What changed in IEC 62305:2024

To see why a revision of the standard can change an answer, it helps to hold the shape of the method in mind. In IEC 62305-2 each risk is built from components, and every component is the same three-step chain: how often a dangerous event occurs, multiplied by the probability that the event causes the damage, multiplied by the loss that results. Written plainly, that is rate times probability times loss. The risk for a type of loss is the sum of its components, and the verdict is that risk compared against a tolerable level the standard sets.

A revision matters when it touches one of those links. If the edition changes how the rate is computed, every component scales with it. If it changes which probabilities a measure is allowed to reduce, the design conversation changes. If it changes how losses are grouped into a single risk, the sum lands somewhere new. The 2024 edition does all three, which is why it is more than an editorial pass. The sections below map each change onto the link it moves.

The tolerable line did not move, but several of the inputs feeding the risk did. A computed risk that sat just under the line on the 2010 method can sit just over it on the 2024 method, and the structure that read as safe now reads as needing protection. That is the whole reason the edition is load-bearing.

The most consequential change is in the rate, the first link of every component. Earlier editions drove the rate of dangerous events from a lightning flash density, written N_G, the number of cloud-to-ground flashes per square kilometre per year. The 2024 edition replaces it with a ground strike-point density, written N_SG, the number of points struck on the ground per square kilometre per year.

The distinction is physical, not bookkeeping. A single lightning flash does not always reach the ground at one tidy point. One flash can have several ground strike points, each capable of doing harm, so counting flashes undercounts the places a strike actually touches down. By counting strike points, N_SG describes the hazard a structure is exposed to more faithfully. A collection area pulls strikes from the ground around a structure, and what matters for risk is how many of those ground contacts it gathers, not how many parent flashes they came from.

Because the rate of dangerous events scales directly with this input, swapping the density term changes every component at once. Where the local ground strike-point density works out higher than the flash density figure that was used before, the computed rate rises across the whole assessment, and a risk that was comfortably under its tolerable value can be pushed over it. The change is not guaranteed to move a given building in one direction, but it is guaranteed to be the term the whole result scales with, which is why it is the change most likely to flip a verdict.

The risk of injury to people, death included, gathers several ways a strike can harm them: injury from touch and step voltages, and the consequences of physical damage such as a fire that a strike sets off in an occupied building. The 2024 edition brings injury to people and loss from fire into a single combined risk view, so a structure that carries both an occupancy hazard and a fire hazard is assessed as one risk rather than as two partial pictures looked at in isolation.

This is a change to how the components are summed into the verdict, the last link of the chain. For a building that holds both people and a fire load, a warehouse with staff and flammable stock, a venue packed with a crowd, a workshop with a combustible process, the combined treatment can push the computed risk of injury to people higher than an approach that weighed the two hazards separately. A structure whose fire risk and life risk each looked tolerable on their own can exceed the tolerable line once they are assessed together, which again can turn a pass into a fail.

Not every site is dominated by the threat to life or the threat of a fire. For a data centre, a control room or any operation where the cost of a strike is measured in downtime, the thing being protected is the availability of the internal electrical and electronic systems. The 2024 edition adds a measure aimed at the frequency of damage to those systems, giving the assessment a way to reason about how often a strike would disrupt the electronics rather than only whether it would injure someone or burn the building.

This recognises that an internal-system failure can be the dominant concern in its own right. A strike does not have to start a fire or hurt anyone to take a facility offline: the electromagnetic pulse a strike radiates, known as LEMP, can induce a surge that knocks out controls, servers or instrumentation. For sites where that is the real exposure, the new measure makes the assessment speak to what they actually care about, and it can change which measures are specified, since coordinated surge protective devices, shielding and bonding become the components that move the number rather than fire measures or an air-termination layout.

The 2024 edition formally recognises thunderstorm warning systems, in line with IEC 62793, as a measure that reduces risk. This acts on the probability link of the chain. By detecting an approaching storm and triggering temporary precautions ahead of it, moving people to safety, pausing an exposed activity, clearing an open area, a warning system lowers the probability of harm to people during the window of greatest exposure.

The 2010 edition gave no credit for this, so a warning system did nothing for the computed risk under the old method even where it genuinely reduced the danger to people. Under the 2024 edition it becomes a recognised way to bring the risk of injury to people down, particularly for sites where people are exposed in the open and can be moved in time, such as sports grounds, open-cast sites and outdoor venues. A structure that relies on such a system can compute a lower risk under the current method than it could under the old one, which is the reverse case: a change that can turn a fail into a pass without any physical change to the building.

Three of them, the density term, the combined risk and the warning systems, move links in the calculation that decides a verdict. The frequency-of-damage measure adds a concern the method could not previously express. Taken together they are why an assessment on the Edition 3.0 (2024) method and one on the Edition 2 (2010) risk model are not interchangeable for the same building.

Put the changes together and the conclusion is direct: an assessment built on the 2010 method can compute a different answer for the same building today. The density term feeding every component is different, injury to people and loss from fire are now summed into one risk rather than weighed apart, and a warning system that earned no credit before can now reduce the risk to people. Nothing about the building has to change for the verdict to move. The change is in the method, and the method is what the tolerable line is applied to.

That has practical weight beyond being current for its own sake. An auditor, an insurer or an authority having jurisdiction checks an assessment against the edition in force, which is the 2024 third edition. An assessment computed by the older method can be rejected as out of date, and because the method genuinely changed, it can also leave a structure under-protected or over-protected relative to the current standard. A building that reads as safe on a 2010-era calculation may need protection under the 2024 method, and signing off on the old number is a real exposure for whoever put their name to it.

The safe position is to compute on the current edition, with the reasoning traceable clause by clause so a reviewer can follow it. Where an assessment is being relied on for approval, insurance or a client specification, or where the structure or its surroundings have changed, recomputing it under the 2024 edition is how you confirm the verdict still holds. The edition is not metadata on the report. It is part of the answer.

Lumex's Voltrace engine computes the IEC 62305-2 risk method on the current 2024 third edition: it drives the dangerous-event rate from a ground strike-point density, applies the combined view of injury to people and loss from fire, accounts for the frequency of damage to internal systems, and credits thunderstorm warning systems in line with IEC 62793. The engine reproduces all three Annex F worked examples to the published figures, and known limitations are published in our conformance register. Every risk is compared against its tolerable value and the reasoning is traceable clause by clause, so a reviewer can follow every number back to its source. See how an IEC 62305 assessment works end to end, or explore the platform.

New to the standard? Start with what is IEC 62305, then read the method in full in the IEC 62305-2 risk assessment.