{"id":738,"date":"2026-09-14T18:55:54","date_gmt":"2026-09-14T18:55:54","guid":{"rendered":"https:\/\/lichens.ai\/?post_type=realisation&p=738"},"modified":"2026-09-14T18:55:54","modified_gmt":"2026-09-14T18:55:54","slug":"a-ramp-is-excavated-once-the-simulator-as-many-times-as-you-need","status":"publish","type":"realisation","link":"https:\/\/lichens.ai\/en\/realisation\/a-ramp-is-excavated-once-the-simulator-as-many-times-as-you-need\/","title":{"rendered":"A ramp is excavated once. The simulator, as many times as you need."},"content":{"rendered":"\n

An operating underground mine had to decide how many passing bays to excavate in its access ramp, where to place them along the axis and what length to give them. The ramp is single lane: two vehicles travelling in opposite directions can only pass each other in a bay built for that purpose. The production plan called for a sharp rise in truck traffic over a few years. The mandate addressed the question through simulation, with a model that reproduces the geometry of the ramp and the site’s actual traffic rules, vehicle by vehicle, and makes it possible to compare layouts before excavating them. <\/p>\n\n

The ramp is excavated once. The layout chosen commits haulage capacity until the end of the production plan. <\/p><\/blockquote><\/figure>\n\n

Scope<\/td>Modelling, sweep of the possible layouts and recommendation<\/td><\/tr>
Deliverable<\/td>A configurable simulator and a layout recommendation, quantified in waiting time and construction cost<\/td><\/tr>
Stakeholders<\/td>Mine planning, engineering, haulage operations, management<\/td><\/tr>
Out of scope<\/td>Calibration against telemetry, fleet variability, intersections and alternative routes<\/td><\/tr>
Drift model<\/td>One vehicle per direction per segment<\/td><\/tr><\/tbody><\/table><\/figure>\n\n

The challenge: on a single lane, capacity comes down to the number of passing bays<\/h2>\n\n

On a single lane, capacity does not depend on how fast the vehicles travel but on the number of passing bays. When two vehicles meet in opposite directions on the same segment, one has to reverse or wait, and the time lost propagates to those behind. <\/p>\n\n

The relationship is not linear: traffic that quadruples does not multiply waiting time by four, because blockages become both more frequent and longer to clear. Intuition therefore becomes unreliable at the very moment the decision is taken, once, for a traffic regime that will only exist later. <\/p>\n\n

At the start of the mandate, the number and position of the passing bays were set by the mine plan, with no measure of their actual effect on traffic, and the planned bays were grouped on a single portion of the ramp..<\/p>\n\n

The approach: each vehicle decides with what it can see<\/h2>\n\n
    \n
  • Fixed time step:<\/strong> the model advances in regular steps and represents each vehicle individually, with its dimensions, its speed, its priority rank and its behaviour when vehicles meet.<\/li>\n\n\n\n
  • Site rules:<\/strong> the traffic procedure is applied as written, including the safety exceptions, which override everything else.<\/li>\n\n\n\n
  • Local decision:<\/strong> no vehicle consults the global state of the simulation. Each one decides from what it can see and from what it is told. This latency reproduces that of underground radio and makes the calculated waiting times conservative. <\/li>\n<\/ul>\n\n