What to do about Lightning
An established client for whom I had done a lot of work on the protection of explosive devices against electrostatic discharge and RF (radio frequency) interference asked for help in dealing with a change in site policy with regard to actions to be taken when lightning threatened the area.
A new lightning sensing device had been installed at a different part of the site and the new policy was to be that when the people there got a reading that looked 'dangerous' they were to phone the guard house who would phone every other operation on the site and shut them down.
When the reading dropped to 'safe' level, they were to phone the guard house again and then the guard house would phone every operation on the site and tell them they could start again.
It sounded reasonable except that lost production time went from typically an hour or less to half a day and sometimes the guardhouse crew didn't remember to phone everyone.
Needless to say, the production manager for the operation in question wasn't too happy about the new methods.

The lessons of the past
The site had been struck many times in the past, it was in a hilly region and included some high ground, but no one was hurt and being an explosives site, properly arranged, even though one magazine had blown up many years ago, there was no propagation to any other since they were all kept well apart.
Over the years a simple practice had worked well. Whenever there was a thunder storm in the area, a watcher was posted at the door who timed the arrival of thunderclaps and work was stopped and people moved to a nearby safe (no explosives inside) building whenever the strikes got closer than 2 miles (10 seconds).
The particular buildings in question were protected by lightning rods, they had totally enclosing metal walls which formed a highly conductive Faraday shield. It was part of the oral history that they had been struck by lightning in the past. The equipment had been, as normal, left powered but not running, electronic control systems included. When the people went back in, there was not a trace of any trouble; they just carried on as if it had been a normal lunch break.

Investigation
At the operation manager's request I investigated the new instrument which was measuring electrostatic field strength as an indicator of stress in the atmosphere. No one had done any calibration work and the operators didn't understand it. The settings were, according to the manufacturer's guidance, OK for a very restrictive policy but no one had looked at the overall operational consquences. The location was on the same site but about 2km away from the operation of interest.
We called in a firm of lightning arrester experts, commisioned them to measure the quality of the grounding at the buildings. It turned out that they had never done such a survey before so in effect I had to run the tests (a sort of Kelvin double bridge test for low resistances). All OK. The building had a big grid of conductors installed under the surface of the ground all around and the resistance measurements were well within recommended limits.
Looking at the theory of lightning strikes and the corresponding design of the protection provided for the buildings it was apparent that they were well protected, including the similar lunch room to which the people retired while waiting for the storm to pass.
It was also apparent that the field measured at the remote location wasn't a particularly good measure for the buildings in question which would generally see the storms first.

Result
So by combing an understanding of the instrumentation, the mechanisms of lightning strikes, a practical hands-on approach to testing and the geography and history of the site, we were able to develop a policy that wasn't far from the old procedure. By the time the site was closed a few years later there had been many thunderstorms but no incidents.

Where we are:
This sort of hands-on approach generally needs to be provided by someone based nearby. My base is in Northern Delaware, USA, but there are strong links to the UK where we have clients too.
Washington DC, Philadelphia PA, Trenton NJ, Baltimore MD, Arlington VA, Camden NJ, Wilmington DE and New York NY are all easy to get to. Using Philadelphia International Airport (PHL) gives access anywhere, though we find much of our computation and analysis work can be done remotely using the internet saving travel costs.

Feel free to contact me if you want to discuss this sort of problem. No charge for a preliminary confidential chat.