article

Hi, DESS is very frustrating. :))
7 days of trying it made me stop using it.
It has 2 rules it implements all the time that I consider counterproductive:
1. it delays the charging from solar until it is too late and it keeps charging from grid. Most of the time this comes over other surprise consumption - so it never gets the target SOC - hour by hour
2. the target SOC is never achieved in the hour. Even is most of the time less than ”I would do” it never gets ti it even there is plenty solar and no consumption to delay the charge. It does not start charging or it is charging at slow rates (1-4Ah)

This is today and before NOW I started charging it in ESS mode as it was clear it will never get to 100% even in the morning this was the ever to be miss target.

So, it charged many times with expensive grid power with a behaviour like ”UPS, I forgot to charge”.
My MP2 - 5000/48/70 - 5KWh bat - 4.3KW solar on AC-out1 with two 112 (grid and solar)

I did not receive an update today regarding the day ahead prices. Am I the only one?

It is 19:37 so the should have been there. Is there a way to actively pull them?

I'm sorry having to confirm some of the often mentioned issues and I turn off DESS for now. The biggest problem for me could be solved by a simple additional rule:

• never ever feed any power into grid if the battery still can take it!

This must be obeyed if “Can you sell energy back to the grid? No” is selected and also if you can sell, at least in Germany feed in rarely is a better deal than storage and self consumption.

Also annoying: if I switch on high loads like washing machine or heat pump for hot water intentionally in times with very low prices I don’t want them to be supplied from battery – which will be recharged afterwards at higher prices – at least if not very much solar is expected.

I’m sure the calculation of the optimal SOC for every hour with unpredictable conditions is not easy to do and I thank Victron Energy and Dirk-Jan Faber very much for this development!

Right now it seems the algorithm strictly optimizes to reduce cost, as exemplified here:

Right now, it really optmizes a single day for reducing cost and optimizing revenue: and of course that means, exhaust the battery to the minimum of 15% SOC at the end of the day. But that is not what I want. It misses the rest of the night (which it will probably only add 'in scope' once rates are out at around 2 pm) - but then it is too late to fully charge the battery. Also, I want the battery usually rather full instead of just charging the minimum to be better prepared for grid outages or occasionally higher consumption over night.

As such, I (as a house owner; boat, RV, commercial or industry could be different) would like to see my LFP battery to rather charge to 100% every day (NMC could be 80% charge goal in warm&sunny months), and charge slowly to extent battery life and reduce midday-feed-in-peak.

As such, the goal should be extended to:

Dynamic ESS is an algorithm that aims to minimise the costs made on the grid and battery, based on following expectations:

- aims to charge the battery every day to "Default Charging Limit" [LFP=100%, NMC warm&sunny=80%]

- Aims to charge the battery until after midday-peak to slow/steady battery charging for battery longevity, reduce midday-feed-in peak and reduce AC->DC conversion losses.

- Aims to hold the battery SOC level rather higher than lower, for situations of grid outages and unexpected (higher) consumptions.

Thanks for reading, Michael

Dear Victron team.

Frank energie/Bliq and zonneplan are offering ESS system from different brands which are compatible with trading on the “onbalans markt” this would also be a huge benefit if Victron can provide this function in cooperation with an energy supplier.

Hi, quick finding.. dyness ignores max charge and discharge at the moment.. anyone got the same?

Site c0619ab38e24

I tried Dynamic ESS on Beta a couple of months ago but as it kept exporting PV instead of charging my batteries, I switched it off. Today I updated my GX to the latest firmware, updated to the latest ESS assistant and re-enabled Dynamic ESS. I can't understand why, when I get paid nothing for export, DESS is exporting 1300W to the grid rather than using it to charge the battery. Yes, there is plenty of PV forecast for later in the day, but what if the forecast is wrong? Why waste all that power?

Even stranger is that the battery is lower than the target SoC, so I'd expect it to be charging right now. Unless 62% is where it wants to be at the end of the hour?

I'll persevere with it for a few days and see how things go, but when import prices are high all day and export is not paid, surely the goal should be to hit 100% battery charge as soon as possible? Then, if PV output is lower than forecast or loads are higher than forecast, there is some spare change in the bank, so to speak!

My last 7 days experience with DESS on Zonneplan with a 5KWh battery was very frustrating. I gave it a chance to be better every day, hopping. Unfortunately it sticks to some (bad) rules:
1. It sells to grid solar most of the day delaying battery charge. From the first day I disabled the sell price and battery cost hopping to start charging the battery sooner. Not a chance.
2. It starts the charge of the battery mostly in the last 2 of sun hours even if it is obvious it will not be able to charge to 100%. Even so, during those hours is charging at random speeds (1-3Ah at times, even idle) and at some point it goes on full throttle. Then as the price gets up and as soon as the sun goes down goes full charging at the high price from grid.
Like ”Uuups! I forgot to charge!”
Ex. A day with 5 hours at 18c and full solar charged 0.4 on solar all day then at the end of the day 3KWh from grid at 24c then 26c just to discharge them in the next hours at 26-28c.
My system: v3.40~1 MP2 48/5000/70 - 5KWh battery - 4.3KW solar - 2x 112 sensors one for the grid and one for the solar on AC-Out1

Today the initial plan was to sell at 0 cost solar to grid (!) and delay start charging around 13:00 and go 55% and so on, all the time less than the 2Kwh per hour limit I set. Now, after 15KWh of solar sold to grid it kept the battery at 34% until 15:00. It was obvious the battery will never be charged and I will pay the full price of 0.37 for the evening so I disabled DESS and start charging the battery on ESS mode.

Am I understanding something wrong?
DESS is not charging at the lowest price. It also shows solar to battery, even tho I don't have solar installed at the moment. I will keep monitoring the situation, but this doesn't seem right.

About the menu-option ‘Self-consumption from battery' with the choice ‘all system loads’ or ‘only critica loads’.

Wondering why an external energymeter is mandatory if the multiplus is connected between grid and all loads. The multiplus does measure all values, so why the need for an external energy meter before this option in the menu becomes available?

Some of the things on our todo list:

• Maximizing battery charge by the end of the day to cover consumption until sunrise

• Periodically fully charge the battery to improve battery life

• Adding support for varying service fees based on the time of day

• Allowing users to enter a fixed schedule for grid-battery restrictions

• Supporting fixed prices that change on the half-hour instead of the whole hour

• Introducing (smart) support for high-consumption loads such as EV charging and heat pumps

When can we expect this to be implemented into the VRM DESS? At least that possibility to decide if we want to prefer battery charging and dynamic prices schedules for entering different formulas to specific hours?

@Dirk-Jan Faber (Victron Energy) It seems that the system is not receiving any new price info anymore. It is after 18 o'clock and it is still only the info of today. We had this before (2 months ago?) and then you fixed it. Is this the same general issue, or is it just my system? (c0619ab2e756)

Subject: Using energy from PV inverter to charge battery

For the last two weeks i've been running Dynamic DSS. It works great! There is one thing though i was wondering.

I've got a 3kw PV inverter (via em24) and 3kw of PV chargers. When the sun shines and the dynamic price is low. The battery is being charged by the PV chargers. Occasionally the grid is used for additional charging. The Target SOC has not been above 80% although we had some sunny days recently.

Effectively the PV inverter is pumping into the grid (costing 2ct per kw provider profit) and occasionally the multiplus is drawing current (costing 2ct per kw provider profit)

If the multiplus would use the production of the PV inverter to charge the battery (or the fraction needed to reach the target) the energy would be consumed internally and not registered by the providers meter.

Efficiency and cost wise it should be prioritised like:

1) PV charger

2) PV inverter

3) grid

Is this the way it is implemented? if so, any reason that i'm not seeing this behaviour?

thanks!

ID c0619ab2e85a

Is it possible to add minutes to the Fixed Tariff? For example, Octopus Intelligent Go in the United Kingdom is from 2330 until 0530 for a low rate. Even 30-minute rates would be good! Looks good and I tried setting it up, but this stops me from using as the car will exhaust the battery ... Many thanks!

Dear Victron (Community),

I think the idea of Dynamic-ESS is great, but I'm having problems with the current implementation - or maybe I'm misunderstanding it.

Problem 1: The forecast in cloudy weather is very bad and therefore the prediction is almost wrong. It seems that Victron generally assumes a cloudless sky. Is it not possible to include hourly cloud cover to improve the forecast?

Problem 2: When I switch on Dynamic-ESS, my battery is hardly charged at all until the afternoon. All solar yields are sold almost exclusively to the grid (from the morning tiull early afernoon). In the afternoon, it often no longer manages to reach the desired SOC and sometimes recharges 10-20% of the battery capacity from the grid. This happens several days in a row (and on these days I have sold 100-500% of the battery capacity to the grid in the morning).

Problem 3: When Dynamic-ESS is switched on, the Multiplus discharges the battery at a very low level (usually around 20-50W discharge in the morning) in the morning till it decide to load the battery. This makes almost no difference to the SOC, but switching off the inverter completely in this constellation would save a lot more energy, as around 10% of our home's normal load is wasted on the unused inverter due to unproductive discharging.

Perhaps others have also observed this and can give me a tip.

Boundary conditions - Germany, constant tariff for feed-in and consumption.

Portal ID c0619ab4d1ac

Translated with DeepL.com (free version)

Another observation and comment on the objective: "Minimize cost":

The stated goal for the DESS is:

Dynamic ESS is an algorithm that aims to minimise the costs made on the grid and battery:

• By scheduling charge/discharge cycles of the battery,

• While taking grid limitations, battery specifications and day ahead energy prices into account,

• When it can, it also considers the consumption and solar yield forecasts when scheduling.

So, as an example my "projection" from today:

I believe I can see what's happening here: The algorithm "literally" trys to minimize cost and maximise revenue... as such selling as much as possible, only adding a littlebit to the battery at the end of the day to make the battery SOC excatly meet minimum (which in my case is 15% SOC). Yes, this maximises revenue for this day, but obviously does not do what I really want: it needs to include the rest of the night (but does not at this point (8 am), because (I venture) prices for tomorrow have not been set yet. This leads to feeding all surplus into the grid, and only at the last moment charge the battery just a littlebit. It will obviously run out of SOC from midnight on, which is not what I want.

Additionally, even if it would calculate the exact need until the next morning when some PV surplus is happening again. I also want to my battery to protect against (seldom/occasional) grid outages, and in these cases I want to have as much SOC left as possible. Also, the actual consumption in the night might be occousionally higher, so I want to be prepared for that as well.

In short, that means, in my mind I have a "Default target SOC" each day that I aim to achieve. As I have LFP chemistry in my battery, I aim to charge to 100% every day. [If I had NMC chemistry, in summer, I could aim to only charge to 80% if that is easily enough to get me through the night for battery longevity].

Also, what I am doing manually right now, when charging to 100%, I aim not to charge full speedright from the morning sun, but rather limit charging speed (through DVCC/"Maximum charging speed") which helps to reduce AC->DC conversion loss and reduces the charging Ampere into the battery. Not quite sure on this, but it feels reducing charging speed of the battery helps its longevity as well. In my mind, I always try to minimize charging speed so that my battery reaches 100% SOC around 2 pm. Why 2 pm? Then it reduces my lunch peak feed-in while still having 2 pm until sun-set to fully charge my battery in case didn't fully get charged over lunch because of reduced sunlight. Obviously an algorithm running every hour so can do a much better job at this. This would help us all in reducing battery wear and lunch-peak mgmt.

So, as such, for me as a house owner (might be different from boat, RV, commercial or industrial), my stated goal for the algorithm should be:

[Text to be continued in a separate Post]

There is no price schedule and no battery forecast for tomorrow at my sites and and also not at all my customer's sites. Time: 20:12

Hello guys,

will be any posibility to choose in Dynamic ESS if you want to buy/sell energy or just to have dynamic minimum SoC according weather forecast in the future? It would be really nice feature. Because I do not care about prices - I just sell/buy energy for fixed price. Currently I am struggling in Node-RED with this (I am not so skilled in it) :(

Thanks for feedback and have a nice sunny days! ;)

@Dirk-Jan Faber (Victron Energy)

we have bad weather for a few days now and the heat pump is switching on. The time pattern is always the same, the heat pump usually runs from 00:00 to approx. 06:00 at just over 2 kW. After that, the daily consumption always drops significantly:

But the consumption forecast does not recognize this pattern. It assumes that today and tomorrow there will be consistently high consumption of over 2 kW:

It is therefore assumed that consumption is significantly too high, which leads to poor planning. Is no pattern recognition implemented in the current implementation of the consumption forecast? In other words, is the temporal progression of consumption over the last few days currently not taken into account?

You wrote in the first post that the following improvement will be implemented: "Introducing (smart) support for high-consumption loads such as EV charging and heat pumps". Will this help in such cases?

VRM b827eb273733

One hour later the forecast for the current day ist a little bit better:

But for the next day it's still bad, the 2 kW consumption of the heat-pump starting at 00:00 doesn't exist at all any more:

Charging a car in the middle of the day completely ruines the DESS planning for the remainder of day and the next day.

The original plan was to dump the battery at high prices (end of the day), but when the carging completed an hour before the end of the day, the picture was completely different: the systems just assumes to feed consumption from the battery. And since the huge consumption is no longer there, at least a complete hour of high prices will be missed by not dumping to the grid.

Also the picture for the next day is completely ruined by a charging session:

Perhaps it is possible to modify the consumption planning algorithm, to just take the deviation from long term consumption as granted, but do not assume the same will be happening for any next hours.

Hello,

after many trials and following of the all comunity from the beginning of DESS, I decided 3 days ago to install node-red and try the latest possibilities.

I am very happy that there is already possible to swich on "green mode" and to define second SOC limit (b_goal SOC @ b_goal hour). Seems that there is good improvement, but nevertheless I would like to report some troubles that I can observe. Just smal info - at my instalation at the moment no unexpected consuptions as car charging, water heater or heat pump.

1. I do not fully understand to option green mode:

The system still feeds from battery to grid from battery. Which is not according to description in the node-red, but on the other hand is fine from me, it is from solar excess production. Would be nice to have better expain in documentation.

2. I have at the moment clearly lot of excess solar power, nevertheless the system is time to time during export time!!! stopping export, stopping use of battery and start to use grid to power the consuption. Further more to start to charge the battery!

This just today caused that more than 50% of the export earning is lost in the pointless power import.

3. The energy consumption from the grid is not shown at the DESS node-red system neither at the VRM graphs:

4. Yesterday night when there was yet "green option off" the system was in contrary with the plan exporting still ~500w during night. That was maybe cause why the system was under the (b_goal SOC @ b_goal hour). System started to use grid and start to charge battery from the grid, nevetheless was still above minSOC.

Hepefully this can help you or you can help me :)

Dear @Dirk-Jan Faber (Victron Energy), might I dare to ask when does Victron plan to release some updated version of Dynamic ESS? I think that all the complaints about the current (non)functionality have been thousand times mentioned here and Im pretty sure you know about them all. As for me, the most bothering seems to be the fact that in the end of the day the battery is drained to the minimum which causes that system takes all the (expensive) energy from the grid untill the solar production in the morning starts to be enough to power the loads. You claim that system aims to minimize the total energy costs, but with my setup (15kWp on Fronius Symo 15 with 3x MP-II 5000 and 58kWh battery) in this part of a year, I wouldnt need to take any energy from the grid (which works with standard ESS very well) and there is even plenty of energy left to feed into the grid. But with current version of Dynamic ESS the system actually ends up taking quite a lot of energy from the grid during the night which of course results in much larger costs than when DESS is turned off.

My company installs Victron equipment in residential buildings and boats and I have always claimed that Victron is the best, but in this case I must confess that Im a little dissapointed that with such major issues in functionality Victron released the official version of DESS...while in beta version this would be completely understandable, the official version should not in my opinion have such issues. Thank you in advance for the answer.

Ich muss nochmal etwas auf Deutsch loswerden. Ich bin seit vielen Monaten hier aktiv am Testen des DESS. Ich habe über so manche Anfangsfehler weggesehen und war immer der Hoffnung, dass es irgendwann richtig läuft. Ich habe es jetzt deaktiviert!!!!!!! Es gibt nach meiner Ansicht noch eine große Baustelle, bei den Verbrauchsprognosen. Ich hatte oft den Eindruck, es habe sich etwas verbessert, aber letztlich ist es so für mich nicht praxistauglich. Ich möchte ein paar Gedanken dazu formulieren, die vielleicht berücksichtigt werden.

- Wenn große Lasten, wie das E-Auto geladen oder Kochen in Betrieb sind, geht die Prognose extrem hoch und zwar für die nächsten 12-24 Stunden. Das ist völlig unrealistisch und wird in der Praxis selten so vorkommen. Wenn die Prognose für die nächste Stunde hoch geht, wäre das OK, aber nicht für den ganzen Zeitintervall. Im Gegenteil, der Ladevorgang des Autos ist dann für diesen Tag erledigt. Ähnlich beim Essen Kochen oder ähnlichen Verbrauchern.

-Der durchschnittliche tägliche oder auch wöchentliche Verbrauch sollte Einfluss haben und die Prognose müsste viel mehr auf diese Durchschnittswerte zurückgreifen. Die meisten E-Autos dürften eine annähernd gleiche wöchentliche Fahrleistung haben. Auch wenn ich Montag und Mittwoch lade, ist die Wahrscheinlichkeit, dass das Auto am Donnerstag völlig leer ist, gering. Andersherum wenn 3 Tage nicht geladen wurde ist die Wahrscheinlichkeit sehr groß, dass ein hoher Verbrauch ansteht. Beim Kochen oder Wäschewaschen sollten ähnliche Regelmäßigkeiten berücksichtigt werden.

-viele uns betreiben mittlerweile aktives Lastmanagement. Das heißt, wenn der Strom entweder besonders günstig ist oder viel über PV produziert wird, geht der Verbrauch nach oben. Ich habe das Gefühl, eine solche Logik berücksichtigt der jetzige Algorythmus überhaupt nicht. Für mich wäre es daher logisch, dass der Algorithmus noch andere Parameter als Wochentag, Uhrzeit, aktueller und durchschnittlicher Verbrauch berücksichtigt. Interessante Parameter wären auch Ein-/Verkaufspreis (jedenfalls extreme Unterschiede), PV Überschuss, Akku Soc und anderes.

Ich hoffe, meine Vorschläge werden nicht als pprinzipielle Kritik aufgefasst, sondern als konstruktive Vorschläge.

Als, erst mal Danke, dass Victron kontinuierlich an der Implementierung/Verbesserung von Dynamic ESS arbeitet. Gerade mit der Aufteilung in den Green Mode / Trading Mode können die verschiedenen Anwendungen von DSS besser in den Griff bekommen werden.Doch ich hab ein paar Bitten

- Bitte berechnet die voraussichtliche Akkulevel für die nächsten 7 Tage (und nicht nur für die kommenden 24h) und stellt dies in dem entsprechenden Diagramm dar
- Bitte schaut nehmt Euch (endlich) auch der Verbesserung Berechnung des „Vorraussichtlichen Verbrauchs“ an. Wenn dies nicht zufriedenstellend allgemein (für alle) gelöst werden kann dann gebt uns Anwendern bitte die Chance ein für uns entsprechendes Modell (mittelwert über die letzten 2 / 4 / 12 … Wochen auszuwählen.
Der derzeitige Algorithmus ist IMO eine absolute Zumutung. Und wenn ich den langen Thread hier lese nicht nur für ich)

Beispiele:

Der Verbrauch ist immer >1,8KWh/Tag

Nicht jedoch in der derzeitigen Vorausschau

Update: dann die Verbrauchs-Vorhersage eine Stunde später:

noch unbrauchbarer :-(

und die Vorraussage des stündlichen Verbrauchs (immer > 50W/h )

(ich hab leider keinen Screenshot von dem anderen Extrems, wo auf einmal ein Verbrauch dreifachen des normalen täglichen Verbrauchs für die nächsten 7 Tage angezeigt wird. (Meist, nachdem an einem Tag mal aussergewöhnlich viel verbraucht wurde)

(Ich passe mein Verhalten (Manuelles Netz-nachladen, größere Verbräuche ) ein wenig der Vorhersage an, doch dies sollte mit einem Mittelwert pro der letzten 4 bzw 12 Wochen für einen pro Wochentag halbwegs passen.

Die derzeitige Implementierung ist eine Zumutung, und für mich der Grund Dynamic ESS meister ausgeschaltet zu haben,
(Diese bewirkt leider auch, dass die Ladezustandsberechnung (fast) nicht mehr aktualisiert wird.
Bitte gebt uns eine Möglichkeit, dass auch bei deaktivierter DESS der Ladezustand für eine möglichst lange Zeit (>>24h) vorberechnet wird .

Trotzdem Vielen Dank an Euer Team, denn vieles läuft echt gut bei Euch, und irgendwann ist auch Dynamic-ESS mal wirklich anwendbar ;-)

Hi, it seems that the prices I See in VRM do not reflect the actual price I have to pay:

While VRM shows me around 18ct, Awattar asks for almost 2€ :-)

My settings should be fine:

I won't import energy today but it would be quite bad to not have exact pricing when relying on imports.

Best regards Christoph

Hi, I am sharing another 2 instances of the wrong DESS decisions, which actually costs money:

1. buying from grid to "shave peak" instead of using battery power
2. buying from grid to catch-up target SoC, instead of "waiting for solar to finish charging"

It seems to be, that it takes SELL Dynamic price to decide if buying from grid instead of BUY FIXED prices. Maybe this is not a common situation in your testing cases, thus I pointing it out:

• I have plain FIXED BUY price (pretty high) -- the same price no matter the time or date
• I have DYNAMIC SELL price (significantly lower than the buying price)
• DESS was in configured in the TRADE mode, battery Restrictions were turned OFF
• I have pretty large system, able to run in off-grid mode

Situation 1: BUYING from Grid to cover internal consumption instead of using battery power:

• morning, 6:30, some cofee and tea machines turned on, the peak consumption goes to 2.5 kW, while the Solar production is "just" ~400W.
• DESS plan was to keep battery at 52% SoC until around 10:00, when it wanted to start slowly charging
• ----------
• Expected behavior: the consumption peaks will be covered from battery power (because the SoC > Minimum SoC). If the battery SoC decreases below the target, it will be recharged afterwards from Solar limiting selling to Grid for a low price
• Observed behavior: the consumption peaks were covered from Grid, buying for high price (0.28 buy vs. 0.10 sell), making unnecessary spending.

Situation 2: BUYING from Grid to catch-up SoC plan instead of waiting for Solar to recharge batteries

• Afternoon, starting from ~13:22. The DESS plan is to recharge the batteries to Full because the selling price is the lowest (the plan itself is good)
• Feed-IN to grid was correctly stopped at 13:00
• Unfortunatelly, some clouds came in, and the Solar production between 13:00-13:20 was lower than predicted (still, solar production ~3000 W, while consumption was just ~500W, so Battery Power was around ~2500W -- batteries were nicely and steadily charging, though not as fast as DESS has predicted).
• -----
• Expected behavior: the charging will continue at a natural Solar-production pace, Feed-In to grid will be stopped "a little longer" to compensate for clouds and slower charging pace [see Fig. 3]
• Observed behavior: the system decided to take additional power from the grid, even though the BUY price is high (0.28), and there is NOT any other time point where the bought energy can be sell back for the higher price. [see Fig. 4]

My guess is, that somewhere in the implementation there is a mistake, that it incorrectly calculates with DYNAMIC SELLING prices instead of FIXED BUYING prices to decide when to buy from the grid.

This example happened on 26.6.2024. My VRM ID = c0619ab5fb1b

Please see the attached screnshots for getting grasp on what was going on

(EDIT: for some reason I can not upload images, it tells me "No link in upload response" :( ).

Hi, i also cant activate it. VRM ID c0619ab4bbc2

I think it is unlogic to charge the battery between 14 and 15 and discharge The same bit between 18-19 given the price scheme..

Id c0619ab39902

@Dirk-Jan Faber (Victron Energy) Great now make it like go-e charger and add dynamic pricing information for Awattar!

I see its missing.
Simple formula:

(EPEX SPOT AUT * 1,03 + ,0,015) * 1,2
Quote of Awattar:

Oh I see its there with another name already ;)

Please add Awattar. Seems like a copy paste job for smart energy.

@Dirk-Jan Faber (Victron Energy) Second comment. VERY COOL FEATURE.

Activated it but ran into a problem:
"Battery capacity unset". I read the dynamic ESS manual fully and expected it to work.
Venus OS shows dynamic ESS in Mode AUTO. So the update works.

Is something off? Why the error?

It's great that DESS has now been officially released, which will significantly increase the number of users and feedback, but will of course also cause you even more work ;-)

The most important points on your to-do list in my opinion:

• Maximizing battery charge by the end of the day to cover consumption until sunrise

• Periodically fully charge the battery to improve battery life

And as I had already noted in the beta phase, DESS obviously ignores the grid set point configured in the remote console. I always set the grid setpoint in ESS slightly negative (-10W) to avoid grid reference as much as possible. DESS seems to ignore this and seems to adjust to 0 W or slightly positive. This causes unnecessary expensive grid consumption over a longer period of time.

The rest: great!

Great that we have now the official release! Many small (or bigger) improvements have been made during the last couple of months with a lot of feedback via this forum and quick response from Victron @Dirk-Jan Faber (Victron Energy) . Thanks!

Now for the next release :)
- Please do not be so strict on target SOC
- Able to set a minimum SOC for DESS. It is now taking the minimum SOC of the system resulting in that the batteries are empty many times for longer periods. And there is a lot of capacity to play with when it charges only up to a good 50%. See picture below. (from node-red because it gives a nicer picture to show this)

Hi everyone,

First of all congrats on the amazing functionality!

What time zone should I use to insert the hourly rates (fixed or variable)?

For instance I am located in Bulgaria (EET) and my local night rate is between 22:00-06:00 local time. Do I have to set this rate between 21:00-05:00 (CET)?

The selling rates are variable and are with delivery time set in CET.

Thank you

Hi

It was mentioned before that once Dynamic ESS was rolled out the VRM grid data would include 'Grid from Battery' data, but I still see only Grid to Battery data e.g.

I assume one does not have to have Dynamic ESS enabled first. Will a "From Battery" box be added to above?

Many thanks.

Hi,

would it be possible to see the dashboards of the DESS after configuration without having the DESS enabled e.g.: via a toggle button? The KPIs are quite useful for other custom implementations as well.

Hi!

Why does the system sell electricity when it is deactivated?




ID c0619ab313dc

Does it still make sense to always install the latest Venus OS beta version, as was the case in the DESS beta phase? That would currently be v3.40~1, for example.

Or now that DESS has been officially released, stick with the official Venus OS versions?

It’s great to see DESS live. With your comment on 30 min pricing being in the works, will that also include a link to the Octopus API, as that provider is the most common variable pricing energy provider in the UK, and it would be great to be included.

Currently I’m doing the setup manually every day when I get the day ahead prices.

I understand the system is leaning, but Is it normal during the fist 28 days for DESS to charge the battery during the day when rate is high, and hardly put any charge in battery over night on cheap rate. Rates are setup in settings. For example, currently generating solar with forecast for sun to continue this afternoon, SOC is already at 76% but the system is currently charging from grid at 1600w on peak rate electricty.?

Simon

Hi,

I have boiler connected to AC2 Out which is switched on based on SoC, it is configured inside the Multiplus and uses its SSR, it turns on phase 1 at 95%, phase 2 at 97% etc., is it possible to make DESS working with that? When I switched DESS on, battery does not reach above 90%, it starts feeding the grid when SoC is 80%. I suppose it can't work together (boiler control and DESS active) or can I somehow tell the system to charge the battery above 90% when there is enough PV power?

When there will be some news about DESS functioning, where is the best place to subscribe to be informed? And second question - when there are some improvements, is it needed to update Cerbo GX to make them active or does this happen on some Victron cloud and those changes are effective when released?

Wir hatten das Thema schon einmal bei der Beta Version. Ich habe heute mal wieder beobachtet das das DESS beim parallelen Laden aus dem Netz und PV Produktion, die PV Produktion runtergeregelt wurde.

Forecast is not quite reasonable.

Very high consumption forecasted for tommorrow. I had high consumption 3 weeks ago but not that high as forecasted and last two Sundays it has been less.

And solar irradiation forecast for tomorrow is about 2 times less than for today but DESS thinks that solar production will still be higher. 6,4 for tomorrow and 4,8 for today. Maybe it will be, we well see tomorrow but it does not seem to add up with irradiation forecast.

Have to disable DESS tomorrow because there is nowhere to put solar if target SOC will be 100% for most of the day and not much real consumption.

Site id 48e7da86cd8b

Any plans to add the UK to the list of supported countries and it's suppliers who now have a wide selection of smart tariffs.

Some smart tariffs have up to 3 different pricing periods for both import and export, some have variable half hour pricing during a 24 hour period.

Feedback on the solar forecast element of DESS - I've been observing the daily forecasts for several months while DESS has been in Beta. For our location - Chester, NW of England, the forecasting is often miles out and nowhere near reality.

Also now in the UK, some energy providers are offering smart tariffs where the energy provider controls the charging and discharging of our batteries to automatically charge the batteries during periods of low grid pricing during the day and discharge during the peak 4-7pm period.

Example Octopus Energy - Intelligent Flux tariff, currently in beta with Givenergy batteries.

Any plans to offer similar integration?

Hi all, I've configured Dynamic Ess, reloaded the config with the new assistent in my Multiplus. But now the Multiplus does not sell to the grid or provide power for self consumption. Charging goes well. Do you know what I did wrong?

@Dirk-Jan Faber (Victron Energy) Last Saturday afternoon during a DESS planned "buy" event the MPPT charger was throttled back to zero for about 5mins. Battery was almost full but charge current from multi's was still very high. In fact, it started charging again before the pylons reached 100%.

I you have a moment, 23rd of march 15:00 - 15:10.

# c0619ab2ef52

Can @Dirk-Jan Faber (Victron Energy) please explain me, why DESS behaves like this on my setup? I attach screenshots of settings - during the highest energy price it covers the consumption of home from grid and not from battery.

When I look on tomorrows plan, it seems it just wants to use battery for covering morning consumption, when the price is high and then it does not want to use battery at all.

@Dirk-Jan Faber (Victron Energy)

I see that the target SOC @24h is changed from Min SOC to higher value. Looks like it is calculation off planned consumption till next production is high enough. What is the correct way. Only the multiplication x1.25 (transformation losses DC->AC) is not made.

I would even make this correction higher factor 1.5 or even 2 to cover also unplanned extra consumption. Do you look also to next day (production - consumption) > 0 ? Because if there is not enough sun next day you need to keep more in the battery.

We have a 10x cost difference between off peak and on peak, 4.5p at night 12-5am and then 5-23:59 - 44p. So we very much avoid all peak usage. I have have had dynamic ESS on for a few days with sensible values and today I have woke up to a 14% charged battery over night. Dynamic ess is buying and selling in the off peak period back and fourth and I have no idea why.

Any ideas before I need to switch it off ?

In the VRM general settings, there is an option for US Dollars but not for Australian dollars. Can Australian dollars be added?

On an (obviously) related note - if I change the currency unit in VRM generally, does the Dynamic ESS menu currency setup inherit that change automatically?

Is it normal that DESS ignores the minimum SOC and optimized without battery life?

My minimal SOC is 10% and battery life is off, but the system stops sometimes at SOC 30 or 50 and i use grid instead battery. It seems that battery life is working although it is off. I want to use my minimum SOC 10%, because when my PV wont load my batterie the DESS should load from the grid when the price is low. But at the moment i start with over 40% battery in the day and my PV charges the battery now immediately to full. The rest of the day i sell my PV to grid and in the night i buy from the grid while my battery is over SOC30.

How can this be explained:

- Hardly charged overnight at all on cheap rate.
- At 6am was already importing from the grid (4 times the price).
- Sent almost all excess PV (18kWh) during the day to the grid.
- Exported battery -> grid 1-3pm (when not enough SoC for evening)
- Importing from grid again at 4pm.
- No where near enough in battery for evening usage. Grid usage planned all evening.

Weird thing is that yesterday it seemed to work really well. Today this logic makes no sense at all though.

c0619ab38079 (26/03/24)

Can't get into VRM settings page. When I go to Settings -> Dynamic ESS both buttons "toggle on/off" and "start configuration" are greyed out.

The system: 6 x MPII 5kVA, 3 fase, FW 508, assistant 018C. Venus OS 3.40~1.

Did I miss something?

ID: c0619ab33ea2.

Id: c0619ab42937

I am very pleased with the DESS. However, yesterday (26/03) I saw on the Energy graph that the system loads the battery from the grid for small amounts of time/KWh although there is enough solar (see the red circles on the graph). Is there an explanation for this behaviour? Thanks.

I'm from Spain and my energy provider is Octopus. When I try to select energy providers, none appear, although I would be fine with using the indexed price.

This is the formula I use for the electricity calculation:

(p+0.005)×1.235

Energy provider profit share = 0.005 VAT + Taxes = 1.236 (21% VAT, 2.5% tax)

This is the formula my provider includes in the contract:

Energy (€) = ([Σ(p=1 to n) (OMIEp+CMp+RCp) * (1+PRedp)] + 0.005) * M + ATRe) * Consumed Energy p

The simplified formula would be: indexed price + 0.005

In Dynamic Energy System (Dymamic ess), it always shows that the price is less than 0.01€.

s shows that the price is less than 0.01€.

@Dirk-Jan Faber (Victron Energy)
ID c0619ab313dc
DESS activatet,
Bad Weather,
Akku65% (2x280AH),
2 hrs cheapest Tibber price (17 cent),
no battery charging from the grid,

Maybe I'm misunderstanding DESS, but I always thought that charging from the grid is also part of this logic.

I tried DESS again. Unfortunately, I discovered that since one of the last DESS versions was feeding into the grid very often, even though I had prohibited feeding into the grid with 0W and in the settings.
Unfortunately, the grid-feed-function is deactivated in that moment, when activating DESS. Unfortunately, this cannot be used for me in Germany.

The Grid feed in Maximum in the ESS setting do not seems to work for AC pv inverters on AC out, max is now 300 W still pushing to the grid (at very low price ?)

gr Hans

VRM id : c0619ab4d86f

Hi, my system does not Charge when Dynamic ess is on.

Vrm ID:c0619ab4bbc2

This happens very often and I don’t know why.

DESS schedule does not obey "grit current limit"

@Dirk-Jan Faber (Victron Energy)

Cerbo GX v3.30, MP2-5000

VRM portal ID: c0619ab4b370

I have a MP2 and DESS running.

Grid current limiting set to 3A, which is obeyed. Max charge/discharge is around 600-700W.

DESS is working as well. On low price grid is used to charge battery. On high price battery is used primarily. Thanks for the great work!

Maybe I found a small flaw. The DESS schedule seems to show higher scheduled values for „Grid to Battery“ (2,5kWh) compared to what is possible in respect to the grid current limit set to 3A -> 0,7kWh.

I assume the capability of the MP2-5000 is used for scheduling and not the grid current limit set.

Today I noticed that the forecasted consumption went haywire, this wasn't the case in previous days with dynamic ESS enabled. This also don't represent other historic consumption (see previous days). I don't charge an electric car or have other giant loads that consume 5kw/6kw.

The DESS planning is struggling to plan to buy energy today at the low prices because of the high forecasted consumption en already started buying energie hours before the lowest point. An hour ago I disabled DESS and will enable DESS at 12:00 to take-in grid power to the battery at the lowest price.

You would think that at some point DESS would realize that the forecast consumption is way-off and would recalculate this.

System ID: c0619ab24334

Other info for today

Any explanation, why the DESS is feeding consumption from grid to when batteries are charged and there are peak price hours now, it does not make sense.

Portal ID: c0619ab1e051

UPDATE: After a one hour, consumption forecast somehow changed to 0 consumption (compare it with forecast from screenshot above where the forecast was correct) and SoC is scheduled to hold to 100%, not sure what is happening, but it is quite strange resulting to permanent grid usage and those consumption forecast "drop-offs" are very often completely disturbing almost perfectly planned DESS schedule...

Update 2: Because of that I disabled DESS, after a two hours, consumption prediction completely changed to extremely high for a change, see next screenshot. Hope this helps for better troubleshooting. It seems, that when I enable DESS, consumption forecast goes down to almost no predicted consumption, when I disable DESS, it reverts back to high consumption forecast.

Ok, I started playing around with dynamic ESS. I will watch the algo and its results for a couple days, but before I comment on this, a couple other notes:

- In the pricing configuration, you should state that the price should include any transfer fees, charges, taxes etc. Essentially the full amount one has to pay. For example, in Austria, with Awattar, yes, the formula is correct [price *1,03+0,015)*1,2], but it should actually include the "Nebenkosten" of around 8 to 9 cents as well. I did that by changing the increment to 0,095 and thus reflecting this additional cost element.

- the battery has not only cost in terms of usage of lifecycles, but also in terms of rount-trip losses of 10 to 20%. This should be part of the algorithm and ideally the settings. essentially this makes the battery more expensive, and should lead to less "trading".

- You guys should state the principals of the algorithm, and perhaps (on a schematic level) disclose the computation process. this itself will clear out many concerns and questions. Also, I believe Victron would philosophically make the Dynamic ESS an option that virtually every user should have turned on (and perhaps become and industry expectation?). Building trust is crucial to this point.

- The graphic at the bottom is great:

I would just try to add the cost element against another y axis in a different color, and also include the SOC line in there. With this one graph, you essentially combine three other graphs on this VRM landing page! And, once the prices are available for the next day (from 2 pm on) I would immediatly show today and tomorrow (48 hrs). Then everybody can see everything happening in one glance. Very powerful! :-)

These are my comments before I go into the algorithm in a couple days/weeks once I glean more insights there after watching (but not yet activating it).

Final note from your post: High volume consumers (Heat pump and EV charging) are indeed crucial. For example, in our household we are EV charging 98% of the year surplus PV only... so any form of Victron Wallbox surplus energy charging should be excluded from the consumption numbers (or an option to include/exclude all or surplus only)... otherwise it distorts the consumption numbers which otherwise look ok'ish. :-)

So I enabled DESS two days ago, this morning first the battery is discharging into the grid because of high prices which is very good and understandable (between 8 en 9 am). But then the system starts charging the battery from the grid, still at a high price. This seems very counter intuitive to me, especially since the system plans to discharge again the next hour.

Can someone explain this behaviour?

These are the relevant screenshots from VRM:

Hi there,

DESS should be charging this hour to target soc, but does not. It seems that it expects PV to charge, but there is only a cloud in the way.....The solar expectations are way to high for days and DESS is not charging to the target soc. In the next hours DESS will charge I think, but will never reach the total charge of the battery in the cheap hours anymore.

After 2 hours charging the third hour was again not charging (even discharging a little).

Had this behavior in the last days, and sometimes it even discharges below minimum soc also....

VRM -id: c0619ab4b607

I have another issue to report: the grid limit is not being respected as it should be. Today the battery has been charging at it's max speed and I am also charging our cars because of the low prices at the moment.

I have a 3 phase grid connection with 25A per phase, a Victron multi 5000VA on Phase 1 with the battery, an Opel Corsa that can charge on 1 phase only and a Skoda Enyaq that can do 3 phase charging.

I initially set my max import power at 15,75kW, figuring that this would be divided by the 3 phases. When I start a 1 phase charge with the car, phase 1 quickly exceeds 25A limit and because my car charger (a Myenergy Zappi) is also load balancing then stays at 25A. The same happens with a 3 phase charge, none of this time the total limit of the 3 phases as a whole is exceeded of course. But even when I set the max import power to 7kW or even 4kW the multiplus keeps charging and the import power exceeds the set limit by a very wide margin!

I would think that setting an amperage limit per phase would make much more sense, or at least to divide the import limit by 3 and reduce charging if the limit/3 is reached on one of the phases. But even if the limit is exceeded the multi keeps charging...

Dirk-Jan: do you know what is happening here?

My site ID is: c0619ab343b0

Screenshot with 1 phase car charging:

Screenshot with 3 phase car charging: