question

rvicev avatar image
rvicev asked

How does the AC PV Frequency Shift algorithm work?

I have been following this Victron Community platform for over 2 years now, read most of the documentation, studied the presentations on Victron Professional, etc. Learned a lot, but there are still some aspects to frequency shifting that aren´t totally clear to me.

Let´s assume following configuration :
- off-grid MP II 5000/48
- AC-coupled PV inverter 5000 VA (compliant to factor 1.0 rule)
- LiFePO4 battery, dimensioned as per minimum Victron specs (or bigger)
there may be MPPT(s) also, but that seems less relevant for the following.

If the battery is nearly full the MP II will frequency shift to signal to the PV inverter to lower its output.

It seems unlikely that this is based on SOC, as is often mentioned on this community platform, but rather based on voltage. SOC is a calculated value and less accurate. Also,I remember having read somewhere in the Victron docs (ESS manual - grid failure ?) that the battery charging rate is decreased from 100% @ 54V to 10% at 57.6V

However this can not be the only criterium.

Let´s suppose (theoretical) :
- battery isn´t nearly full at all, say SOC 60% (voltage roughly 52.3V)
- sunny day, PV inverter is producing at full blast (5000 VA)
- only small loads active (500 VA or less)

The MP II 5000, according to spec, has a maximum charging rate of 70 A, or in rounded figures roughly 3500 VA. So the MP II can only process 3500 VA, but has 5000 - 500 = 4500 VA coming in at AC-OUT.

So where does the remaining 1000 VA go to ? (after all, it will take quite a while to raise the SOC from 60% to 100%, so what happens in the mean time ?)

It therefore seems to me that frequency shifting must also be triggered on incoming load (not to be confused with consumers) at AC-OUT, even if the battery isn´t (nearly) full ?

Furthermore, I assume that since the MP II´s output capacity de-rates with temperature, it´s charging capacity also de-rates with temperature ?

This would mean that there are at least 3 criteria :
- battery voltage
- incoming load at AC-OUT
- temperature
for triggering frequency shifting.

Is there any info available on the frequency shift triggering algorithm ? Are there any other criteria I overlooked ?


offgrid
6 comments
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

rvicev avatar image rvicev commented ·
Did I ask a silly question ? Or is everybody who should know more about this still on holiday ?
0 Likes 0 ·
Alexandra avatar image Alexandra ♦ rvicev commented ·
Too many variables. In the question and yet not specific.

Some batteries send a do not charge signal to the system, the system shifts frequency if offgrid and throttles or stops the PV inverter.

Not all batteries behave the same some send a power charge amps signal some reduce to 0A.

Not all pv inverters behave the same either some are quite binary, some throttle and run weird (flickering lights). Not all use the same frequancy range either.

Then there is the load demand trigger as well, battery full yet produce for loads. Again some pv inverters behave differently here as well and dont do this but wait for rebulk.

Some don't throttle at all hence the requirements for frequency shifting response.

I do know for sure that the MP charging derates as well when the inverter is running warm. In fact charging from AC PV is a bad idea. It is best to rely on DC charging for battery and AC PV for loads.

0 Likes 0 ·
rvicev avatar image rvicev Alexandra ♦ commented ·
@Alexandra : the question seems pretty specific and is related to what happens if the battery is not full and there is excess AC PV coming in, i.e. more than the MP II can handle.


It applies equally when using DC for charging and AC for loads ..... that´s why I mentioned in the beginning ¨there may be MPPTs as also, but that seems less relevant ...¨
0 Likes 0 ·
Alexandra avatar image Alexandra ♦ rvicev commented ·

@rvicev

By not specific I was refering to brand of AC PV and battery/bms manufacturer. Behaviour is specific to the combination.

The answer will be the same generic.

If direct to loads, the AC PV will/should run that, if the battery has not throttled it down. Excess will charge batteries if PV has not been throttled for heat/derating or battery charge reasons.

The response of the ac pv the frequency shift will determine how binary (i e. on / off or just slow down) referring to the throttling response is of the pv inverter. And the signal from the battery. Some batteries throttle the charge current down, some request 0A.

With DVCC the system will take into account the dc mppts as well for max charge.

But i have seen with some brands of pv inverter if you have dc mppts, it favours the DC for charging and if it is enough even loads.

Some will only dc charge and ac pv runs loads and not charge from ac pv.

Some do the combo. It strongly depends on the set up.

Mostly the logic is - if this then that.

0 Likes 0 ·
rvicev avatar image rvicev Alexandra ♦ commented ·

@Alexandra : in my example the battery is @ 60% SOC, so the battery/BMS is not very likely going to send a signal to throttle down. So the question is independent of the type of battery/BMS.


Still, there can be excess AC PV coming in, simply because the MP II charging capacity is limited. The DVCC system must then somehow throttle down the AC PV inverter, and I assume this is based on AC PV coming in at the MP II´s AC-OUT .....


I understand that how well this all works depends entirely on the type of AC PV inverter, but even if it is a Fronius, it can´t make the excess 1000 VA in my example disappear into thin air : the Fronius must be throttled, even if in this example the battery is only at 60%.


If, in some cases as you say, DVCC favours the MPTTs for charging, then the situation gets even worse (excess AC PV would become 4500 VA in my example) and the AC PV inverter must definitely be throttled down ! If it would switch off the AC PV inverter (binary), then AC-coupling wouldn´t work. That seems to me the obvious reason why one could prefer an AC PV inverter recommended by Victron, such as Fronius.

As for temperature : the MP II AC-OUT is bi-directional. If it de-rates output with rising temperature, it seems to me that it also has to de-rate input (and therefore battery charging capacity) with rising temperature ....

This frequency shifting mechanism seems thus far more complex than often suggested.

0 Likes 0 ·
Alexandra avatar image Alexandra ♦ rvicev commented ·
The MP derating for both loads and charging is well documented. So I didnt address it that much. Bi directional also well documented and advertised.

The frequency shifting happens in steps.

If it needs to charge then it is working. How much it is working depends on throttling in DVCC (also well documented as the battery way of controlling the system)

Some pv inverters have a very narrow frequency range so their steps are less. On or off. Some like fronius are on/throttled/off.

If throttle then not going to produce the full 4500w of pv in your example. It is a power generator will only supply the demand. Same way a normal fuel generator does.

0 Likes 0 ·
2 Answers
Guy Stewart (Victron Community Manager) avatar image
Guy Stewart (Victron Community Manager) answered ·

Hi @rvicev,

Nice to hear you have been making the most of the resources, and still thinking about how it all works.


The AC PV control functions are based on DC voltage at the Multi. Not to be confused with the battery voltage. I'll explain the difference.

Even though load and temperature are involved in deciding how much power is required, in the end they get 'reduced' to a DC voltage signal.


The system does not need to do any accounting for AC load. This is done automatically by physics :)


As a large load is applied to the system, it adds resistance to the AC circuit, and the AC voltage will drop accordingly.


This AC voltage drop will mean the voltage at the point of the load will be lower than the AC voltage at point of the AC PV inverter, and current will flow from the AC PV inverter to the load without any control necessary from the Victron system.


There wil always be less resistance along the AC line than the across the DC to AC inverter bridge, so as long as the AC PV inverter is operating at full capacity, power will flow from AC PV to AC load first automatically.


The exception to this is when the AC PV inverter has already ramped down.


That is why you can see some delay when the battery is full, the AC PV is limited (or zero) and a large load comes on, it will first be supplied by the battery (and battery inverter) until the voltage of the battery drops a little bit. Then the system will signal for the AC PV inverter to begin produciton again.


That brings us to the next part of your question. If the AC PV inverter is producing too much power for the battery and load, then the battery inverter will signal to the AC PV inverter to reduce its output.


There is a correlation in the Multi between the AC voltage and the DC voltage (via the torroidal transformer and FET bridge).


If the AC voltage rises because there isn't enough load then the DC voltage will also rise, and once it hits the target, the (correctly programmed) Multi will send the signal for AC PV inverter to reduce its output (by raising the AC frequency).


If the Multi DC voltage is higher than the battery voltage, then current will flow from the Multi to the battery, and vice versa.


The Multi charge current limit is mostly related to the ability for the Multi to disappate heat, it's not a hard limit, and it can be reduced by temperature derating.


The multi regulates how much current flows though it by controlling the voltage differential between it's DC voltage and the batteries DC voltage. The bigger the differential, the more current will flow.


If the charging current is too high for the multi, the multi will do what it can to reduce the DC voltage so less current flows to the battery, and if programmed that there is an AC PV inverter, that involves raising the AC output frequency.


When the AC frequency rises, the AC PV inverter will then increase the voltage on the PV panel side (using it's Power Point Tracking control), and as the PV side voltage increses, the PV current decreases even more, decreasing reducing the total power output.


Note the limits of this control loop is one of the reasons why we have the factor 1 sizing rule, so the system has enough time to make the adjustment (and doesn't recieve more power than it can handle too quickly).

3 comments
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

rvicev avatar image rvicev commented ·
Thanks @Guy Stewart (Victron Community Manager) for this extensive answer.


I always like to understand how ´mission critical´ equipment works, before deploying it, and this is very useful input towards that understanding ! :)
0 Likes 0 ·
Guy Stewart (Victron Community Manager) avatar image Guy Stewart (Victron Community Manager) ♦♦ rvicev commented ·

Me too, I am endlessly fascinated by how a complex system comes together.


If you haven’t seen it already, this Video by Johannes is an even deeper dive into how the Multi is able to create an AC sinewave from a DC voltage, something I also did not understand very well before his explanation.

https://youtu.be/UPfUn5ki7OM?si=EWxZAVci3314pO3I

0 Likes 0 ·
rvicev avatar image rvicev Guy Stewart (Victron Community Manager) ♦♦ commented ·
@Guy Stewart (Victron Community Manager) : Yep, that was one of the first videos I watched :)
0 Likes 0 ·
Paul B avatar image
Paul B answered ·

Hi, the frequency shift is based on battery VOLTAGE only, and is based on your voltage settings that you install/setup in the ESS or the PV assistant.

soc has no imput and ac or dc load has no input


except where the batteries BMS takes FULL control then the bms will tell the multiplus when to increase the frequency ie a hi cell voltage then frequency will go the high point, or temp as well. the bms also can control the max amps as well, however in thevreal world max amps is a bit hard as what do you do if you are 1 amp over the max amps turn the inverter off as you cant control the amps flow easly, so basically everything is dc voltage controlled, from my point of view

I use the 123smart bms on my cells and it works very well in conjunction with the ess assistant and the 123smart software installed on the cerbo. (using a USB inteface.)


again of note I am not talking about a battery that has a internal bms as thats not connected to the system interface usually. as these internal bms systems have there own internal shutoff relay/fets. and they dont trigger a high or low cell event out to the MP/cerbo.


hope the above is of some help in answering your question



1 comment
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

rvicev avatar image rvicev commented ·

Thank you, but no, not really.-

I have no doubt whatsoever that it works and I know voltage settings are the only parameters you can configure - but that doesn´t mean there is not more to it.

Again, as pointed out before, my question does not relate to the situation where the battery is (nearly) full, so type of battery or bms is not relevant in my example.

It´s about excesss AC PV coming in, off grid, and the charging capacity of an MP, which is limited and presumably de-rates with temperature as well (although I have not found any spec on that). In that situation the AC PV inverter needs a signal to throttle down as well ....

0 Likes 0 ·

Related Resources

Victron Offgrid page