Hi, I currently have three SMA Sunny Island inverters (in off grid)with two Fronius Primo AC coupled PV inverters. I have made some frequency monitoryng devices that I use to control a SSR and dump excess power to heaters proportionaly acording to the rise in frequency due to frequency shift.
I am now switching to some Victron Quattro, will still use my two fronius primo inverters with frequency shift, but I will also add some more solar panels with a Smartsolar MPPT 250/100. Is there a way to use excess solar from the MPPT controller as with frequency shift?
What I like of frequency shift to dump power is that I do not need any other wiring to control. I can just plug into any AC outlet and connect my load. Plus, I am providing electricity to around 8 buildings, so I can dump loads into space or water heaters in any building.
Will frequency also rise to dump the PV energy of the MPPT to the AC loads? or will it just stop generating?
Frequency shiting is used to throttle AC connected Inverters. MPPTs are throttled to reduce their output through the bus/VE Direct connection - whenever feed in is disabled, and batteries are full.
Since you use the frequency shitfting to generate âmore loadâ, you are basically hooking into that process at a very late time - when the System starts to attempt to reduce production.
With my system, I noted that the System prefers to throttle AC-connected PV and completely disable it, before it starts to throttle the MPPTs. Or to be more precice: When there is more power required, than the system currently produces, MPPTs are spinned up first, over re-enabling AC-connected stuff.
So, an educated guess would be: If you start to generate AC-demand, while the system is trying to reduce itâs generated power, it will try to keep the loads powered from the mppts, and only increase AC-coupled power output (by reducing the frequency again), when the power generated by the mppts is not sufficent.
However, it may be better if you control your consumers in a different way - to generate power demand, before the system is starting to throttle. I could imagine that this will lead to a unstable up/down otherwise:
Your consumers âstart consumptionâ cause frequency is increased. more power is required, frequency is lowered. Your consumers therefore turn off, frequency is raised again âŚ
Thank you for your information, it is very usefull.
I actually manage to increase AC loads before it starts to throttle the AC couplede inverters. You can manage this with a wider range of frequency shift. Otherwise when you turn on the load, frequency will drop and load will disconnect. This graph shows how I load my dumploads according to increase in frequency and how PV generation is reducen with frequency.
This works with my SMA Sunny island because frequency is increased progresivly. It doesnât jump directly to 51Hz to start to throttle. So by when you reach 51Hz and PV is starting to throttle, all dump loads are already at max power. I imagine frequency will increase progresivly on the Victron Quattro too. But I can easy tweek this on my PV inverter and my dump load controller too.
I was just trying to find a similar way to use the MPPT excess power without having to be wired to the aux relay since inverters are very far from where loads usually are.
You could use node red and smart switches. Or smart switches driving contactors if they are bigger loads. Since you can use information for either device.
It will in theory change the frequency since now there is no need to throttle. So to trigger an on condition is ok (maybe soc is better) maybe use other triggers to stop the load instead.
Kia ora @Matthew did you manage to get the Quattro to raise the AC frequency gradually, the same way the Sunny Island did? I am looking for something similar.
I have a python script / toolbox which contains a solar overhead distributor, so an infinite amount of consumers could request power (through mqtt) and the distributor will set allowances, based on priority, available overhead, stepsize etc.
For my use case this works perfectly by looking at grid feed to determine available overhead - but this apparently does not work offgrid, or with 0 feedin setups.
So still looking for alternative ways of knowing when consumption requests can be approved and when they have to be revoked.
Hi @spicedreams, I live in quite a far place, so I do not have them installed yet. They have just arrived but will be installing in one or two weeks. So will let you know then.
Does yours not increase in a progressive way?
@Matthew it looks to me (based on the graphs of Hz coming out of VRM) that the EasySolar (==Multiplus II) raises the frequency to the top of the PV-Assistant set range immediately when the battery opens the charge relay because it reached 100%. I do see smaller âblipsâ in frequency occasionally, but donât know what explains them.
That looks quite interesting. I will be installing my Quattros in a few more days and will start to get into that. For what I could find, I think it will have to be through MQTT or Modbus and find how to adapt it to an off grid
@spicedreams I will check when I install my Quattro inverters and will report back to you.
With my Sunny Islands it has been working great for a few years, but I had lead acid batteries. So charge current would decrease gradually and frequency increase gradualy too.
Now with my lithium batteries the Sunny Island has a software bug that when generator is off, it will set the charge voltage 1V less than what it receives from CAN bus from the BMS. So it will start throttling my PV inverters quite soon when my betteries hit 51V instead of 52V. So, being fixed at 51V charge current also starts to decrease gradualy. This is how my frequency looked for today for example (PV throttling starts at 51Hz as in the graph above)
When I have my new inverters installed should be able to charge to 52V at a constant current, so I imagine it will be more abrupt when I reach 52V and switches to float voltage. So will update here on how frequency and dump loads behave, but I still think it should work ok
Edit: I get these graphs polling through Modbus every 5 seconds. I do not remember how often VRM polling is. So maybe it is just a lack of resolution
)