VFDs and Prospective Short-Circuit Current

[Shackled Designer]

What should I make of this?

Recently, I found a Schneider Electric tech bulletin published in 2013 entitled "Variable Frequency Drives and Short-Circuit Current Ratings." (One may quickly find it on the web.) In it, the paper defines some basic terms such as Prospective Short-Circuit Current (PSCC), which I frequently see called available fault current (AFC). Then, the paper starts discussing the input ratings of VFDs and makes the following assertion:

The level of PSCC can have a significant thermal impact on a VFD's input diodes and capacitor bank. The VFD's input current increases significantly as the level of PSCC rises. This is caused by the input diodes conducting only when the input voltage is higher than the DC bus. The current is then limited only by the system impedance. Applying a VFD on an electrical system with a higher PSCC than the VFD input rating may cause overheating of the diodes and capacitor sections, and reduce the life expectancy of the VFD or damage the VFD.

The paper then provides illustration of this with some nice graphs showing peak current at about 30A for an AFC of 5kA and nearly 70A for an AFC of 100kA.

Before reading this, I would have rather expected that there was a bit more current limitation built into the input circuitry of the VFD. Also, having read previous posts in this forum regarding VFDs, I see that many here are of the opinion that modern VFDs do not so strongly affect supply harmonics, but I wonder at this if, as the paper maintains, VFD input current draw is a function of available fault current.

Is the paper correct? Or is it probable that the bulletin is focusing on older VFD designs?

If the paper is right, then there does seem to be more justification of the use of a line reactor on the infeed, not so much for harmonic mitigation of the drive, but for controlling the VFD from drawing excessive current.

I welcome the comments and discussion.

Best regards,
Shak

 

[Besoeker]

What should I make of this?

Recently, I found a Schneider Electric tech bulletin published in 2013 entitled "Variable Frequency Drives and Short-Circuit Current Ratings." (One may quickly find it on the web.) In it, the paper defines some basic terms such as Prospective Short-Circuit Current (PSCC), which I frequently see called available fault current (AFC). Then, the paper starts discussing the input ratings of VFDs and makes the following assertion:



The paper then provides illustration of this with some nice graphs showing peak current at about 30A for an AFC of 5kA and nearly 70A for an AFC of 100kA.

Before reading this, I would have rather expected that there was a bit more current limitation built into the input circuitry of the VFD. Also, having read previous posts in this forum regarding VFDs, I see that many here are of the opinion that modern VFDs do not so strongly affect supply harmonics, but I wonder at this if, as the paper maintains, VFD input current draw is a function of available fault current.

Is the paper correct? Or is it probable that the bulletin is focusing on older VFD designs?

If the paper is right, then there does seem to be more justification of the use of a line reactor on the infeed, not so much for harmonic mitigation of the drive, but for controlling the VFD from drawing excessive current.

I welcome the comments and discussion.

Best regards,
Shak

Yes, the current into the DC link capacitors can be very "peaky"
Fitting input line reactors is one way of limiting the peaks. A DC link reactor can be more effective.

 

[GoldDigger]

A DC link reactor can be more effective.

Just because you can use one shared reactor after the rectifier bridge instead of three on the phase lines, each of which is active only 2/3 of the time?
Or is there more to it than that?

 

[Besoeker]

Just because you can use one shared reactor after the rectifier bridge instead of three on the phase lines, each of which is active only 2/3 of the time?
Or is there more to it than that?

A little.
A reactor on the AC side can have limited inductive reactance value because it drops the input voltage.
The DC reactor doesn't have that limitation.

 

[GoldDigger]

A little.
A reactor on the AC side can have limited inductive reactance value because it drops the input voltage.
The DC reactor doesn't have that limitation.

If you think of what the line reactor sees as alternating short duration pulsed DC the apparent difference seems to me to go away.
The time scale you look at is important, and each line reactor is not seeing a sine wave by any stretch of the calculation.
Both the line reactor and the DC reactor will also tend to stretch the trailing edge of the pulse, which is good.

 

[Besoeker]

If you think of what the line reactor sees as alternating short duration pulsed DC the apparent difference seems to me to go away.
The time scale you look at is important, and each line reactor is not seeing a sine wave by any stretch of the calculation.
Both the line reactor and the DC reactor will also tend to stretch the trailing edge of the pulse, which is good.

Not quite like that.
For the purposes of an explanation take a three phase input.

Imagine the DC choke to be infinite inductance. So the current in it will be level.
In that case the current in each input line is a rectangular 120deg pulse. With sharpish edges.

Of course it isn't quite like that in real life. The current takes time to commutate from one phase to the next owing to supply and any additional inductance and semiconductor reverse recovery times- overlap angle. But, in my experience, it isn't very long. Usually some tens of us at most. Sort of think I've checked quite often so I'll see if I can dig up some recordings.

The difference is that the DC inductance can be many times that of a line reactor. We sometimes use both (for different functions) and typically the AC reactors are maybe 50uH or less. The DC choke, several mH.

 

[GoldDigger]

I was just thinking that to the extent we are talking about current pulses near the peak of the AC waveform (corresponding actually to a lightly loaded condition?) then the effect of the DC choke in impeding the current flow (and hence the voltage drop across it) will be comparable in magnitude to the voltage drop on the AC side for the same current pulse.
In the extreme where the choke is carrying current continuously, I do see your point. Yet there will be a reduced voltage at DC bus corresponding to the fact that the choke will limit the voltage on the DC bus to the average value of the full wave bridge DC waveform rather than the peak.
Much like the difference in output voltage between a choke input and capacitor input power supply driven by the same AC input and running at or near design load.

 

[Jraef]

I've never heard of this issue referenced to the AFC (PSSC as they call it), but I was taught that it relates to the kVA size of the transformer, which I suppose boils down to the same thing in general. Originally I was taught that if the transformer kVA was over 4-5X the kVA of the VFD, the potential for rectifier damage is greater unless there is a reactor ahead of the drive. More recently I attended a VFD power quality training seminar where we were told the rule of thumb to use is 10X the kVA. In that lecture they were however discussing transients, not normal current draw, so that may account for the difference. They showed an example of a 25HP 480V VFD, rated 40A with diodes that were actually rated 60A, being fed by a 1000kVA transformer. During grid switching voltage transients, they recorded peak currents through the rectifier components of 805A. By adding just a 3% Line Reactor, those peaks were knocked down to 55A.

The problem with a DC bus choke in that scenario is that it is located BELOW the rectifier, not ahead of it. So while it might help mitigate the normal current draw issue, it does not provide the same level of transient protection.

 

[GoldDigger]

The rectifier would be exposed to transient voltages, yes. But any resulting transient currents would flow through an inductor either way unless the voltages were high enough to break down the rectifier diodes.

 

[Besoeker]

The problem with a DC bus choke in that scenario is that it is located BELOW the rectifier, not ahead of it. So while it might help mitigate the normal current draw issue,

Which was the point raised by the OP. And a DC choke addresses that much better than line chokes.

it does not provide the same level of transient protection.

The input bridge ought to be equipped with transient overvoltage protection but that is a quite different function.

Actually, I spent a week or so in Hong Kong dealing with exactly that problem when a major manufacturer, a household name to most, was having an unacceptably high rate of attrition with input rectifier bridges.

It was quite an interesting application as it happens. Seawater pumps providing cooling water for the high rise buildings in the business districts. This meant that any tests had to be done outside business hours. So from about 9:00 pm until 07:00 am. And the drives were underneath the docks. Access was via a manhole cover - about 800 mm diameter or 2ft 7 in. And then a vertical 3 metre ladder. Not a challenge unencumbered but if you need to take instruments and tools, just a bit more so.

Anyway, a few snubbers in the right place fixed the problem. I don't know who eventually got hit with the costs and, perhaps more importantly, the reputation fall out.
I was there as an independent just being paid to do a job. But I did get to sample some of the HK night life that I hadn't previously experienced............