[SOLVED] Generator woes in 3p hybrid AC (Fronius)/DC system

Question

question

petajoule asked Oct 10, '20

[SOLVED] Generator woes in 3p hybrid AC (Fronius)/DC system

Summary

Problem1: Generator feed-back

Cause: Gridcode (Germany) & LOM type B enabled for AC-IN1 (where the Generator is connected. Symptom was that there never really was any demand on L1.

Solution: Disable LOM on AC-IN1

Problem2: Generator was not demanded beyond 4A per Phase, although AC Input current limit was set to 7.5A

Cause: Unknown, probably firmware bug in either Quattro 469 Firmware or the 2.60 Venus Firmware

Solution: Wait 3 hours, turn Generator on again, then it respects the AC Input current limit again

Problem3: Dynamic current limiter does not work (description see below)

Cause: Gridcode enabled

Solution: Disable gridcode, thanks to @Daniël Boekel (Victron Energy Staff)

I would have expected a Gridcode to apply only for the AC-In defined as "Grid" and not for the one defined as "Generator" (and I still do), but as the genset functionality is more important for our off-grid installations, disabling gridcode altogether is an acceptable trade-off.

So the question is why. I would have expected the Quattros to pull the defined amps from the genset symmetrically on each phase and to juggle real-world consumption (or excess) on AC-Out.

Instead, I see a constantly skewed L1 with feedback to the genset - see picture.

Evidently the Fronius (Symo 6) has something to do with it, because when that thing is turned off, the feedback is gone. So I wonder what's wrong there or if it falls into the "It's not a bug, it's a feature" category.

Moreover, the generator runs less-than-smooth. All load changes seem to go directly to the genset, instead of being compensated by the Quattros (power assist or whatever), so the genset Amps go up and down like a kangaroo.

I would assume that defining an AC-IN source as "generator" means that "feed-in" is an absolute no go for the quattros. So why isn't it?

Generator Fronius

1602336879214.png (149.1 KiB)

2 |3000

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

Modifications Space

Node Red Space

Answer 1 · 2020-10-10T13:50:21Z

Daniël Boekel (Victron Energy Staff) answered · Oct 10, 2020 at 01:50 PM

Hi @PetaJoule

In this picture it looks like L1 is not charging at all.

Are you using ESS? a grid code?

5

2 |3000

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

petajoule commented · Oct 10, 2020 at 02:15 PM

No ESS, Grid Code Germany, but generator is on AC-IN1, Grid on AC-IN2 (and disabled)
Fronius just departed with Error 105 (frequency too high)
No wonder - AC-In:

does no good things on AC-out:

Max amps on AC-out are set to 7.5A, I would have expected the genset is used up to that (and this level constantly used). Instead the generator goes up and down as if the quattros would extremely vary demand.

The genset is an

https://ademax-strom.de/ITC-POWER-Full-Power-Stromaggregat-Diesel-7-kVA-DG7500SE-T-230-400-V

Actually the first one we got to run at all...

AC Input power goes cuckoo as well. And from time to time both the genset and the Fronius go out. Well, the fronius is of course set to MG50 - I can imagine it does not like the frequencies it gets.

@Daniël Boekel (Victron Energy Staff) it's installation 53255 if you want to have a look. e.g. right now (16:39), L1 was demanded 16.6A where the cap should be 7.5A

It looks in general that L1 does not respect the 7.5A threshold. L2 and L3 do.

0 ·

1602338760970.png (9.2 KiB)

1602338804276.png (11.0 KiB)

1602339883483.png (110.5 KiB)

Daniël Boekel (Victron Energy Staff) ♦♦ petajoule commented · Oct 10, 2020 at 03:06 PM

Hi @PetaJoule

The first thing would be to disable the grid code! why are u using a grid code?

from there look further: generator stability for example.

0 ·

petajoule Daniël Boekel (Victron Energy Staff) ♦♦ commented · Oct 10, 2020 at 03:14 PM

Why is the grid code relevant???
A generator is a generator is a generator.

Of course on AC-In2 there is a German grid - hence German grid code.

Why would we want to disable that? It has to be there.

The system knows (at least somewhere someone has told it) that AC-In1 is a generator. Why on earth would be a grid code relevant for a generator?

0 ·

Daniël Boekel (Victron Energy Staff) ♦♦ petajoule commented · Oct 12, 2020 at 09:28 AM

Hi @PetaJoule

if you don't use ESS you don't need to select a grid-code

can you try running without a grid code?

also, I assume you have LOM disabled for the generator input?

0 ·

petajoule Daniël Boekel (Victron Energy Staff) ♦♦ commented · Oct 12, 2020 at 10:47 AM

We did an even lower-level test by putting the genset directly on a 3p 4.5kW heating rod:

6.5A long-term rock stable per Phase, so I am fairly confident now it's not the genset.

We have to keep LOM on AC-IN2 (grid), but I disabled it on AC-IN1 (generator), formerly it was type "B". Hell knows why.

NS Log said something about "RoCoF too high" which is not surprising.

Deactivating LOM brought a slight improvement. In quantitative words:

Before

AC-IN1 current limit 7.5A, L1: ~ 150W; L2&L3: ~ 1500W each

After

AC-IN1 current limit 7.5A, L1&L2&L3: ~ 1000W each

<edit>

this graph shows nicely that the AC genset power influx is ~3200W before and after, except now it is distributed more evenly:

</edit>

input current limit is not respected anymore. I put in 7.5A, Quattros take ~4A (swinging 2.7A to 5A) per phase. I put in 10A, Quattros do the same ~4A

"Weak AC Input" is disabled, dynamic current limiter is enabled. Ideally we'd want the genset to run delivering pretty constantly its 7.5A per Phase no matter what.

It can't be a Limit in the DC Charger, because from grid 8A are taken per Phase (set up limit) when charging batteries.

Situation seems to get better, although not optimal yet.

0 ·

1602498162799.png (3.7 MiB)

1602503158048.png (71.1 KiB)

Answer 2 · 2020-10-10T14:33:48Z

petajoule answered · Oct 10, 2020 at 02:33 PM

Tried to capture the "madness" in a short movie.

genset-unstable.mp4 (642.9 KiB)

2

2 |3000

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

Daniël Boekel (Victron Energy Staff) ♦♦ commented · Oct 10, 2020 at 03:07 PM

Are your AC loads really fluctuating that wildly?

0 ·

petajoule Daniël Boekel (Victron Energy Staff) ♦♦ commented · Oct 11, 2020 at 09:49 AM

Usually not. We believe the main culprit is the frequency fluctuation of the genset which the Quattros sync onto and then pass that frequency to AC-Out - which breaks havoc with the Fronius and its MG50 setting.

Things got somewhat better when we turned off the Fronius (no feed-back to genset), but the overall instability and load fluctuations on the genset remained.

If you are referring to this:

Then the 2 peaks with 2.4kW per phase is a heating rod where we wanted to look if the Victrons are able to shave off the load (like power assist) from the genset. They are not, the full load went to the genset, which turned off temporarily.

That's why one sees inverting and no power assist. The rest I would not call "fluctuating wildly".

0 ·

gentest-201010-15-16.jpg (189.2 KiB)

Answer 3 · 2020-10-11T08:47:00Z

petajoule answered · Oct 11, 2020 at 08:47 AM

I have to tell you, this situation is utterly frustrating. For the time being, the working hypothesis is, that you cannot operate a 3p Generator on a 3p Victron system.

We have 1p installations, where you can perfectly operate a quality generator (Honda inverter generator) in the system, we also have installations with a CHP generator, but since February 2020 and 2 diesel generators later, it seems impossible (and increasingly frustrating) to manage to include such a generator into the system.

Victron does not offer any satisfactory solutions in this case. Much ado about "hundreds of parameters" over and over again. Gridcode, LOM, weak AC-Input, dynamic current limiter, rah rah, etc. pp.

You cannot use a 3p diesel generator in a 3p Victron hybrid AC/DC system. Period.

This is what we will tell customers from now on, and this is what we will also document and write about here and otherwise.

1

2 |3000

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

Daniël Boekel (Victron Energy Staff) ♦♦ commented · Oct 13, 2020 at 08:11 PM

There are thousands of 3 phase systems with generators, either on-grid (generator purely for backup) or off-grid

combining small generators with an ESS system can have its challenges due to grid-codes and LOM detection (hence why LOM detection should be disabled for a generator input), this is all explained in this document:

https://www.victronenergy.com/live/multiplus_faq

(also linked from here:)
https://www.victronenergy.com/live/ve.bus:grid-codes-and-loss-of-mains-detection#q3can_i_use_ess_with_a_generator

0 ·

Answer 4 · 2020-10-12T08:37:19Z

petajoule answered · Oct 12, 2020 at 08:37 AM

Ok, more analysis.

We put the house installation on bypass, turned off AC-out, and switched the Quattros into "Charger only" mode. Which means there was only the Generator, the 3 Quattros and the battery.

Input current limit was set to 7.5A and indeed, L2 and L3 delivered(generator)/drained(quattro charger) as much (~ 1500W), but L1 kept floating around 100-400W.

So this falls short some 1500W because of L1. Then we set the input current limit to 1A to see if there was a load problem on the generator.

Nope. :-/ L2 & L3 @ 100-280W, L1 falls behind by a factor of at least 5.

We then rotated the phases L3 -> L1, L1 -> L2, L2 -> L3 to make sure there is not a problem with some coil or somesuch specific to the generator. Result: same! Again L1 is not delivering power (so this time L3 from the generator which did deliver power when it was connected to the L3-quattro)

We then measured - again - all cabling diligently. All same 0.2 ohm.

When charging from grid, everything behaves as it should. While we're really running out of ideas what's wrong here, I would like to reiterate that we will get to the bottom of this one way or another.

edit:

VenusGX Firmware 2.60

Quattro Firmware 469

1602491373380.png (29.6 KiB)

1602491419554.png (32.8 KiB)

2 |3000

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

Answer 5 · 2020-10-12T14:53:21Z

petajoule answered · Oct 12, 2020 at 02:53 PM

Dynamic current limiter not working - the hard way. As soon as load is enabled, Quattros switch into pass-through mode and kill the generator. As soon as the generator is dead, they need to provide the power from battery.

What we would expect to happen was that the generator gets demanded with ~ 7.5A per phase and the rest is provided from battery.

We have also a test without generator killing, namely when the load can be coped with by the generator, but the main problem (Quattros not assisting) remains.

no-power-assist2.mp4 (122.5 KiB)

no-power-assist.mp4 (233.4 KiB)

4

2 |3000

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

Daniël Boekel (Victron Energy Staff) ♦♦ commented · Oct 13, 2020 at 08:04 PM

Hi @PetaJoule

About passthru, this is a state you generally want to avoid, here is some documentation:

https://www.victronenergy.com/live/ve.bus:grid-codes-and-loss-of-mains-detection#how_does_lom_detection_relate_to_passthru

and Q4 here:

https://www.victronenergy.com/live/ess:design-installation-manual#faq

0 ·

petajoule Daniël Boekel (Victron Energy Staff) ♦♦ commented · Oct 14, 2020 at 05:33 AM

It's not an ESS! Never was, probably never will be. Conceptually it's an off-grid system.

LOM is disabled for AC-IN1 - where the passthru happens.

0 ·

Daniël Boekel (Victron Energy Staff) ♦♦ petajoule commented · Oct 14, 2020 at 07:33 AM

Hi @PetaJoule

Than please don't use a grid code!!! you're making it extra difficult on yourself.

select grid code 'none'

when a grid code is selected, a lot of functions work differently due to regulations.

0 ·

petajoule Daniël Boekel (Victron Energy Staff) ♦♦ commented · Oct 14, 2020 at 08:04 AM

Ok, will try. I was under the impression that disabling LOM means there is no "Grid code / regulation" for said AC-In. I mean the warning text when you disable it says so.

0 ·

question