Possible to monitor the extra (unused) power available on a Growatt SPF6000ES ?

[Mitker]

Hi,

I started running a Growatt SPF6000ES inverter, with which I'm globally satisfied, on an On-Grid (but no export) configuration. No batteries at this stage.

In order to optimize the system, meanwhile the primary devices are already running and if there's any power produced by the panels unused, I'd like to automatically launch secondary equipments (ex: oxygenating pumps for the ponds, tank water heater, etc.).
This implies I would need to collect this information (amount of extra power available) first, in real time and continuously.

My question is : Is this a data that can be retrieved somehow from the inverter ?
(I have no idea if the inverter is even in condition to estimate this value)
I didn't find any mention of this on the UM or searching on the net.

Thanks for any info or experience on the topic,

ps: definitely not a geek in electronics, but I should be able to write down and inject some basic code on a board that would process the data and run a few relays accordingly 😉
ps2: I don't want to go the empirical way (progressively increasing loads until detecting the inverter must draw from the grid) - unpractical

[Bandersnatch]

I have 4 Growatt SPF5000ES inverters. The ShinePhone app is pretty rubbish so I have have fitted a Raspberry Pi and using Solar Assistant software and a USB hub to link the 4 inverters together I am able to extract data from all the inverters 

 

 https://solar-assistant.io/

 

 

 

 

 

 

It is possible to push data from solar assistant into Home Assistant https://www.home-assistant.io/ then it is possible to manage smart devices like smart wifi breakers to switch on/off devices based on data from solar assistant.

 

I plan to manage my EV chargers to do that.

 

Here is a video I made a few years ago about smart wifi breakers

 

 

[Mitker]

Thanks for these infos

Interesting to see that from the screenshot you sent, it looks like you get this very data I'm looking for, under the 'solar PV' wattage figure (5232 W).
My only uncertainty it that this value is very close (by a couple of %) to the sum of the loads and the battery charge.
Would this be due that by chance that day your use was that close from what the system was generating, or that what appears under 'solar PV' is just the sum of the uses plus some losses and roundings ? In the later case, it wouldn't inform about the 'untapped' portion of solar power

Does it happen that the dashboard shows value of 'solar pv' wattage significantly higher than your loads + battery charge ?

And yes, I had seen your video about the wifi remotely controlled breaker. Very convenient and I'm currently thinking about something like this for my own system

Thanks again

[Mitker]

and I do agree with your appreciation of the ShinePhone App
I never had battery connected, as confirmed on the inverter screen info, but the app keeps telling me how well my batteries are currently charging (?!)

Useless
 

[JBChiangRai]

It's not normally possible to tell the unused power available for generation.  This is because the inverter doesn't know it, it can only know what the maximum power can be by trying to max out the available power.  In normal use, you're not using the available power and there is no way the inverter can know what that is.

 

What I have done is max out the inverter (by grid export) and graph it by 30 minute intervals.  If the sun is shining, I know roughly what is available from my chart before drawing power from the grid.

[Bandersnatch]

36 minutes ago, Mitker said:

My only uncertainty it that this value is very close (by a couple of %) to the sum of the loads and the battery charge.
Would this be due that by chance that day your use was that close from what the system was generating, or that what appears under 'solar PV' is just the sum of the uses plus some losses and roundings ? In the later case, it wouldn't inform about the 'untapped' portion of solar power

 

I agree with what @JBChiangRai posted above. Any unused solar production from my off-grid solar setup is curtailed.

 

The screen shots I posted were early this morning  - a very dull day here in Surin. 

 

Here are 2 more screen shots that I have just taken, one with one of my EVs charging the other without.

 

 

 

[lom]

3 hours ago, Mitker said:

ps: definitely not a geek in electronics, but I should be able to write down and inject some basic code on a board that would process the data and run a few relays accordingly

Then install a CT (current transformer) to such board and measure the imported power from grid and let that regulate how many of those additional consumers you enable. Basically add consumers until there is an import from the grid and then remove until import disappear.

[lom]

Shopee has a kit for measuring up to 100A, communicating over RS485-ModBus

https://shopee.co.th/product/274610962/24942397874

 

 

 

[Mitker]

3 hours ago, JBChiangRai said:

It's not normally possible to tell the unused power available for generation.  This is because the inverter doesn't know it, it can only know what the maximum power can be by trying to max out the available power.  In normal use, you're not using the available power and there is no way the inverter can know what that is.

 

What I have done is max out the inverter (by grid export) and graph it by 30 minute intervals.  If the sun is shining, I know roughly what is available from my chart before drawing power from the grid.

That's an interesting approach I didn't think about.
Thanks

By modelling a theoretical curve of the power generated by the system in sunny conditions on one side (as you did), and by comparing it with the data provided by a small panel as an indicator of the real conditions on the other side, maybe that combining both information would make it possible to estimate quite precisely the total power available at a given moment

The next step would be to deduct the loads at that moment to have an idea of what's left available

[Mitker]

2 hours ago, Bandersnatch said:

 

I agree with what @JBChiangRai posted above. Any unused solar production from my off-grid solar setup is curtailed.

 

The screen shots I posted were early this morning  - a very dull day here in Surin. 

 

Here are 2 more screen shots that I have just taken, one with one of my EVs charging the other without.

 

 

 

Thanks for these

They seem to confirm what @JBChiangRai was explaining about the extra potential production being unknown by the inverter
(my hope was that, by having to 'discard' this excess some way, it would be able to figure it out..)

This being said, I'll consider the Solar Assistant app for the other features
 

[Mitker]

1 hour ago, lom said:

Then install a CT (current transformer) to such board and measure the imported power from grid and let that regulate how many of those additional consumers you enable. Basically add consumers until there is an import from the grid and then remove until import disappear.

In the long term I would rather avoid this empirical approach as I mentioned, but it may come handy in order to establish this theoretical generation curve as a start

[Mitker]

18 minutes ago, lom said:

Shopee has a kit for measuring up to 100A, communicating over RS485-ModBus

https://shopee.co.th/product/274610962/24942397874

 

 

 

Thanks for this 😉

[Fruit Trader]

3 hours ago, JBChiangRai said:

It's not normally possible to tell the unused power available for generation.  This is because the inverter doesn't know it, it can only know what the maximum power can be by trying to max out the available power.  In normal use, you're not using the available power and there is no way the inverter can know what that is.

 

What I have done is max out the inverter (by grid export) and graph it by 30 minute intervals.  If the sun is shining, I know roughly what is available from my chart before drawing power from the grid.

 

That's correct and has been discussed here before. An independent reference is required to determine potential power not being used. 

[Bandersnatch]

36 minutes ago, Mitker said:

The next step would be to deduct the loads at that moment to have an idea of what's left available

 

Couple more of my videos for you. I am now only buying devices that work with home assistant 

 

 

 

 

 

[Crossy]

As an aid to prediction how about solcast?

 

 

You'll still need some form of real-time power monitoring to control your loads mind.

 

[JBChiangRai]

Assuming you can tell the amount of power your inverter is producing at any one time and you're running hybrid with PEA topping up any shortfall.

 

Then every 30 minutes max out the power used by running more than your solar can produce and record the output from your solar system.

 

That will give you the kind of data I have to work from.

[Mitker]

Sorry for this long silence. I've read all your posts, have been looking for info on the web and did a little of processing of all that.

I'll post here what's my plan at this stage and, if the topic doesn't get closed due to low activity, I'll post later how it takes shape practically since I think it may be useful for other readers in the future facing the same question.

I made a simple flow chart (attached) to illustrate how I see it after checking the leads you gave me - thanks a lot to all of you for these !

It starts obviously by reading the current supplied by the grid to the inverter, since any positive reading means all the PV supply is used up already : case closed
I'll install a reading device (like these @lom and @Bandersnatch mention) to retrieve the info.

Even if at first I rejected the idea, I'll go the way of empirically testing the amount of unused solar power the panels can supply at a given time by progressively turning on a set of power-using devices (cfr @JBChiangRai and @lom posts)

The data processing and control of the devices will be managed by an Arduino board.


... and in a later stage :

In order to optimize the scheme, there are thus a few more options:

- establishing a reference curve (gathered from real conditions readings) of the maximum potential output of the arrays in full sunny conditions and exploiting it to improve the model. cfr @JBChiangRai approach

- since using only the real time unused power can lead to constant switching on and off of the devices, estimating the general weather conditions of the moment would be useful to add to the data to be taken into account. 
It could be an app (like the Solcast @Crossy mentions) or it could be a hardware setup reading the luminosity in a cone - like 45° wide - around the sun position. That way, the board can smoothen the cycles based on objective conditions.
The first option means being able to extract the infos from the app to be exploitable by the processor. The second means getting my tools ready.

- I've also been told that the PV voltage reading of the PV supply could help estimate the unused power, the rationale being that it was a direct result of the inverter restricting the PV input to its current needs, and that the right equation would give the answer (arrays specs being the main part of it). 
The explanations I got far exceed my skills in the field so that I won't go that way...

Thanks again 😉