Hydrogen has a very broad range for combustion in air, with a LEL (lower explosive limit) of about 4% H2 in air and a UEL (upper explosive limit ) of about 96% H2 in air (or 4% air in H2 if you prefer) - well, actually a more recent reference from Matheson indicates 4% to 75% - I've learned something new today, thanks! It also has a very low ignition energy (energy input required to get it going). So, your exhaust system can fill up with lots of H2 and then get set off easily. This can quickly go from a small flame zone to deflagration in the exhaust system (sub-sonic) to detonation (supersonic reaction front), and the reaction will continue until you are outside of the LEL/UEL range.
Look over the Matheson reference, and you will see that there are very few other gases with such a broad LEL/UEL range. For example, methane (or natural gas) is only 5% to 15%, propane is 2.1% to 9.5%. So, if you fill up a space with methane or propane, only so much of it will burn, then it will quench as the mix goes out of the LEL-UEL limits.
The implication is that a hydrogen-air mixture can sustain detonation/deflagration across a much broader range, so it can (and by Murphy will) make a lot more flame and shock and awe. Bottom line, don't have anything in the way of all that excitement exiting the facility as quickly and completely as possible.
An article in Chemical Engineering discusses the issues briefly, although doesn't go into the engineering details (a bit odd for an engineering magazine?).