Issue with ESS after upgrading

Q&A and troubleshooting
, ,
1

I am probably a bit late to the party on this, and maybe this is the wrong thread but I have noticed an odd behaviour with my ESS system after updating to V3.67:

The system has been in use for nearly 10 years on the same Quattro inverter, at the end of December the old CCGX was replaced with a new Ekrano unit as well as a new and much larger battery (3x16kWh Soluna units).

  • So the summary of the installation is:

    Single phase 230v in the UK

    Quattro 8000 firmware V497

    MPPT150/40 with 2,2kw of solar PV

    8.6kW of PV on old SMA inverters on input of Quattro metered via a Gavazzi EM24 meter on Phase 2

    Ekrano GX running standard firmware (not large, not using node red)

    3x16kWh Soluna batteries

With the firmware that was in the Ekrano as supplied (3.55) everything commissioned fine, at some point I upgraded to 3.67 and everything still seemed to work as expected.

Today I was taking a look at the dynamic ESS option (using the VRM implementation) and noticed that even though the battery was at nearly 100% SOC only about half of the load was being met from the battery:

Firmware 3.67 example one
Firmware 3.67 example one1127×478 36.3 KB

The load here is a constant resistive heating load, and whilst there were small changes to the inverter output the majority of the energy was coming from the grid for many minutes. With an increase in load the inverter output increased but was still short of the load.

Firmware 3.67 higher load example
Firmware 3.67 higher load example1096×515 35 KB

I looked through all the ESS settings but nothing had changed that was obvious (although the new firmware menu options are different so I may have missed something) The battery BMS was advertising a maximum discharge current of 660A so there was no limit from the BMS.

I admit to being utterly baffled and spent a lot of time looking on the forum for other users who may have had this, but similar issues normally related to current or power limits (and sometimes Dynamic ESS settings) but I was using “normal” ESS.

Getting desperate I rebooted with firmware V3.55 (the version in the Ekrano memory so no need to download) and as soon as the system rebooted it worked normally again, fully meeting the loads from the battery:

Firmware 3.55 example
Firmware 3.55 example1012×477 33 KB

I have no idea why this fixed it, is it a bug in 3.67 with my system or was a reboot needed and the version of the ESS is not critical. For now I am happy with V3.55 and the system is working as intended but am curious if there is any input as to why this was happening.

If you need any more information let me know and I will try and help

Nick

Venus OS v3.67~2 available for public testing
2

@NickJenkey Have a look at the latest beta and see if this may be a contributing factor?

“Peak Shaving, fix two bugs related to peak shaving while Battery Use is set to “Output loads”

  • One, under these specific conditions peak shaving did not work correctly:
    System is configured to feed excess DC PV into the grid
    System is charging
    Battery use is set to output loads only

  • Two, under these conditions a large positive AC Power setpoint could overrule peak shaving:
    System is in self-consumption mode, OR charging but not configured to feed in excess DC PV
    Battery use is set to output loads only”

3

Interesting, I have “DC feed excess to grid” set to on as at this time of year we use any surplus for heating water, it is turned off with the radio button in the summer.
I note in your reply @ejrossouw the post refers to peak shaving which is not a feature I use, and is disabled in the system so not sure how that affects it?

Since posting yesterday everything has worked as expected with firmware V3.55, I am happy that for now the system is operating as expected and keeping the grid import to a minimum, I will try upgrading at some point to a later a version and see what happens but I don’t really want to break anything at the moment!

Thanks

Nick

4

An update on the issues above:

I upgraded the ESS assistant in the Quattro as the version that was in there was about 4 years old (still leaving the firmware at 497) and checked all the settings were correct.

After that I re-installed Venus 3.67 and there seemed to be no issues immediately.

I have been watching the system quite carefully over the last few days and the issue seems to recur intermittently, there are times when the amount of power being drawn from the battery is significantly less than the load for significant periods, not just a short term adjustment but a constant import from the grid of around 50% of the load.

This will be for maybe 20 minutes or so and gradually the power from the batteries is increased until the system is behaving “normally” with fluctuations of import and export.

I even got a video of this as it transitioned from “half grid” to more “normal” behaviour

There seems to be no obvious trigger for the behavior, it just seems to occur periodically, persist for maybe 20 or 30 minutes and gradually resolve.

I was going to look at the “Advanced” graphs for the times it does it but those are currently not working on the VRM for this system

Any thoughts would be appreciated

Nick

5

The VRM data is back so I was able to look at the data for the same period as the video in my last post:

image
image672×1133 67.2 KB

As there graphs show there is the “normal” behavious of the system up to about 15:50 and then the draw from the grid increases significantly gradually rising over the next hour to a peak around 1645 before falling back to “normal” around 17:20 there are a couple of big peaks after that but only for a short while and then the system settles down again.

The behaviour is happening at a time when the load is high but is not correlated directly:

image
image723×1075 51.6 KB

In this zoom the load is a consistent 4kw for most of the period but at the begining of the period (16:50) the grid is supplying most of the load and by 17:15 the load is met from the batteries again. The big increase to 7kW when the kettle went on is mostly met from the batteries and after the 4kw baseload is met from the batteries almost entierly.

There is some correlation with the end of the solar generation for the day:

image
image661×1150 67.9 KB

As the solar output falls off towards dark the grid consumption increases, once there is no solar output at all the batteries meet all the load. Is the fluctuating PV output on the AC inverters as the light fails somehow affecting the control loop for the ESS?

The problem definitely seems to be happening as the AC solar output hunts for power at dusk.

6

@NickJenkey
With all the under the hood changes Update the inverter firmware and reprogramming is tha answer.
You can’t partially update the components and wonder why its not cohesive.

7

A fair point, I will upgrade the inverter firmware when I get a chance, I was following the general advice from Victron to leave alone if its working (which to be fair it basically is!) but We will see if it solves the issue.

8

Lol. But you didn’t leave it alone. :laughing:

There are only three reasons to update.
You clearly wanted a new feature. It is one of the three.

9

I have tried to illustrate a technical issue that I noticed, and which on further investigation seems related to low power PV output towards dusk, rather than the Venus OS version as first thought.

The update happened originally because I had had auto updates on the Ekrano, so the update originally happened automatically, not because I wanted a new feature in particular.

I will upgrade the firmware on the Quattro as per my previous post when I am able to do so, and then see if the issue improves.

In the mean time I would appreciate not getting sarcastic answers from you @lxonline your reply to my post is an unnecessary flame and does not help me or anyone else resolve issues.

The rules of this forum are clear: to be polite and professional not mock other users because you feel you can.

10

I think you are misinterpreting a tongue in cheek comment that was intended to be lighthearted.
Forum dwellers do need to have a slightly thicker skin, but it is helpful not to interpret directness as rudeness or mockery, especially when it is directed at volunteers who try help as many as possible and the lack of time makes for shorter and sharper comments.

11

It was not my intention. You mentioned giving DESS a try. So I assumed it was something you were interested in looking into. New feature…new interests.

It followed this comment. Which to some degree- as you attribute the action to yourself- implies intent on your part.

DESS was introduced may 2024 so ideally firmware from v551 would be expected on a system.

As an FYI dess decides not to use the battery if the cost of energy from the grid is cheaper than 1kW from the battery. It is about profits. And some normal ess settings can’t be used with dess.

Either way if there are still weird behaviours, you still may need to upgrade as tou are now troubleshooting. Rollbacks don’t alway work as expected.

12

I also see an Issue.

Trading with the low differences on the day ahead over last month is very low.
Did change the battery cost /kWh from 5 cents to 1 cent but even this doesn’t make a big effect.
Could it be that the 5 cent battery cost /kWh is still active?

13

Hi Ron, I am not sure I fully understand your point, I think you may be assuming that we are on some kind of flexible tariff with half hour metering? the system is in the UK on an “economy 7” tariff which gives 7 hours of electricity between 0000 and 0700 at less than half the daytime rate (14.03p/kWh compared to 33.1p/kWh)

The differential between overnight kWh and daytime kWh on my tariff is this £0.19 which is significantly higher than the battery cost per cycle.

The battery is effectively approx 40kWh if discharged to 20% SoC (48kWh nominal) the lifecycle costs based on 8000 cycles (warranty is 10k cycles) would be £0.71 per cycle or £0.017pence per kWh if we round this up to £0.02 per kWh and allow for some losses in conversion the cost per round trip kWh is approximately £0.04p which means that the saving in using battery storage to cover daytime use in the winter is around £0.14 per kWh which is the reason for using the overnight charge.

Also I am not using DESS I am using normal ESS with a scheduled overnight charge so any price set per kWh should not affect the operation unless I am missing some aspect of how the ESS works here.

I am about to update the inverter firmware as suggested so will see if the same issue happens around dusk today

14

@lxonline Sorry to have misinterpreted any comments

I am going to upgrade the firmware to the latest version in the quattro this afternoon and see if the behavior recurs:

I did have a look at DESS out of interest and the graphs are very nice but as in my reply to @ron_h below I am not using it, as the tariff we are on probably does not warrant it as there is such a big difference between day and night rates and the electricity consumption here varies hugely day to day. Some days it is just office work others it is large amounts of energy going into ovens and machinery so from the reading of the manual for DESS it sounds like it would not work as well because of this.

I use a series of scheduled charges depending on the time of year and a bit of weather watching, as well as consideration of the next days activities.

Being this far north (53 degrees) and in the mountains makes for very dark winters!

15

Everything is now on the latest firmware.

image
image1029×219 21.1 KB

I will see if the behaviour repeats around dusk tomorrow the, system seems to be behaving “normally” at the moment but it is nearly dark and the firmware update was done around the time the issues happened yesterday, it took a bit longer than expected as I had to go looking for my grid code password which I had put somewhere “safe”

Anyway we shall see if the strange behaviour repeats tomorrow and hopefully it was just an interaction between the newer Venus OS and the older Quattro firmware

Thanks for the input everyone and apologies for any confusion over the “looking at DESS” it was literally looking at the options and trying it briefly whichj meant I was on the VRM page for longer than usual and thus saw the behaviour.

16

Dess isn’t usable for you, scheduled charge and for the rest of the day zero grid load is the best you can do.
Hope after the update your issues are gone!

I would expect Dess to buy or sell if there is more than one cent difference (did make battery cost one cent).
But it looks like the previous 5 cent battery cost is still somewhere stored,…
Can Victron look at this?
Looks like in the new VRM environment the battery costs in the Dynamic ESS setup isn’t working!
I run the latest Venus OS V3.70~92
Or how can I check the battery costs settings?

17

When you programmed your battery bank value on the vrm?

I am also an avid user of sheduled charge.

DESS is not applicable where we are as our tarrifs are usage based. More kilowatts total equals more cost. Over 400Kw per month is a 48c per unit. Less than 50 units is 4c.

18

Yes in VRM under settings → Dynamic ESS → Battery → Battery costs.
Seems the old 5 cent is still active.
Because I see only buy/ sell activity when there is more than 5 cent difference, and not the 1 cent how should now be the threshold.

I have a dynamic contract at Zonneplan, and buy and sell price for every 60 minutes is the same so there are none costs to add onto the battery costs how do affect trading.