PV Inverter cutting off array sunny days

[RonanIrl]

I agree the inveter cooling sink can get hot but accoring to SOLIS and the installer no hotter than any other installtion they have done and not the cause of the problem.

Hello M - just triggered me on the conversion efficiency of the Invertor. For sure no Invertor is 100% efficient and these readings imply 1- 44/560 = 92%. So 8% in heat, hmmm ... smoking gun?
View attachment 60838

 

[marconi]

This afternoon I observed your real-time power figures for Generation, Grid, Battery and Consumption. It has clearly been very sunny where you live because the generation figures have been up in the 5-6kW range. The temperature of the inverter has exceeded 60C and you have had a shutdown of generation.

What I observed was 4600W output from the inverter to the CU; 900W consumption from the CU; 3700W export.

4600 = 900 + 3700.

With such high solar generation the battery should not be being discharged yet it was at a rate of 170W - this is not right. There is ample solar to supply the load and any export.

Here is the wrinkle; your installation is not set up to export because the backflow power is set to +0000W. And yet it is exporting - why?

I reckon the current sensor which monitors the glow of grid power into and out of your home is orientated the wrong way around - they have an arrow to indicate the supply and the CU direction - you could check this easily.

It might be though that the current clamp is orientated the correct way but its leads have been wired into the meter (which relays data to the inverter meter comms socket via an RS485 cable) the wrong way.

The wrong orientation of the current sensor or a switch over of its leads will reverse the sense of the data input on grid power flow so the inverter will interpret an export as an import and an import as an export.

For the self consumption mode of operation, any grid import is offset by the injection of inverter ac power of an amount conditional on the kW limit of the inverter pertaining at the time and the incident sunshine on the panels. The inverter does not use power from the battery for export to the grid - this battery is solely for home use.

A negative feedback loop controller ramps up PV generation until the ac output power cannot be increased any further. As the ac output power increases the imported power should reduce and if there is sufficient pv generation to zero import. However if the sense of the signal is incorrect the imported power signal rises and rises because it is actually representing export of the very same increased pv generation - a positive feedback loop - ie: export power becomes maximised. The inverter starts to get hotter.

When one studies the ac output power plot it cycles between maximums and a minimums in a way determined by the temperature related power output derating scheme, incident sunshine changes and load changes in the home. Eventually the inverter shuts down (or becomes unstable at higher temperature and looses sync with the mains) . It cools down, restarts and then carries on again.

Power figures right now: Gen 3.25kW, Load 0.9kW, Export 2.25kW Battery discharge 0.15kW

 

[marconi]

RonanIrl - you have seen this in my email to you.

Yes I did see the shutdown. I was observing the data and 'fortunately' I saw it happen. I hope my explanation in my last post of the EF makes sense; you must have covered positive and negative feedback which will be used in aircraft system controllers.

What made me explore positive feedback through an incorrectly orientated grid power flow sensor was the apparent oscillation of the total ac power output between peaks and lows when there was little home load; it could only mean this high solar power generation was being exported because it was not being consumed nor charging the battery because as noted the battery was discharging - something you noticed.

Once I had confirmed you were exporting shed loads of power in the self use mode in an oscillatory manner the most likely reason was going to be the sensor mistaking import for export and changing negative feedback into positive. The other give-away was the gradual temperature rise - it went over 60C when ac output was some 6kW - which instigates the derating of the power output to prevent damage to the power electronics.

An indicator was the pv1 current mimicked the total power generation (which was being maximised for export) and not the home load - for self consumption, once battery charging is subtracted it should follow the home load only - negative feedback control system in action to minimise imported grid power but never export.

Other key factors - all worked properly overnight when there was no pv generation available for export - as noted the battery will not be drawn on to export. At the moment the 150W odd discharge is making up for the inverter power losses and measurement errors.

What I do need to check further is you seem to be charging your battery from the mains sometimes - need to confirm but I don't suppose you want that to happen - I would not unless you have cheap rate electricity overnight that makes it worthwhile. I saw it during the day too though.

 

[RonanIrl]

All,

So the enginerr just left the house after testing and replacing a few things.

1. All firmwares updated
2. 1 shunt remoewd and 2 new wired in as they now belive the shunt was faulty and cutting off the PV array due to overheat, asked them to install one each string thsi time and move them lower in the attic to a cooler area which they did. Model Projoy PEFS-EL55-4
3. All wiring and connections checked inclduing all of clamps which there are 3, 2 for th eedi and one for the battery system allocated in my CU, They all checked out and were correctly installed. He also confirmed gauge and lengths for me... inverter to CU 6mm at 12M length, Cu to Meter 16mm approx 7 M Length, DC strings to Inverter 4mm at 7 M Length, not posisble to check Meter to Transformer as it is sleeved.
4. All EPM Backflow and bettery mode setting checked and were all correct.

They are almost 100% sure the shunt was the problem so lets see. Seems they cut off at 70 degrees temp inside the enclosure and frankly my loft where they were installed right at the peak could be 35 ambient on a super sunny day so perhaps that added to the power flow heating caused the shutdown. Although the manual says the shunt if it shuts down for temp must be manually reset by removal and reapplication of the AC voltage which seems to indicate power on and off cycle in the hosue which hasnt happened to date yet it had come on by itself again so who knows!!

If it is not the problem then they will adjust the country setting to increase to outbond voltage to be higher just in case it too low and our inbound is higher when exporting. Although we need to ask the network provider for permisison to do this but they always say yes if told why.

He did mention there is a fault code associated with grid voltage too high and it hasnt thrown that code ever OV-OG or something so he doesnt think that is the issue.

 

[marconi]

Thank you for the update. If it alright with you I will log in every so often and observe the system? - the next very sunny day here (maybe where you are too) is tomorrow Sun 20 Sep 20.

 

[RonanIrl]

Guys just to close this issue the movemmeg of the fireman switch shunts lower in my loft and the installation of one on each string seems to have solved the issue.

Appears they were exceeding their operating temp and shutting off on sunny days due to heat trapped in the loft.

Wont know 100% until nect summer but so far looks good.