logging software that can show data like line voltage and frequency, and something like battery discharge current which will tell you when the problem is happening? Then you can hunt for records of when it happened, and see if there is anything unusual about the line voltage or frequency (etc) just before it happens.

The logging software does not record any anomalies before this happens - voltage and frequency are normal.  My current hypothesis is that the voltage stabilizer is causing this. I have a large 12kw voltage stabilizer  -it is installed on the grid input and it then feeds into the inverter. When the inverter switches from bypass to battery it (inverter) makes  a clicking noise about 1 time per sec. I can also hear a clicking noise inside the stabilizer. I have now switched off the stabilizer in order to test if this remedies the strange behavior of the invertor.

I can confirm that the problem above was due to a slow step stabilizer. Basically, when the stabilizer switched to a different circuit, the inverter was able to detect this small disruption of electricity and consequently switched to battery mode. The solution I've used is to replace the step stabilizer with a modern inverter-based model, that provides continuous flow. So, to anyone out there considering to install the new VMIII behind a stabilizer - they are much quicker than the older models and this can create trouble for you.