Adiabatics not complying on short-length circuits/faults close to CU

Firstly, please go easy as I'm still a trainee! Working on my L3 2365
I'm having trouble understanding the rationale behind adiabatics calculations for CPCs, because it seems to me that if you had a fault close to the CU on a longer circuit, or a fault on a short (in length) circuit, you'd have a very low Zs that would fail the equation in lots of instances.

Say, for example, you have a lighting circuit in 1.5mm T&E that is 20m long, protected by a 6A MCB. Let's say Ze is 0.35Ω, and we can calculate that R1+R2 is 0.6Ω via Table I1 in OSG. So expected Zs is 0.95Ω, and PFC is 242A.
If we plug it into the Adiabatic with 0.1s* then we get: sqrt(242 * 242 * 0.1) / 115 = 0.66Ω min CSA required - all good!

But, what if the fault occurs just 2m into the circuit? Nail through a cable etc.
Suddenly, the Zs is down at 0.41Ω, the PFC is up to 561A, and if you do the adiabatic it reckons you need 1.54mm on the CPC (yes, I know 7671 table 54.7 caps it at the line conductor CSA, but still)

Or, what if you had a short-length circuit (say 2m) - how would you get it to comply without using much beefier cables - and how does that make sense?

Am I missing something here, or doing my calculations wrong? What's the point in doing this overheat calculation only for the end of the cable, when a fault part-way along will likely fail it and the CPC overheat in the duration before the OCPD triggers?

*I've used 0.1s from the graph of a 60898 breaker, but most discussion of adiabatics uses the earthing arrangement timings - e.g. 0.4s for a final on TN system, which makes the requirements much harder to reach. Not fully sure which I should be using.

First point is the usual graphs for MCB & RCBO often only go down to 0.1s but the reality is they open in under 20ms if you hit the "instant" magnetic trip point. This shorter time is very important as it explains why most cables are fine for most PFC.

Generally fuses are good at energy-limiting, basically the wire burns out faster than the cable and largely nothing you do to push more current through stops that, short of the fuse physically exploding. So your risk with fuses & adiabatic is low faults where they take longer to heat to the point of runaway and then blow.

MCB only really limit energy much once you hit the magnetic point, so for B-curve at 3-5 * In for example, and then the operating time is not very sensitive to current. So while to do provide some more energy limiting as the PFC increases, you don't get the same level I2t limitation of a fuse. You can get a situation where you meet adiabatic for current from the magnetic trip only to a certain PFC.

Here is an example from the Hager catalogue showing how the I2t of a C-curve MCB varies with fault current:

Thank you - that makes a lot of sense for both fuses and MCBs
I've just had a look at a few manufacturers and the quality of info about this on datasheets varies, but what you've given is very helpful in terms of figuring out whether a CPC is truly inadequate.

I was beginning to think that this might make adiabatics a bit OTT for domestic - but I've just checked my Ze at home and it's 0.04 (substation within 3m) so some quick calcs suggests my lighting circuits might not make the cut even if it's 20ms (or even 5)! Good to understand - cheers.

For most domestic work the "standard circuits" in the On-Site Guide Table 7.1(ii) are all pre-checked for adequacy in these aspects, also there is the Table B7 with CPC sizing guides for similar situations.

Having Ze = 0.04 ohms is a bit brutal for smaller cables but as long as you are using reputable energy-limiting MCB they will be safe. While that is pointing to 6kA sort of PFC your incoming fuse will help protect the MCB against the worst-case as well, that is almost what defines a "consumer unit" rather than a general distribution board, that is it coordinated with the DNO fuse types to be safe to supply PFC/PSCC to 16kA.

In the more general case, if you check the better supplier's catalogues and data you will see "cascading values" that give you the breaking capacity limit for two devices in series, such as a supply fuse and MCB/RCBO/etc, or an incomer MCCB and outgoing MCB, etc.