There are numerous models that can be used to assess the consequence and risk of leaks and releases.
One such model is HYRAM which is publicly available from Sandia and the US DOE.

The suitability of flame arrestors depends on the design of the system, but generally flame arrestors are
rarely needed for hydrogen systems when there is a 100% hydrogen atmosphere upstream of the vent,
and when the downstream vent system is designed to withstand internal ignition. Flame arrestors can
also cause potential blockage or restriction of flow, so relative risks need to be carefully assessed when
they are installed on vent lines.

There is no consensus on the “correct” answer. Small leaks of short duration have a much lower
probability of ignition compared to large releases. Ignition probability is affected by the operating
conditions, whether the release is from a leak or vent stack, and the surrounding environment. Since the
probability of ignition is high, hazard analyses will usually assume the hydrogen will likely ignite and
should consider the potential consequences of ignition to the system and surroundings.

Frequency and severity off consequences are situational and subject to the safety review team’s best judgement. One measure of severity is an estimate of the energy released if ignited. Assuming the worst-case mix to be stoichiometric, the energy content of a 500 mL of hydrogen in air is about 0.2 Wh (700 Joules), comparable to the energy release of a wooden, blue-tipped matchstick (~1kJ or 1 Btu). This may not be very significant in a safe location such as an operating fume hood. See Risk assessment and risk ranking at H2Tools, Best Practices: Ranking Risks, for more information. 

There are many designs of storage systems where multiple vessels might be needed to obtain the required storage quantity. Regulations differ between vessels and modules which are intended for stationary or transportation purposes. Similarly, there are differences in codes globally. The issues of requiring shutoff valves on individual vessels and requiring TPRD’s are linked since generally vessels need to be in communication with the pressure relief system. A valve on each vessel will require one or more TPRD on each vessel.

There are trade-offs between the risk of having a large number of valves and TPRD’s, each of which is a potential leak/release source, and having a large number of vessels manifolded together with fewer valves/TPRD’s but then larger banks that contain larger quantity of hydrogen that will be released in a single event. The main disadvantage of multiple vessels being connected to a single TPRD is that a much larger release is possible since each vessel can’t be isolated. This may also require a much larger device to be able to vent the multiple vessels which then means a higher release rate with the associated larger vapor cloud, risk of explosion, and radiation profile. The decision is dependent upon the risk assessment and might be
different depending on the location and application. When a tube has its own, individual TPRD and isolation valve (which must be closed during transport), a smaller release of hydrogen would occur.