Consequences of common mode circulating currents

[somlioy]

Hello

Say you have this topology in an industrial supply:
LiFePo4 Battery 600V/200Ah Nominal -> Inverter (400V 3ph) -> LCL/Sine Filter -> Distribution board <- 400V/400kVA diesel gen.

What is the consequences of common mode currents and voltages in the dc-part of this system?
What levels of noise is to be expected on the dc-bus?

[capt bullshot]

Depends on the architecture of the Inverter. Some of them can have the + and - rail of the battery "jumping" around at high dU/dt rates and switching frequency including heavy overshooting (way higher peak voltages than expected). Others may be rather quiet. High frequency stuff on the battery DC bus can lead to all kinds of capacitive leakage, causing all kinds of EMC issues or overheating capacitors (especially Y rated caps from the DC rails to earth). Control circuitry of the inverter usually should be isolated from the DC and AC buses, so less problems here.

[floobydust]

Where is the physical earth-ground connection here?
Diesel gen-sets are usually (portable) on a skid and no ground rod.

[SparkyFX]

Some inverters just can´t handle being supplied with power at the output. It overloads the body diodes in the FETs of the output stage (if it has FETs).

Additionally one needs to be master for the grid AC frequency. Solar inverters are capable of doing that, but diesel generators might not.

[somlioy]

All this equipment is mounted on a steel barge so battery cabinet (not the batteries it self) and generator is both earthed not far from eachother. However the cabel to the distribution board is probably around 20-30 meters from each.
Not quite sure if the neutral point of the generator i grounded. Sometimes it is sometimes it isnt.

The inverter is an Siemens inverter with active front end so it is capable to driving currents in both directions. (Active) Power is NEVER flowing into the genset.

As you can see in the attachments the dc bus has quite alot high frequency noise. Two first pics are volt/amps when charging the batteries, and the last two are volt/amps when discharging.
Would replacing the inductor in the LCL filter closest to the distribution board with an 1:1 transformer eliminate the common mode noise "rotating" in the batteries and instead "rotate" in the transformer? (Having the "loose end" of the system in the transformer).

[capt bullshot]

Your data looks like a long term recording, but I cannot identify the time axis' scaling (is it seconds?).

In this case, the recording is rather meaningless in terms of distortion caused by common mode currents, I believe it's just the normal fluctuations caused by whatever the load / generator does. A transformer wouldn't change this behaviour. If nothing gets unusually warm (cables, filter capacitors, inductors), I'd consider the setup OK.

If you suspect too high common mode currents, you'd have to measure the actual currents in the earth / neutral / DC bus using a current clamp and suitable equipment (True RMS DMM, scope). Also measuring the inverters AC side with a good and safe differential probe might be interesting, but not necessarily useful for you. For measuring the common mode battery current, you'd have to put both conductors (Bat + and Bat -) through the clamp, so only the common mode current remains to be seen, assuming the battery doesn't have any more connections to the inverter.

[somlioy]

Time axis is in milliseconds. So about 3.5 min/div for charging, and 1.5 min/div for discharge.
There isnt much fluctuations in loads. Atleast during those recordings. Lights and heating is the only loads.
What about running the DC-cables to the battery through a series of ferrite rings? I'd expect those to get warm but will it be effective?

[capt bullshot]

OK, a rather long term recording. I don't have experience with your particular setup, so I can't tell if the noise band is normal or not. Do you have a reasonable suspect something is wrong with your system?

If your problem is circulating line frequency currents, ferrite won't help. If it is switching frequency, yes they may get hot and may or may not help. Anyway, it's important to apply them common mode, running both (and all) + and - cables through the same opening to cancel out the operating current. The operating current would immediately saturate the ferrites and render them useless.

I suppose the battery itself is isolated from ground, so only capacitive leaking is possible. Normally such a setup should not have an issue with that. Otherwise a field engineer from Siemens should have a look at your setup.

[SparkyFX]

If i get the question and setup right you get at least the forward voltage/RDSon drop of your inverter/charger as a common mode offset on the DC bus, that would be normal, depending on what your reference is. It won´t be lower as that.

[somlioy]

Reason I asked about the transformer is because of this guy claiming you need a transformer in a system like this: (From 6min 56s.)
https://youtu.be/2KVmg34kopU?t=6m56s

[capt bullshot]

OK, I see.
There's two reasons claimed:
- Voltage adaption for optimum operation point
- The battery vendors don't like the common mode potential on the battery

Well, the second reason really depends on the battery technology and construction. You should ask your battery supplier about that, I can't tell any general applicable rule for that. It also depends if you are doing battery storage for 3 phase or single phase operation. With single phase you'd have way more line frequency common voltage mode than with your 3 phase operation. There might also be safety reasons to isolate the battery (applicable for lower voltage, e.g. 48V, systems).

If there aren't any concerns by the battery supplier, I'd say the non-isolated operation mode is OK, and I don't think the noise in your recordings is related to common mode issues.

Edit: there's a lot of historical grown good practice of having the positive terminal of a battery earthed in UPS systems. This mostly applies to stationary lead acid batteries and has its rationale in the effects of electrochemical induced corrosion. I don't think this is applicable to a properly built LiFePo4 in the same way, but better ask your battery supplier.

[SparkyFX]

Well they pictured a transformer as an impedance converter - sort of, maybe it even acts as an isolation transformer, which gets rid off any common mode voltage in an AC signal. No voltage differential for the coil -> no current -> no magnetic flux (none that rotates in the transformer).


On a boat with (salt) water and steel walls any ground fault might have more drastic implications than usual, maybe that part is what is referred to as common mode as well, so that part is isolated from ground and therefore makes it less dangerous.