Electrical – Sizing AC fuse for a DC device

4:40am
Rushing.

If you search prior material on Stack Exchange EE you'll find a substantial amount of material on this.
The figures you cite are in the order of right.
Fuse blowing current and fault clearing current are different.
HRC (high rupture capacity) fuses exist to deal with this difference. The ceramic bodied fuses you see in better equipment are HRC.
A glass fuse may blow but sustain an arc of 100's of amps long enough to kill you.

If your pole fuse is 100A and your neighbours is 100A and ... what is the street cct able to supply?

If you draw 50A from your home mains supply and it sags 1%, what current would you expect it to supply if you hard shorted it?

At 50 Hz, 230 VAC, what inductance do you need to add say 1 Ohm reactance to your house feeder circuit. What inductance do you think the feeder has?

A friend had an electrician (stupidly) reverse phase and neutral when wiring up their house.
Steam came out of the cold taps due to electrical heating in the grounded copper "cold" water pipes as current flowed from mains phase via switchboard ground to copper pipes and thence to ground. (really)
and worms crawled out of the ground (really)
and they tell me that the house made groaning sounds.
I imagine that that was probably from water boiling in the cold water pipes.
What current do you think flowed :-) :-( ?

HRC fuses - there will be somje ueful links there.

Wikipedia - fuses

• The breaking capacity is the maximum current that can safely be interrupted by the fuse. Generally, this should be higher than the prospective short circuit current. Miniature fuses may have an interrupting rating only 10 times their rated current.
  Some fuses are designated High Rupture Capacity (HRC) and are usually filled with sand or a similar material. Fuses for small, low-voltage, usually residential, wiring systems are commonly rated, in North American practice, to interrupt 10,000 amperes.

Wikipedia - breaking capacity

• Miniature circuit breakers and fuses may be rated to interrupt as little as 75 amperes and are intended for supplementary protection of equipment, not the primary protection of a building wiring system. In North American practice, approved general-purpose low-voltage fuses must interrupt at least 10,000 amperes and certain types useful for large commercial and industrial low-voltage distribution systems are rated to safely interrupt 200,000 amperes..

ADDED

Stack Exchange:

Similar material.

Fuses: What are the practical differences between Ceramic and Glass cartridge fuses

What is the Thévenin equivalent of the mains power supply? - 1st approximation - a piece of copper busbar :-)

The Impact of Mains Impedance on Power Quality

Useful. See fig 6.
Note transformer impedances specified as a % - this is the % drop in output voltage at rated load.

Added 2:

Thanks for the clarification of breaking capacity and highlighting reactance. I still think 2000 amps is over the top. 200 amps I could understand.

I'd guesstimate that 2000A would probably be getting on the high side in a residential situation. But 200A is far too low.
Far far too low.

If you can get 50A intended current at your home's distribution board and your neighbour's lights do not flicker, what would you get if you shorted it?

People have died from mains arc discharge that was improperly interrupted.

Standards typically allow a 5% V drop at the farthest outlet from the distribution board in a home at rated load.
At 20A rate that implies available current is ~+ 20A/0.05 = 400A.
And that's worst case on house wiring!.

Inrush current is greatest when operating no-load on secondary

Is it a problem of having no load?

Yes, it might be significant.

Most likely the problem is "inrush current" and this is due to saturation of the transformer core. At this point some folk may be shaking their heads and not believing this and they may point out that "surely the current will be greater when the transformer secondary is fully loaded and surely this will produce more saturation in the core?"

The magnetization of the transformer core is due to the inductance \$X_M\$shown below (with a blue box around it: -

This is the equivalent circuit of a transformer and shows other stuff like iron loss and copper losses. When the transformer secondary is open circuit, the copper losses are extremely small and can be ignored; the upshot is that \$X_M\$ receives the full supply voltage because there is minimal volts dropped across the copper loss components and leakage inductance.

This means that the magnetization of the transformer is at its highest because when secondary current flows (due to load), the volts on \$X_M\$ reduce slightly due to \$R_P\$ and \$X_P\$.

But why should the core saturate? Here's what the magnetic field looks like when power is applied when \$V_P\$ is at the top of its voltage cycle: -

The voltage waveform is in black and the magnetic flux (or current) is in red. Forgive my drawing ability. Two scenarios and one produces normal peaks of current (as if the supply had been connected forever) and the other scenario causes the current to hit a big high-spot then settles down.

There's nothing mystical about this - this is what would happen in an inductor if you simulated it (look at the current).

If the magnetizing current peak is bigger, then so is the driving force of the magnetic field. If you looked at what the flux density is like when the ampere-turns are bigger you are likely to see the flux reach saturation and, when this happens you also draw a lot more current which in turn makes the problem bigger. This settles down after a few cycles of the power waveform of course.

What you will probably find is that sometimes the current limit trips and sometimes it doesn't - it's all down to where on the voltage cycle you applied power to the transformer.