I have system with 6 multiplus II 5000KVA in a 3 fase configuration with 90 kWh battery capacity and a 3x25A grid connection
I expected that during the price peak between 18.45 and 19.00 h my batteries will be discharged to the grid to the max a little bit under 25A per phase. However DESS discharged my batteries to the grid around 15A per phase. See picture below
And one hour later when the price was much lower between 19.45 - 20.00 h The batteries were discharged to the grid at the max (almost 25A) in a flat line
See picture below
In my opinion this behaviour is not logical en cost my money.
My question is: Is this a bug or intended behaviour, and if so why is this inteded behavior. From the viewpoint of a user is this far from logical
Or what can I do to change this behavior?
First thing that I notice is that your system isn’t running the latest Venus version (3.72 vs 3.79). Can you update to the latest version to make sure we aren’t going to chase bugs that have already been resolved?
I have 6 Multiplus 5000kVA ll in 3 phase configuration. The system is able to deliver 35A continue per phase. My grid connection is 3x25 A.
My expectation is that during the price peak the system will deliver 25 A per phase in a flat line. At least when the price is at the highest level.
What I observe is that in a quater when the price is higher than the quater before the current to the grid is lower.
As an example the price between 18.45-18.59h is lower than the price between 19.00-19.15h, but the current between 18.45-18.59 is higher.
The current between 18.45-18.59h is a flat line a little bit lower than 25A, but the current between 19.00-19.15 h is controlled to a lower level.
It seems also that the pattern is alternating. See picture below
I expect with a high price profile a flat line of the current to the grid at max level for some hours
It is quite a normal effect of the soc-non-linearity. DESS targets for reaching a certain SOC at the end of the schedule slot. Which results in a non linear current (depending a bit on where on the curve of the battery we are).
There is the logic. DESS tries to account for that as good as possible and work is being done on further improving that.
Basically you reach your target SOC earlier than expected by DESS, which is why the current tapers off. You can correct this by increasing the max discharge power in DESS settings. For example if you have a max discharge of 17kW at the moment, try increasing it to 19 or 20kW. You need to play around with it a bit, it helped on my system with a 156kWh battery.
Thanks for the feedback in your spare time. I appriciate that.
What I also did when this happened I looked to the target SOC. Maybe there is good reason for, but what I see is that the target SOC is too conservative during discharging when the prices are high and also the SOC is too conservative during charging when the prices are low.
Hopefully you can use this input for further improvement. And if you have something to test, I am willing to do that
The discharge power of a MultiPlus 5000 is approximately 4,000 watts. Across three phases, that means a maximum of 12 kW. So you need to adjust the formulas accordingly. There is no point using 20 kW in the formulas if the inverters can only deliver 12 kW.
Harold, so he has 6Ă— MP5000, 24 kW in total. The limiting factor in this setup is the 3Ă— 25 A mains connection. Naturally, he wants to make maximum use of it.
Yes, I had arrived at the same idea. I increased the battery capacity from 90 to 100 kWh because I felt that DESS was charging too conservatively and that the prices were already rising again before the battery was full. However, I suspect that too high a battery capacity has the opposite effect when discharging. In that case, DESS assumes that a given change in SOC represents more kWh than it actually does, and will therefore reach the target SOC sooner.
It would probably only work properly if you entered a higher battery capacity when charging and a lower battery capacity when discharging. But that isn’t practical to do manually.
That’s why I always use a higher discharge power than I have available in practice. This ensures that the target SOC is not reached during discharge, so the inverters continue discharging at maximum power for every full quarter-hour.
Thanks for the feedback. My maximum import and export power is set to 17 kW in DESS because of the 3Ă—25 A connection. I have my maximum charge and discharge power set to 24 kW. Are you suggesting that I set the discharge power to, say, 26 kW? Would you do the same for the charge power?
Ah, I hadn’t noticed the 6x5K either. In that case, the system parameters remain at 17 kW import and 17 kW export.
For the inverter, enter 24 kW discharge power and 22 kW charge power—that is, the AC power ratings, not the DC power ratings.
Then check the BMS data. Which BMS do you have, what are the charge/discharge parameters, and have you checked them against the timestamps?
The inverters can request a discharge current of up to 500 A, but if the BMS limits this, you’ll still run into problems.
And as Dirk-Jan already says, DESS is not a linear system.
There are 6 batteries of 15 kWh connected in parallel. The maximum charging current per battery is 180 A, and the maximum discharging current is 200 A. The charging current is 60 A.
For the battery pack as a whole, the maximum charging current is therefore 1,080 A, and the maximum discharging current is 1,200 A. The total charging current is 360 A.
I do not know the type of BMS used in the batteries. The above are the battery specifications.
Set your export and import capacity to 18 kW. You don’t need to worry about your main fuse going above 25 A as long as your peak shaving isn’t set too high. Keep in mind that when you set a limit of 25 A for peak shaving, Victron always maintains a safe margin of around 2 A, so in reality you are limiting the connection to approximately 23 A. You can of course resolve this by setting peak shaving to 26.5 A.
Hello,
Thanks for the information.
I have a Multi RS and saw the same behaviour yesterday, also due to a difference between the SOC target and actual SOC, which caused the power to drop during peak prices.
(This is visible in V3.80 and 3.79, by the way.)
I’ll have a play around with it and see whether I can influence it.
However, I would prefer a proper solution in Victron’s software, which is why I’m also reporting it, with supporting evidence, as input for product improvement.
