PeterDawson

Hi All
This is my first post, but I've been a long time lurker, so first off, thanks to the creators of PowerForum and to all the contributors for making this a truly great information resource.
Background:
In Oct last year I installed a small off-grid system in the SA bush.  
Axpert MKS 5kVA. 9 x 260W panels (3x3). 1 x Huawei ESM-48100 battery (48V, 100Ah, 4.8kWh). Generator connected to AC in, but only used to charge batteries very infrequently. ICC with BMS-702 used for monitoring and remote setting adjustments. 90% of the time the house is vacant and only the fridge draws power from the system.  
For the most part the system has worked very well.  There are however some strange behaviours from the Axpert (and I suspect the MPPT) that I have not been able to iron out.
The first of these strange behavious is illustrated in Fig 1 below showing the charge period of the battery in the morning.   The SolarWatts (blue pen) shows the fridge's compressor periodically switching on and off.  Batterywatts (orange/yellow pen) shows the power flowing either into or out of the battery.  Notice how the pvwatts (green pen) sometimes drops right off when the fridge starts up.  This behaviour is not consistent as can be seen at 08:15ish.  Any ideas?

Fig 1. PV power drops to zero when fridge compressor switches on.
 
The second, and more annoying behaviour is illustrated in Fig 2 below.  Sometimes, the MPPT seems to shut down and the inverter switches to battery mode.  The panel voltage jumps to Voc (103V for panel Voc of 36V), and the current drops to zero.  The trend in Fig 2 shows this happening at the same time the fridge compressor switches on.  This is also not consistent.  Fig 3 shows it happening when the fridge compressor switches off. It does however seem to always be associated to a change in inverter load, but the size of the load does not seem to matter.  Its not worse when the draw is higher, say when the dishwasher element or the kettle switch on or off.  Sometimes things resort to normal in a few minutes or hours, but typically only the next morning when the sun comes up.  If I happen to notice that the MPPT is not drawing any power from the panels, I can log onto the RPI remotely and bump the Bulk Charge Voltage setting by .1 and that kick starts the MPPT again.  
This behaviour does not seem to correlate with battery SOC nor with the panel output.  Sometimes we can go for weeks without it happening.  Today it has happened twice, once early this morning when panel power was 175W (Fig 3 @06:57), and then again just after I bumped it back to life with panel power output at 1,300W (Fig 4 @09:37).

Fig 2. MPPT seems to switch off completely and inverter switches to battery mode when fridge compressor switches on.
 

Fig 3. MPPT seems to switch off completely and inverter switches to battery mode when fridge compressor switches off.
 

Fig 4. MPPT seems to switch off completely and inverter switches to battery mode when fridge compressor switches on and panel power at 1,300W.
 
I'd be very grateful for any ideas or advice on solving this.
Thanks
Peter
 
 

The 07:00 to 09:30 is when the SCC was in its "tripped" state.  The 11:00 to sunset is as you suggest, full battery and solar is merely running the fridge and the float of the battery.
Here is another example of the SCC "tripping".  It went down at about 11:33 when the washing machine element switched off.  Fig 2.  
It then stays down until I noticed at 14:30ish and got it going again.  Notice that I ran the generator to charge the battery up from 16:00 to 18:00ish.

Fig 1. Whole day with relatively heavy system use.

Fig 2. Showing trip as washing machine element turns off.