An independent lightning risk assessment recommends measures, not products
An independent lightning risk assessment is run by someone who gains nothing from the amount of protection it recommends.
The standards ask for enough protection, not the most
Every risk method in this family works the same way at the end. You compute the risk, compare it with a tolerable value, and add protection until the risk is within it. IEC 62305-2:2024 clause 7.3 asks for the risk R to be reduced to no more than RT, zone by zone, and clause 9.3 asks the same of the frequency of damage F against FT. NFPA 780-2026 L.6.5 compares each relevant type of loss with its own tolerable value. AS 1768:2021 gives typical tolerable values per type of loss in Table 2.1, and its worked example at C.3.4.2 builds protection up step by step until the risk becomes acceptable.
None of these procedures requires more than that. An owner can still choose extra protection, and NOTE 3 to IEC 62305-2:2024 clause 7.3 says it can be useful or appropriate, but that is the owner's call, not the assessment's. The inputs carry judgement, and so does the choice between measures. When the assessor also sells the protection, those judgement calls can lean one way.
The least protection that meets every tolerable value
| Standard | What counts as compliant | How the least protection is chosen |
|---|---|---|
| IEC 62305-2:2024 | R within RT in every zone (clause 7.3), and F within FT where it is judged (clause 9.3) | |
| NFPA 780-2026 Annex L | Each relevant type of loss within its tolerable value (L.6.5) | Sets are tried by the number of measures changed, one, then two, and so on. The first count with a compliant set wins, and within it the set leaving the highest compliant risk, the least protective one |
| AS 1768:2021 | Each type of loss within its tolerable value (Table 2.1) | Protection levels are tried from the weakest up, with surge protection added in the order of the C.3.4.2 example. Where a lower level with more surge protection and a higher level with less both work, both are shown and the engineer chooses |
Measures named by the standard, never by brand
No AI computes or ranks a candidate. The search runs the calculation for each set of measures. The AI layer can describe the chosen set in plain language, but it cannot change which set was chosen or any figure behind it.
Where Lumex™ chooses, and why it tells you
A ranking of measure types
Your fire provisions stay yours
Built from what is there
For IEC 62305-2:2024 and NFPA 780-2026, the search starts from the protection already installed and only adds to it. For AS 1768:2021 it starts from no lightning protection system and no surge protection, as the C.3.4.2 example does, and also tells you whether the protection you declared already passes.
A recommendation, not a design and not a sale
Related principles:
why figures must be traceable and reproducible,
why the engineer signs and the reviewer checks,
why the current edition is held as data and
why the audit log is only added to.
For the methods themselves, see
Questions answered
Why should a lightning risk assessment be independent of protection products?
Does Lumex™ recommend specific brands or products?
How does Lumex™ decide how much protection to recommend?
What if no combination of measures meets the tolerable risk?
Is the Lumex™ recommendation a protection system design?
Does Lumex™ rank cheaper measures above dearer ones?
Get started today