SAG-mill feed chute blockage: the power ramp before the alarm

A SAG mill does not usually announce a blocked feed chute with a bang. It announces it with a slow, monotonic rise in power draw that looks, for the first several minutes, like competent control responding to harder ore. By the time a conventional overload alarm trips, the mill is already carrying a charge it cannot circulate, the chute is packed, and the operator has lost the lead time needed to intervene before mechanical stress escalates.

This article dismantles that one failure mode — feed chute blockage on a SAG mill — and frames the detection problem in IEC 61511 terms: diagnostic coverage, proof-test interval, and the difference between a process alarm and a safety instrumented function.

The mechanism: how a chute blocks and how power responds

A SAG mill feed chute delivers ore and grinding media from the feed conveyor into the mill inlet. Blockage typically begins as a partial restriction: wet, sticky ore, an oversize rock bridging the chute throat, or a foreign object such as a broken liner bolt or a section of conveyor belt. The restriction reduces the mass flow rate into the mill while the control system, seeing a low mill load, increases feed rate to compensate.

The mill responds in a predictable sequence. As the charge level rises without adequate throughput, the mill draws more power to lift and tumble the increased load. Mill power climbs — often gradually, sometimes in steps as the charge redistributes. Bearing pressure on the trunnion rises. The recycle or pebble circuit loads up. Discharge flow becomes erratic. Only after the charge reaches a point where the mill can no longer circulate it does power spike sharply, and only then does a conventional high-power or high-bearing-pressure alarm annunciate.

The critical window is the ramp. During that period, the process is drifting outside its safe operating envelope while every indication available to a standard control system remains within the range that control logic considers normal. A high-power alarm set at, say, 110% of rated power provides no warning at 95%, 98%, or 103%. It is a threshold device, not a diagnostic.

Why conventional overload alarms miss the ramp

Three structural reasons:

**Alarm setpoints are static.** A high-power trip is configured against a fixed limit. It cannot distinguish a slow ramp from a step change, and it cannot distinguish a genuine overload from a transient caused by a hard ore slug. Operators learn to acknowledge and reset it, which erodes its value as a protective layer.

**The ramp is inside the control band.** Mill control loops are designed to vary feed rate and speed to maintain throughput. A rising power trend is exactly what the controller is supposed to produce when ore hardness increases. Without a trend diagnostic, the control system cannot tell the difference between a legitimate response and a developing blockage.

**Detection is downstream of the cause.** By the time power or bearing pressure crosses a threshold, the chute is already blocked. The alarm reports a consequence. It does not report the loss of feed function that caused it.

IEC 61511 framing: diagnostics, not devices

IEC 61511 addresses this class of problem through the safety instrumented function (SIF) and its associated diagnostic coverage. The relevant question is not “is there a high-power alarm?” but “what is the diagnostic coverage of the SIF that detects loss of feed?” A threshold alarm on mill power has, in practice, low diagnostic coverage against a slow chute blockage, because the dangerous condition develops well before the threshold is reached.

Two concepts matter here:

**Proof-test interval.** A proof test verifies that a SIF can still perform its function. If the only means of detecting a blocked chute is a high-power trip, the proof test must demonstrate that the trip still actuates at the correct setpoint. It does not demonstrate that the SIF detects the ramp. A proof test that passes while the diagnostic is blind to the dominant failure mode gives false confidence.

**Diagnostic coverage against the specific failure mode.** IEC 61511 requires that the diagnostic be evaluated against the failure modes it is credited with detecting. For feed chute blockage, the dominant failure mode is a slow, monotonic reduction in feed mass flow with a corresponding rise in mill power. A diagnostic that detects only the terminal spike covers a small fraction of the failure trajectory.

The practical implication is that a feed chute blockage SIF should be architected around a trend-based diagnostic — rate of change in mill power correlated with feed rate, bearing pressure, and discharge behaviour — not around a single threshold. The diagnostic must be proof-tested against a simulated ramp, not just a simulated step.

What this means for instrumentation and proof-testing

For mining reliability and OT engineers, the actionable points are:

• **Instrument the ramp, not just the limit.** Mill power, bearing pressure, feed conveyor mass flow, and recycle load should be trended and evaluated together. A single tag is insufficient.
• **Define the diagnostic in the SIF specification.** State the failure mode (slow feed restriction), the detection principle (correlated trend), and the diagnostic coverage claimed. Do not credit a threshold alarm with coverage it cannot deliver.
• **Proof-test the ramp.** A proof test that injects a step change validates the trip. A proof test that injects a realistic ramp — over the timescale of an actual blockage — validates the diagnostic.
• **Separate process control from the SIF.** The control loop that increases feed rate in response to low load must be independent of the diagnostic that detects the resulting blockage. Common-cause failure between control and protection is a recognised hazard.
• **Document the demand rate.** If blockages occur frequently, the SIF is operating in a high-demand mode, and the architecture and proof-test interval must reflect that.

Where the capital argument sits

The commercial context matters because detection upgrades compete for capital. As MINING.COM reported on 12 October 2026, geopolitical risk is driving capital toward private-sector investment, with analysts noting a shift in where mining capital is deployed (source: live-verified). For reliability engineers, that shift is an argument for framing feed chute diagnostics as risk reduction with a defensible IEC 61511 basis, not as a discretionary instrumentation upgrade. A SIF that detects the ramp before the overload alarm reduces the demand on mechanical protection and shortens the window in which an operator must respond.

Getting started

Audit your existing SAG mill feed chute protection against the failure mode described here. List every tag currently credited with detecting a blocked chute. For each, ask whether it detects the slow ramp or only the terminal spike. Where the answer is the spike, the diagnostic coverage is low, and the proof test is validating a trip that fires too late to protect the mill. Re-specify the SIF around a trend-based diagnostic, define the proof test to inject a realistic ramp, and document the demand rate. That is the work that closes the gap between a power ramp and an alarm.

Grounded in a verified live source: Geopolitical risk drives capital to private sector: analyst (2026-10-12) — source: live-verified.

Talk to S5 System about your site

S5 System designs and manufactures ultra-low-power industrial IoT sensors for mining, agriculture, infrastructure and utilities.

Discuss your IoT monitoring project — book a call.