2026-07-05 · Field note
What re-entry actually costs an underground mine
Every development heading in a drill-and-blast mine runs the same loop: drill, charge, blast, wait, inspect, re-enter, muck, repeat. The waiting part looks free on a Gantt chart. It is not.
The anatomy of the wait
After the blast, three things have to happen before the heading produces again. Ventilation has to clear the blast fumes to safe levels. Someone has to verify that it actually has, at the face, not at the fan. And someone has to look at the ground: misfires, rockfall, damaged support, water.
In most operations, steps two and three are a person walking in. That person is performing the single most dangerous routine entry in the mine, into ground that was explosively rearranged less than an hour ago, and the schedule protects them the only way it can: by adding margin. Conservative re-entry windows are not inefficiency, they are the correct response to inspecting with human bodies.
Price it for your own operation
We will not invent your numbers, but the arithmetic is short. Take one heading:
- How many blasts per day does it take?
- How many minutes does your re-entry window hold beyond ventilation clearance, because verification needs a human on foot?
- What does a producing heading earn per hour in your plan?
Multiply the three, then multiply by your number of active headings and your working days. That figure, which your own planning department can produce this week, is the annual cost of verifying re-entry with people. It does not include the exposure hours themselves, which your safety system already counts and would rather not.
What changes when the verifier is a robot
A robot that enters at ventilation clearance, reads gas at the face, images the ground, and reports before a human reaches the muster point does not make re-entry instant. It makes the margin unnecessary, and it moves the human entry from "first eyes in" to "informed and documented."
That is Sarthi's M1 mission, and it is the reason we describe our product in minutes and exposure hours instead of robot specifications. When we run a discovery conversation with an operations team, this calculation is the first thing on the whiteboard, with their numbers, not ours.