Principles

How Lumex™ verifies its risk engines

A risk figure is only worth signing if the engine behind it gives the same answer as the standard. This page sets out the worked examples each Lumex™ engine is tested against, the figures the standards print, the tolerance each check allows, and what is not yet covered.

A lightning bolt striking behind an industrial building, the exposure a lightning risk assessment quantifies

Every Lumex™ engine is tested against the worked examples printed in its own standard.

A lightning risk standard is a long chain of equations and tables, and a small slip anywhere in that chain changes the verdict. The only outside check that counts is the standard's own arithmetic. Where a standard prints worked examples, Lumex™ runs the same inputs through its engine and compares the result with every figure the standard prints. Where it prints none, the engine is compared with an independent implementation instead, and this page says so.

Each check below is a test in the Lumex™ engine test suites. The published figures were read from the standards themselves: the published IEC 62305-2:2024 text and the AS 1768:2021 standard and its risk tool. For why this matters to a reviewer, read traceable, reproducible risk.

At a glance

What each engine is checked against

IEC 62305-2:2024

All three Annex F case studies

The house, office and hospital studies, each unprotected and protected, compared with the printed collection areas, dangerous events, risk components, zone risks and frequencies of damage, within 1.5 percent.

AS 1768:2021

The tool and Figures C.2 to C.8

400 cases computed by the Lightning Risk Assessment Tool v5.0, the case saved in its workbook, and every risk printed in the seven worked examples of Appendix C, all matched exactly.

NFPA 780-2026

An independent implementation

Nine cases computed by a separate implementation of Annex L, written from the annex rather than from the engine. The 2026 annex prints no full worked example to compare with.

IEC 62305-2:2024

The Annex F case studies

Annex F of IEC 62305-2:2024 (Ed.3) takes three structures through the whole method, first with no protection and then with the measures the study chooses. It prints the collection areas, the number of dangerous events, each risk component, each zone's risk and, for the office and the hospital, the frequency of damage. Lumex™ runs each study through its engine and checks every one of those printed figures.

Annex F prints to three significant figures and warns in F.1 that rounded intermediate values move the last digit. So each check allows 1.5 percent of the printed figure, or half of the last printed digit where the annex prints too few digits for a percentage to be meaningful. Risks are printed in units of 10⁻⁵ per year, and the tables below keep that unit.

F.2

The house

A house of 15 by 20 by 6 m with an NSG of 8, no lightning protection system, and an aerial power line and an aerial telephone line. The study protects it with SPDs of LPL IV at the entrance of both lines.

Figure checkedPublishedAnnex F tableVerdict the study reaches
Collection area AD2.58 × 10³ m²Table F.4-
Direct flashes ND2.06 × 10⁻² per yearTable F.5-
Risk R, unprotected1.793Table F.8Above RT = 10⁻⁵, protection needed
Risk R, with SPDs at the entrance0.149Table F.9Below RT
F.3

The office building

A reinforced concrete office of 20 by 40 by 25 m with an NSG of 4 in a suburban area, five risk zones, and a buried power line in a high-voltage and a low-voltage section. The study protects it with a Class II LPS, SPDs at the power line entrance and coordinated SPDs of LPL II on both internal systems.

ZoneRisk R, unprotected (Table F.21)Risk R, protected (Table F.23)
Z1, entrance outside0.002Approximately 0
Z2, roof2.2590.113
Z3, archive6.5260.592
Z4, offices0.2020.018
Z5, computer centre0.1560.014

The frequency of damage is checked too. Each of the three inner zones prints F = 0.584 per year unprotected (Table F.22), above the owner's FT of 5 × 10⁻², and 0.014 per year once protected (Table F.24). The protected frequency only reproduces when the SPDs on both the power and the telecom system are counted, as clause 8.5.1 combines them, which is one of the checks that separates a full implementation from a simplified one. The collection areas and dangerous events of Tables F.13 and F.14 are checked as well.

F.4

The hospital

A reinforced concrete hospital of 50 by 150 by 10 m with an NSG of 8 in a suburban area, five risk zones and a buried power line. The study protects it with a Class II LPS, bonding at the power line entrance, a warning notice on the roof and coordinated SPDs.

ZoneRisk, unprotected (Table F.35)Risk, protected (Table F.37)
Z1, entrance outside (RAT)0.0360.002
Z2, roof (RAD)18.3570.092
Z3, rooms block (R)36.5280.692
Z4, operating block (R)108.8340.266
Z5, intensive care (R)306.7870.685

Each protected zone falls to or below RT = 10⁻⁵. The frequency of damage drops from 0.325 per year in Z3 and 0.306 in Z4 and Z5 (Table F.36) to 0.0032 and 0.0006 (Table F.38). Every risk component that Tables F.35 and F.37 print for a zone is checked on its own, so a correct total cannot hide two errors that cancel.

AS 1768:2021

The risk tool and the Appendix C examples

AS 1768:2021 gives the shape of its risk method in Appendix B, and the Lightning Risk Assessment Tool v5.0 that comes with the standard holds the detail. The standard's own worked examples were produced with that tool. So the Lumex™ engine follows the tool cell by cell, keeping the order of every operation, and is checked three ways, each with no tolerance at all.

  • 400 tool cases. Seeded input combinations, each computed by the tool itself, including sites with no power service and sites with no other overhead or underground lines. The engine must agree with the tool to the last bit.
  • The workbook's own saved case. The results Microsoft Excel computed and saved in the tool as distributed, which proves the engine against the tool's native evaluator.
  • Figures C.2 to C.8. The seven worked examples printed in Appendix C. Every printed risk, the direct, indirect and total figure in all four loss categories, and the strike frequencies, are compared as the printed text.
Appendix C

The AS 1768:2021 worked examples Lumex™ reproduces

The headline figure for each example is its risk of loss of human life, R1, against the tolerable value of 10⁻⁵ per year. Lumex™ matches every printed figure in each example, not only this one.

Worked exampleRisk of loss of human life, as printedDirect flashes per year, as printed
Figure C.2, office block, unprotected9.10E-050.166
Figure C.3, office block, protected6.24E-060.166
Figure C.4, hotel in Darwin5.92E-060.062
Figure C.5, historic church6.23E-060.542
Figure C.6, remote pump station1.49E-090.001
Figure C.7, two-storey house8.94E-060.007
Figure C.8, mountaintop broadcasting site3.01E-063.009

One value differs between the standard's table and its tool. AS 1768:2021 Table 2.1 prints 10⁻⁴ as the tolerable risk for loss of cultural heritage, while the tool and every Appendix C example apply 10⁻³. Lumex™ follows the tool, so its results match what an assessor sees in the official spreadsheet, and the choice is recorded against the value.

NFPA 780-2026

An independent implementation, until a worked example exists

NFPA 780-2026 Annex L prints no full worked example of the 2026 method. The only worked NFPA calculation Lumex™ holds was prepared to the 2011 edition, and that method differs from the 2026 annex in almost every quantity: the 2026 annex works from the ground strike-point density rather than the flash density, and computes the collection areas of the lines differently. It would be wrong to tune the 2026 engine to match it, so it is kept as a record of why it is not a test.

Instead, the NFPA engine is compared with nine cases computed by a separate implementation of Annex L, written from the register of Annex L values rather than from the engine. Each case goes in through the same validation an assessment uses. The two implementations multiply in a different order, so the comparison allows a relative difference of 10⁻¹² and nothing more. For the method itself, read NFPA 780 Annex L, section by section.

The limits

What verification does not cover

A passing check proves the engine applies the method as printed. It does not prove the inputs are right, and it does not reach parts of a standard that print no worked figures.

The inputs are the engineer's

The structure, its surroundings, its lines, the losses and the protection measures come from the person running the assessment. Every report shows them, so a reviewer can check them as well as the arithmetic.

Paths the examples do not reach

Annex F does not exercise every option in IEC 62305-2:2024, such as each national departure. Those paths are covered by tests that cite the clause or table each value comes from, rather than by a printed example.

The design is not certified

Lumex™ runs the risk assessment. It does not design, install or certify a lightning protection system, and a passing verdict does not replace the engineer who signs the report.

For how the edition a figure was computed on becomes part of the answer, read the edition is part of the answer; for what a reviewer checks before a report is signed, read the engineer signs, a reviewer checks.

FAQs

Questions answered

Does Lumex™ reproduce the IEC 62305-2 Annex F case studies?

Yes. Annex F of IEC 62305-2:2024 works three structures through the method: a house (F.2), an office building (F.3) and a hospital (F.4). Lumex™'s test suite runs each study, unprotected and protected, and compares the engine with the collection areas, dangerous events, risk components, zone risks and frequencies of damage the annex prints.

How close do Lumex™'s figures need to be to the published ones?

Within 1.5 percent of each printed figure, or within half of the last printed digit where Annex F prints too few digits for a percentage to mean anything. Annex F prints its values to three significant figures and warns in F.1 that rounding the intermediate values moves the last digit, so an exact match on every printed digit is not possible.

Does Lumex™ match the AS 1768 risk assessment tool?

Yes, exactly. The AS 1768:2021 engine is compared with 400 input combinations computed by the Lightning Risk Assessment Tool v5.0 itself, with the case saved in the tool's own workbook, and with every risk printed in the standard's worked examples, Figures C.2 to C.8. The comparison allows no tolerance: the figures must agree as printed.

Is the NFPA 780 engine verified against a published worked example?

Not yet. NFPA 780-2026 Annex L prints no full worked example for the 2026 method, and the only worked NFPA example Lumex™ holds follows the 2011 edition, whose method differs in almost every quantity. The NFPA engine is instead compared with nine cases from an independent implementation of Annex L, written separately from the engine.

Does verification mean Lumex™ replaces engineering judgement?

No. Verification shows the engine applies each standard's method as printed. The inputs still come from the engineer: the structure, its surroundings, its lines, the losses and the protection measures. A correct calculation on wrong inputs gives a wrong answer, which is why every Lumex™ report shows its inputs and the clause behind each figure.
What Lumex™ does, and what stays with you

Lumex™ computes the method of the standard you choose, IEC 62305-2:2024, AS 1768:2021 or NFPA 780-2026, and shows the working. It does not certify a structure. You may not issue or submit a Lumex™ output until a competent person, qualified where the structure is located, has reviewed the inputs and the result and signed it.

The tolerable risk in IEC 62305-2 is not a fixed constant. Clause 7.3 NOTE 1 gives RT = 1 × 10⁻⁵ per year as a representative value of tolerable risk and adds that another value may be set once the case has been investigated in detail. Printed p.12 then lets national or local regulations fix RT, the tolerable frequency of damage FT, and the Annex A, B, C and E calculation rules and parameter values. Every IEC 62305-2 assessment in Lumex™ states the jurisdiction it was computed under and the values that applied.

Get started today

Risk figures you can check against the standard

Get Started Book a Demo