Questionable solar generation traces

[syronos]

Hello forum - first post here.

I have a customer with a 30kW installation (8+ years old) with 2 x inverters (1 x SAMILPOWER SolarLake 15000TL and 1 x Solis Solis-15k). Below are some traces from the generation output. On a good day, all looks as expected (although not quite 30kW, despite panels recently being cleaned). However, on some days the output gets very choppy. Seemingly not because of cloud.

Does anyone have any ideas on how to check what's going on? We have 5-second data being logged via CT clamps.

Here is the past 10 days.



This is a 'good day':



And this is a questionable one:



Many thanks for any ideas.

 

[syronos]

What is interesting is that the days that are questionable are where the output seems to be over ~25kW. Could it be one of the inverters being faulty or some panels that become faulty at high output (and possibly temperature)?

 

[syronos]

For those interested, the last four days traces with local UK-AIR UV station data shows the first three days of as expected data and then the fourth day with very volatile traces.

https://snapshots.raintank.io/dashboard/snapshot/lLbqJAe5tm5SzdtSJjA41QuIjgQfbZX0

 

[johnb2713]

I think you need to monitor the voltage going in to the inverter from the panels to see what's happening. I think you're probably right on the temperature theory, maybe invert cooling fans? Looking at the traces, I would favour inverter issue over the panels, the fault occurs as generation increases and doesnt stop until a lot of the generation has gone. Looking again and thinking out loud, a connection burning up? I think a meter on the DC from the panels is the only way to really see whats happening.

 

[syronos]

I will try and sort out the DC monitoring and get some traces to then compare DC / AC. Just need to work out how to do that, as it would have to be invasive as opposed to CT-based?

 

[binky]

I would vote for the Samil being faulty, they went bust due to too many warranty claims. Turn off the Samil and see if the problem disappears.

The only panel issue you would see in this weather is overheating causing drop in power, but that should be a linear decrease. Panel faults also tend to be permanent, so if it was a panel, I would expect to see it constantly

 

[syronos]

Your thoughts on the panel issue align with ours - we would expect to see the issue more often. I assume we can switch off the Samil and that would just be it. No need to disconnect? It's typically cloudy and rainy at the moment!

 

[binky]

Your thoughts on the panel issue align with ours - we would expect to see the issue more often. I assume we can switch off the Samil and that would just be it. No need to disconnect? It's typically cloudy and rainy at the moment!

ahhh the joys of solar fault finding .

Just turn it off, at the AC isolator, you can do the DC aswell of course, but no actual need.

 

[syronos]

Not sure switching off the Samil system has made much of a change (marked with the line). Will leave it for 48 hours and then switch over the inverters.

 

[syronos]

Not sure switching off the Samil system has made much of a change (marked with the line). Will leave it for 48 hours and then switch over the inverters.

View attachment 15664

We had heavy rain this AM, hence the delay in generation.

 

[binky]

Da

We had heavy rain this AM, hence the delay in generation.

Daft question, but why are you looking at this system at all?

 

[syronos]

Da

Daft question, but why are you looking at this system at all?

No daft. Helping agri client understand and optimise their energy production and usage. Pre-existing installation which may need an update. Not wanting to suggest new inverter(s) for no good reason.

 

[binky]

So the customer wasn't reporting a fault or drop in output?

 

[syronos]

So the customer wasn't reporting a fault or drop in output?

As we put in monitoring for other farm equipment, we also included the solar. Production was below the installed 30kW, so we had the panels cleaned (never been done). I've been looking at the traces to see if there is a reason other than cloud, but I think that's all it is. Switched off each inverter and the output traces seem the same. Will put a light sensor and double check, but I've not looked at solar traces this closely before, so maybe when looking over 5s readings you're going to see the choppy nature if cloud is coming and going. I naïvely assumed it would be a smoother curve!

 

[syronos]

I did question whether it was a particular phase, but they seem to be on parity with one another. (The drop at ~9:30am was the switch-over between inverters.)

 

[binky]

Definetly never a smooth output, but yours does seem rather choppy, but that maybe a function of the monitoring gear. It's surprising how reactive panels are to light changes, but... The inverters have displays that show instant outputs, I would look at those and see if they match your monitoring. My general position on solar is if it's working it's fine, the inverters self check the system on start up, and tend to find any faults.

 

[fourtytwo]

TBH I think your choppy output is just atmospheric conditions, this is a plot from my system on a typical Lincolnshire day, one chop every 10 minutes is caused by a panel scan but the rest is just passing clouds & birds!

Also using UV plots may be misleading as I believe typical solar panels have low sensitivity to it and it passes through
clouds.

 

[binky]

TBH I think your choppy output is just atmospheric conditions, this is a plot from my system on a typical Lincolnshire day, one chop every 10 minutes is caused by a panel scan but the rest is just passing clouds & birds!
View attachment 15714
Also using UV plots may be misleading as I believe typical solar panels have low sensitivity to it and it passes through
clouds.

Panels work off infrared rather than uv