To what extent does HV protection mitigate LV faults?

The HV protection will be designed to protect the transformer, and the tails up to the first LV protection. As with the DNO protecting your first 3m, anything after that may be protected but unless you talk to the designer it is not guaranteed.

However because of the impedance of the transformer coupled with transformation ratios, what looks like a HUGE current at LV may well look like a mere overload current at HV. To compound this some HV protection only provides protection for fault current, relying on the LV side and other devices for overload. Becaue we don't want the HV protection operating at the drop of an LV hat (selectivity etc, particularly given that the HV protection could well be based on HV fuses or time-limit fuses), the upshot is that HV protection for the LV components will be slooooow.

Furthermore, assuming a delta-wye winding, LV earth faults appear to the HV system as phase current, so unless it's a close in and solid fault, may not even register at HV at all.

As others have noted, this means that the energy released (damage caused) before the fault is cleared could be considerable. This is why alternative protection schemes are fitted for more valuable transformers / at risk locations, such as Restricted Earth Fault (REF), which is a bit like an RCD facing backwards to trip very quickly on transformer LV winding earth faults which are very hard to detect at HV but cause a lot of damage if unchecked.

In principle it may be possible, subject to practical limitations like fuse sizes / setting ranges, to provide backup protection for the main LV protection, perhaps and ideally that is done. But it needs coordination and all of the information to be available at the right time, e.g. LV switchboard withstand rating.

The HV protection will be designed to protect the transformer, and the tails up to the first LV protection. As with the DNO protecting your first 3m, anything after that may be protected but unless you talk to the designer it is not guaranteed.

However because of the impedance of the transformer coupled with transformation ratios, what looks like a HUGE current at LV may well look like a mere overload current at HV. To compound this some HV protection only provides protection for fault current, relying on the LV side and other devices for overload. Becaue we don't want the HV protection operating at the drop of an LV hat (selectivity etc, particularly given that the HV protection could well be based on HV fuses or time-limit fuses), the upshot is that HV protection for the LV components will be slooooow.

Furthermore, assuming a delta-wye winding, LV earth faults appear to the HV system as phase current, so unless it's a close in and solid fault, may not even register at HV at all.

As others have noted, this means that the energy released (damage caused) before the fault is cleared could be considerable. This is why alternative protection schemes are fitted for more valuable transformers / at risk locations, such as Restricted Earth Fault (REF), which is a bit like an RCD facing backwards to trip very quickly on transformer LV winding earth faults which are very hard to detect at HV but cause a lot of damage if unchecked.

In principle it may be possible, subject to practical limitations like fuse sizes / setting ranges, to provide backup protection for the main LV protection, perhaps and ideally that is done. But it needs coordination and all of the information to be available at the right time, e.g. LV switchboard withstand rating.