Skip to content
Risks in open-pit mining
Risks › Lightning › The eight questions

The eight questions an operations manager should ask

They separate what a supplier can demonstrate from what it can only claim. They paste straight into a statement of requirements.

From the report Standard and performance Version 1.1

1Does it see every lightning flash that threatens my operation?

Ask for the POD (probability of detection) measured on data from your own site, with the standard’s method. Reference points: networks exceed 90%; field mills fall between 20 and 60% (Sections 4 and 6).

2Does it know where each flash struck — and how does it achieve that precision here, where there is no national network?

Ask for the median location accuracy and what sustains it. In the Andes, accuracy of hundreds of meters is only achieved by adding local sensors to a global network — ask how many local sensors the offer includes and how they fuse with the global network (Sections 5.3 and 7).

3How much warning do I get before the storm arrives?

Ask for the POD with minimum lead time (for example POD₁₀: warnings with 10 minutes or more). A network that sees the storm at 100 km gives tens of minutes; an isolated local sensor, minutes (Sections 4 and 6).

4How often will it shut me down for nothing?

Ask for the false alarm ratio (FAR) and how it was measured — the standard’s definition is lax and two offers can declare incomparable figures. Every needless shutdown costs production and, worse, the alarm’s credibility (Sections 4, 6 and 8).

5What about the storm that forms right over the mine?

Ask how the system covers it: intra-cloud detection (the first discharges of nearly every storm) and meteorological precursors. If you are offered a field mill, three questions suffice: who manufactures it and will support it in five years?, how does it integrate — or not — with the network’s processor?, and can its signal, by itself, shut down the mine? (Sections 5 and 8).

6Does it keep protecting when a sensor fails?

Ask for the efficiency with one sensor out of service. With time-and-angle sensors the network keeps 94.5%; with time only it falls to 60.3% (Figure 3). A sensor failure is routine, not an exception.

7When can I safely get back to work?

Ask how the end of the alarm and the total duration per episode are defined (TTC, time to the all-clear, and TAD, total alarm duration, in the standard): that is where the shutdown hours live — and the risk of resuming too early (Sections 4 and 6).

8Who has verified all this, besides the one selling it to me?

Ask for validations by universities or meteorological institutes with real data. What operators declare systematically exceeds what third parties measure, and the IEC does not certify products: without a third party, everything is self-declaration. And if the offer invokes a “class,” it is citing a cancelled standard (Sections 3, 4 and 7).

The basis for each one is in the report, with its sources.

Not a purchase recommendation, nor an engineering specification for a particular site.
Version 1.1 ·