I think if someone constantly believes the system is making wrong decisions and that I, as the operator, could do better, then I would deactivate the system and control it myself. At least, that’s what I would do.
1 Like
Actually I use real world settings for dess but it still makes strange decisions sometimes which is confirmed by Victron professionals.
The issues are minor but they are there and they are getting improved by the victron team.
Keeping them up to date with my experiences will improve dess even more like my previous issues did already lead to improvements to make DESS better.
Fiy my RTE is 77% and battery costs €0,02/kWh.
Many thanks to the Victron team for improving Dess.
The energy buy and sell price does not seem to be set correct from your screenshots.
With 77% efficiency I almost would say your system setup is too small for the purpose you are using it for.
1 Like
The wrong calculated Target SOC (rounding error) from 8 pass until 15 pass, is the only annoying thing in DESS.
Now when the kWh price is at the highest moment on the day, DESS decides to reduce the discharge from 8,2kW to 2,5kW.
Please victor fix this!
Buy and sell price are set correctly but already set without “saldering“.
In one of the latest beta’s the rounding error has been solved. In future versions these values will have decimals I was told by a victron developer.
1 Like
That’s good news!
Did they told the beta number, tried them all.
Or the next beta will solve it?
No it’s already solved. Maybe you encounter a different issue. The issue I experienced was that 100% soc was not reached because target soc was for example 99.9% which was converted to an integer becoming 99%.
Than this is another issue 
I have often from 8 to 15 minutes after the hour a SOC that jumps 1% to the incorrect direction.
In discharge you see then a positive 1% jump in target SOC instead of none jump or -1%.
And in charge just opposite, looks like they incorrect add or subtract 1% to make a hysteresis.
From 8 until 15 minutes pass the hour /Settings/DynamicEss/Schedule/0/Soc is the active target SOC.
From the full hour until 8 pass /Settings/DynamicEss/Schedule/4/Soc is the active target SOC.
See the jumps in the picture.
Yes, think this is a common issue!
It has a huge effect on the discharge current.
Perhaps the price setting influences DESS behavior a bit different than you expect it to do and which of both are correct, the expectations or DESS? From the screenshots there is no insight of the SOC so its hard to tell why PV did go into grid directly.
I would pick the moment of dissatisfaction, take a detailed snapshot of what i had expected and what DESS decided to do and mirror it against the price deducted with the efficiency and take the forecast, SOC and the power the system has to move electrons around into account. Some trade logic is reflected in what the system actually does and perhaps mirroring this against the expectation could clarify the taken snapshot.
In my case reflecting it this way cleared some unhappiness. Yes, i would have liked a setting that would allow a threshold for trading on the difference between buy and sell, but that is not implemented. Playing with prices just makes DESS behave different than we wish or would expect.
Perhaps in your case DESS is doing a completely unlogic thing and you can share the details of the reflection to the dev team.
As an example, i had set the system efficiency to 79% initially because everybody is yelling it.
But i noticed that a lot of PV this go into Grid at moments i did not expect it and reflected this all.
Checked the real system efficiency and noticed that it was higher with my loads.
Import 94% and Export 94% made an overall efficiency of 88% and set that and my PV problem was solved.
Picture from yesterday morning.
At 15,3% (think 15,25%) SOC target SOC starts to jump from 16% to 15%.
15% target SOC can only be seen at 8 until 15 pass, this is not a feature but a bug!
I agree this is at least bad behavior. This has been discussed before to be related to the INT limitation of the D-BUS scheduled SoC values. There are messages from Victron that they are working on that but I have not seen any concrete results nor any indication of a timeline thereof. For our own system I have build a flow that constantly monitors what DESS is doing and then manipulates the D-BUS SoC values to be rounded more up or more down depending DESS is buying or selling. It is, yet again, a hack. But it works.
Why would fixing a simple thing like this takes over 6 weeks?
They have enough time to add gimmicks like graphical interfaces and Shelly devices.
Would say invest this time to make the system bug free!
This seems to happen because approximately 5 minutes after the the hour a new schedule arrives. The target soc during this 5 minutes was calculated an hour before when the previous schedule was received and is to high causing a charge.
When charging this also occurs, 5 minutes of discharging at the beginning of the hour is sometimes happening because of an old schedule.
This issue is known to the Victron and it will be solved they said.
Correct from 0 until around 8 pass /Settings/DynamicEss/Schedule/4/Soc is the active target SOC.
At around 8 pass the system has contact with the server and the schedule is refreshed and the active target SOC will be /Settings/DynamicEss/Schedule/0/Soc
And in the new calculated schedule /Settings/DynamicEss/Schedule/0/Soc is sometimes wrong.
Fix would be to store old Schedule/4/Soc and reprogram Schedule/0/Soc if the value doesn’t match old Schedule/4/Soc.
But how hard can it be to fix this in DESS !!!
Here is a nice graph showing the issue during charge.
After the hour a wrong tsoc is used for around 5 minutes until the new schedule with correct tsoc arrives causing a discharge while it should charge.
And here I saw a scheduler issue , at17:08, the new schedule arrives but the target soc is the same as the current soc and higher than the previous target soc scheduled, so discharging stops for a while:
And here the target soc is lower than the actual soc, causing unwanted discharge for a short while until the target soc matches the actual soc again:
And here the actual soc jumps to 100% (battery is has reached 100% soc voltage) and because the target soc lags, it discharges:
But as I said, this is a know issue to the Victron Team and they are working on it so we just have to wait for a solution.
Correct, the new calculated /Settings/DynamicEss/Schedule/0/Soc is sometimes incorrect.
And /Settings/DynamicEss/Schedule/0/Soc is the active target SOC after the schedule is updated.
True, because the new /Settings/DynamicEss/Schedule/0/Soc is based on an older actual soc causing a to low or to high tsoc. Also there are no bat2grid or grid2bat restrictions in place, so that’s also not preventing unwanted charging/discharging.
This is also under investigation at the Victron Team.
We just have to wait for a solution.