This article is part of the Ventilation Automation Pyramid series. Read the full overview here.
Level 3 gave the ventilation system reflexes. A gas exceedance triggers an airflow response. A door opens out of sequence and the system reacts. An equipment fault fires and the appropriate zone adjusts automatically.
The problem is that reflexes are still reactive. The system waits for a threshold to be crossed before it does anything. It has no awareness of what is about to happen underground, only of what already has.
Level 4 changes the question the ventilation system is asking. Instead of what just happened, it asks what is happening right now, and what will that require.
From Signals to Presence
The key input that Level 4 introduces is location.
At Levels 1 through 3, the ventilation system has no reliable knowledge of where equipment is in the mine. It knows what the schedule says: a blast is expected at 17h00, a shift ends at 18h00. It knows what sensors are reporting: CO is elevated in heading 730-4, a door on 650 Level is open. But it does not know that a scoop just entered a heading that has been unventilated for the past two hours, or that two diesel trucks are converging on the same ramp from opposite directions.
Level 4 plugs in a real-time location system (typically an RTLS using WiFi tags, BLE beacons, or ToF technology) and begins using equipment location as a primary input to ventilation logic. The system now knows where the diesel load is. Not where it is supposed to be, and not because a gas sensor already crossed a threshold. It knows because the equipment is being tracked in real time.
What Ventilation on Demand Actually Means
The term gets used loosely in vendor pitches, so it is worth being precise about what it means at this level.
Ventilation on demand means the airflow delivered to a zone is determined by the actual ventilation requirement of that zone at any given moment: not by a fixed setpoint, not by a schedule, and not by a threshold exceedance after the fact. At Level 4, that requirement is calculated primarily from equipment presence and type. Each piece of diesel equipment has a known emissions profile, and the minimum airflow quantities required per unit of engine power are documented in provincial and federal mining regulations, so the demand model is not arbitrary. It is grounded in the same numbers regulators use. When the system knows which equipment is in which zone, it can calculate the required airflow for that zone and adjust fan speeds and damper positions accordingly, continuously, as equipment moves through the mine.
The key difference from Level 3: the system responds before air quality degrades. The truck enters the heading. The fans ramp up. The heading receives adequate airflow for the diesel load it is actually carrying, not as a corrective response to a CO reading that already spiked, but as an anticipatory one to a known load entering a defined zone.
A Concrete Scenario: The Converging Equipment Problem
Consider a ramp section where two ventilation zones meet. Zone A is running at reduced flow, the last piece of equipment having left 40 minutes ago and the system dialled back accordingly. Zone B is at normal operating flow.
A haul truck exits Zone B heading toward Zone A. Simultaneously, a drill is making its way into Zone A from the other direction.
Without Level 4, neither piece of equipment has crossed into Zone A yet. No gas thresholds have been crossed. The schedule does not flag anything unusual. The ventilation system continues running Zone A at reduced flow. By the time both machines are in the zone and CO begins climbing, the system is already behind. A Level 3 event will eventually trigger, but the response is corrective rather than preparatory.
With Level 4, the RTLS shows both machines moving toward Zone A. The system calculates the combined diesel load of a haul truck plus a drill, determines the airflow requirement for that combined load, and begins ramping Zone A to the appropriate setpoint before either machine arrives. When the equipment enters the zone, the ventilation is already where it needs to be.
This is the operational difference between reacting to air quality and managing it.
The Telemetry Question: Knowing More Than Just Where
Location alone gets you most of the way there, but there is a subtlety worth unpacking. In the absence of additional telemetry, a Level 4 system has to assume worst-case diesel load for any tracked piece of equipment. If a 65-tonne haul truck is in a zone, the system ventilates for a running 65-tonne haul truck, even if that truck is parked with its engine off or running in electric mode on a battery-electric variant.
That conservative default is correct from a safety standpoint. But it leaves energy savings on the table.
On one project, we were deploying a new generation of LTE-enabled beacons for underground tracking. The original plan was to mount them as fixed infrastructure, but the customer had a better idea: they put them directly on the vehicles. That single decision changed what the beacon could do. It was no longer just a location reference point. It became a mobile asset tag that moved with the machine, doubled as a repeater for the evacuation alarm signal, and because these units had I/O ports available, could be wired directly into the vehicle’s ignition circuit.
That last part mattered more than it might sound. With ignition state available as a live signal, the VoD system could distinguish between a truck sitting idle in a zone and a truck actively running diesel. On hybrid or dual-mode equipment, it could also detect whether the machine was operating in diesel or electric mode. The ventilation setpoint could be calculated on actual load rather than assumed worst-case load. In areas where LTE infrastructure was deployed, which typically extends further into the mine than the fixed electrical network, this gave meaningful, measurable reductions in fan runtime and energy consumption. The savings were not theoretical. They showed up in power draw.
The broader point is that Level 4 is not just about connecting an RTLS to a ventilation controller. It is about how much you know about each asset in real time. More telemetry channels mean fewer conservative assumptions, and fewer conservative assumptions mean a system that can safely do less work.
What the Infrastructure Actually Requires
Getting Level 4 to work reliably underground depends on a few things that are easy to underestimate during planning.
The RTLS coverage needs to be genuinely continuous across active mining areas, not just adequate at the headings. A tag that goes offline for 20 minutes in a poorly covered section is not a minor inconvenience at Level 4. It is a gap in the ventilation control model. The system either defaults back to conservative assumptions for that zone or, worse, has no basis for any calculation at all. Dead zones in the tracking network translate directly into either safety risk or energy inefficiency, depending on which direction you choose to fail.
The mine also needs a well-defined ventilation zone model that the control system can work with in real time. This sounds straightforward but often is not. Coordinate systems, level designations, and heading names used by operations do not always map cleanly into the data structures that SCADA or VoD platforms expect. Getting that translation right is usually more of a project than it looks on the scope-of-work.
Fan and damper responsiveness becomes a real constraint at this level too. Level 4 demands more frequent setpoint changes than anything below it. Fans controlled by VFDs and dampers with reliable actuators are a practical requirement, not just for speed but because continuous adjustment will expose any mechanical reliability issues in the ventilation network that lower levels were never stressing.
Finally, and most often the source of actual project pain: the integration between the RTLS and the ventilation control system. These are usually different platforms, sometimes from different vendors, with different data models, different update rates, and different failure modes. The interface between them needs a defined fallback behaviour for when the RTLS feed is interrupted. A ventilation system that freezes at its last-known state when location data drops is a system waiting to cause a problem.
What Level 4 Does Not Do
Level 4 is still fundamentally a rule-based system. The rules are more sophisticated than Level 3, incorporating location, equipment type, zone topology, ignition state, and operating mode, but they are still rules written by engineers and encoded in the control system. A 65-tonne haul truck running diesel in Zone A demands X CFM. Add a second machine and the calculation updates. The system executes the rule.
What it does not do is notice that Zone A has been consistently underperforming its modeled airflow for three weeks because a damper upstream is not seating properly. It does not detect that fan efficiency is degrading as a unit ages. It does not recognize that the fleet’s actual emissions profile has shifted because a contractor brought older equipment on site than the model assumed.
Those observations require a system that is not just consuming real-time data but reasoning about patterns across time. That is Level 5, and it sits directly on top of what Level 4 builds.
Previous in the series: Level 3 – Event-Triggered Ventilation Automation.
Next in the series: Level 5 – Predictive Control and Network-Wide Optimization.
Working out where your own operation sits on this pyramid, or what the next level would actually take? That’s the kind of evaluation I do. Get in touch. No pitch.
Leave a Reply