Recently I started to give the ESS grid setpoint an offset equal to the total power of EV charger so it does not drain my battery. This works, up to a point where ESS is doing some unexpected behavior which in my opinion is a bug.
Setup :
3phase grid (3x25 A)
Cerbo GX
Multiplus II 6k5 on L1 with input limit set to 32A
3 phase EV charging station on AC-IN
ESS enabled
Peakshaving enabled (normally 25A, but also tested 15A)
Venus OS 3.70
Situation
EV charger draws 10A on each phase –> 7200W
Therefore ESS grid setpoint set to 7200W
Expected behavior: ESS will balance the total grid power at 7200W. There is nothing to peakshave since all phase currents are 10A .
Actual behavior: Multiplus injects 1700W to keep the total grid power at ~5750W (25A*230V)
Maybe there is some limit programmed at 25A right? Let’s change the peakshave current to 15A.
Actual behavior: Multiplus injects 4200W to keep the total grid power at ~3450W (15A*230V)
If the peakshave current is set to 40A everything works fine
This situation is quite annoying, because it makes my system very unreliable when loads are present on L2 and L3. If the charger is connected the battery cannot be used anymore, unless I sacrifice the peakshaving function. Which I don’t want, since it is especially needed when battery and EV charger are working simultaneously.
Do I miss something here or is this a (known) bug?
Just tried the 3.80~4 beta, but same issue. I am also not sure whether the issue is in the peak shaving function or in the “balance the grid setpoint” function.
Unfortunately this has not been resovled by 3.71~3 (or 3.71). If I disable peakshave or increase peakshave current it works immediately.
I see that the control window of a single phase unit with peakshave current of ±25A is ±5750W. In single phase operation it should also not be possible to configure grid setpoints beyond this power window. But in multiphase regulation the grid setpoint window should be ±17.250W since the other phases can contribute equally.
It would be nice if someone of Victron could reply if the above interpretation is correct or if there are fundamental limits to respect.
Do you have a 3-phase grid meter?
If you don’t it won’t be metering the other phases.
You are aware that multi’s max output rating is 6kW at 25C?
iirc feeding back is less efficient.
Also, what grid code are you using?
Yes, I do have a 3-phase grid meter. So, if my EV is charging on 3 phase with 8kW it will be shown on the grid meter in VRM/Venus OS.
I am aware of the power limitations of the multiplus, but the limit is not reached. Let’s assume a situation where another load of 500W is switched on while the EV is charging with 8kW and the grid setpoint is also 8kW. In this case the multiplus only needs to inject 500W. Instead, it refuses a grid load of 8kW in the first place by compensating it always back to 5750W.
I tested recently with Venus OS 3.80~7, but the issue remains. Also tested with different P1 dbus meter instead of Virtual Grid meter, but also here the same result. @jeroen sorry to tag you directly, but you must be able to give some clarity here. Thanks!
I am aware of that, but it would be nice if someone can confirm this is a hardcoded limit / implementation or could be related to the meter type. It’s not only the meter I need to buy, but also an extension of the distribution board because the current one won’t fit any more meters.
And to be honest I have a hard time believing it is related to the meter, because the whole multiphase regulation works perfectly fine, until the single phase peak shave power limit is reached. To me it seems there is still a single phase limit in place whilst it is operating in multiphase (with a single unit). Again, if someone can confirm it’s a bug, or it’s by design I don’t have to look further for additional hardware or other solutions.
I am not sure you will get that level of response.
Does it work fine on 3.67?
Some grid codes limit grid use, the only way to verify that is to change it to something like “other”.
There absolutely was a peak shaving issue introduced from 3.70 but this is resolved in 3.71.
There is a part of the algorithm that prevents setting an AC power setpoint so great that it would exceed the total power we can draw from the grid. That is to prevent people blowing their fuses by setting a very large positive AC power setpoint, but also for users of DynamicEss, so that if the system is incorrectly set up and instructed to charge the batteries at a very high rate, there should be a safeguard against it. It should not be possible to set an AC power setpoint of 7.2kW if the max we can draw is 5.7kW.
This is where it goes wrong though. The max is not 5.7kW.
Since you only have one Multi on L1, the total power we can get from the grid is calculated as 25 * 230 = 5.7kW, but this is incorrect, because technically we can get three times that from the grid (we just cannot control the other two phases).
So functionally your system is ignoring an AC power setpoint over 5.7kW. This is again because, unfortunately, systems with only a Multi on L1 wasn’t fully supported for peak shaving, and even now we’re still finding small issues like these. We even considered just not supporting it at all.
Thanks for reporting it, I cannot promise that it will be easy to fix, but I know where to look now.
Thank you for the explanation, this all makes sense. I understand the implementation of peakshaving function of a single unit in 3phase system is tricky and has a lot of dependencies and pitfalls. At the same time, the peakshaving function is quite essential for me so I would appreciate a quick message as soon as you have more clarity on a fix.
Venus OS v3.72~3 has fixed this issue . It allows now for a high grid setpoint and the right screenshot shows it nicely compensates for excess load. I also tested peakshaving on L1 when a ‘unreachable’ high grid setpoint is given, this also works nicely as can be seen on left screenshot.
. It allows now for a high grid setpoint and the right screenshot shows it nicely compensates for excess load. I also tested peakshaving on L1 when a ‘unreachable’ high grid setpoint is given, this also works nicely as can be seen on left screenshot.