Last year I installed a Victron system including a MultiPlus II 15000/48, an MPPT 450/200 and a Cerbo GX, together with 9x Pylontech US5000. The inverter is configured in ESS mode with PowerAssist, zero feed-in and Spain’s grid code.
After some time, I realized that the inverter had an erratic behavior, where it constantly starts inverting, then stops and starts again, causing sudden current spikes that make lights flicker. Such behavior can be observed in the following video:
The problem mainly occurs in the following transition points:
Early in the morning, when solar yield starts
Mid morning, when battery reaches 100% SoC
Late afternoon, when solar yield approximately matches consumption
When minimum SoC has been reached and the Multiplus switches to passthrough (mainly during winter months)
When minimum SoC has been surpassed and the Multiplus switches to inverter (also mainly during winter months)
The initial thought was that a weak grid was prompting the problem, as I stated months ago in a previous post. However, after doubling the section of the wire from the meter to the main board and ensuring the Neutral-Earth bond was in good shape, we reached a point where the grid is decently solid, but the problem still persists.
Currently I’m more inclined towards a DC problem, maybe something in a control loop that can’t for some reason reach a stability point. I have been documenting some time frames belonging to different instances of the problem occurring, for example this one during the morning:
As additional information, while the problem is occuring, disabling zero feed-in does not change anything. Changing ESS mode from “optimized” to "keep batteries charged has also zero effect. Changing the grid setpoint has also no effect.
I’m trying to figure out what is happening, as I haven’t found anyone online suffering from the same issue.
Its a voltage change (between grid synchronisation and inverting voltage) and harmonic problem.
Listening to your video, last time i heard that we got the power provider to trim all trees touching the overhead lines; clean up the connections for the clients feed to the overhead lines and magically the issue disappeared.
It probably is a long line, however, I refuse to think that the problem comes solely from the grid.
I do agree that the grid is relatively weak and may amplify the symptom, but the event seems strongly correlated with DC-side control: 99% SOC, MPPT in External Control, VE.Bus oscillating between Absorption and Passthru, and Pylontech MaxChargeCurrent observed dropping to 60 A during the event. Island mode also stabilizes it. So I suspect the weak grid is a contributing factor, but not the root cause.
Switch off power assist then for now. And set your ac input limit lower.
Confirm you dont have a grid meter or loads before the inverter?
Do you have solar?
I will turn off PowerAssist, although I already set the PowerBoost factor to the minimum possible. My AC input limit is currently set to 40A, which is the maximum power I have contracted with the utility company (not the size of my main breaker).
I do not have a grid meter nor loads before the inverter, and I’m currently using the internal multiplus CT.
I do have solar, 32 530Wp panels connected to the 4 MPPTs of the RS 450/200.
Is this in ve config or the power assist on the GX? If you are using ess assistant this is usually disabled if you are referring to the one in ve config.
I was referring to the PowerAssist setting in VEConfigure. I currently have PowerAssist enabled with the minimum PowerBoost factor (0.3), but as you say, I understand that with ESS enabled the assistant largely takes over the control logic.
The configuration file says it was off, however I have just updated the Multi’s firmware and reconfigured it again from scratch following Victron’s Pylontech guide.
I still believe there might be something off about the DC side of the installation, if you look closely at the scenarios when it occurs:
Early in the morning, when solar yield starts
Mid morning, when battery reaches 100% SoC
Late afternoon, when solar yield approximately matches consumption
When minimum SoC has been reached and the Multiplus switches to passthrough (mainly during winter months)
When minimum SoC has been surpassed and the Multiplus switches to inverter (also mainly during winter months)
You can see that in all of them the common factor is a change in the sign of the current from the battery (despite it being from drawing to charging or vice versa).
I am not a fan of the stacking pylontec allows and here is why.
15kVa takes 2x95mm² cable (so 190mm²)
You have 3 sets of cable 25mm² from the bank. (75mm²). individual connections here would be better. (9x25 = 225mm² so larger than the 190)
It is supposed to draw what it needs through less than half of the cable diameter? It is less about the voltage drop now and more about speed of saturatio etc.
@Lcs_26 As an aside, I noticed the one ve.can 2 port is not terminated. Your MPPTs on VE.Can require termination at both ends of the serial string. With the Pylontech’s termination is at the Cerbo end only as the batteries terminate.
