Tax breaks on exploration spending could mask level drift in dewatering circuits

The 2026-10-12 MINING.COM analysis “Geopolitical risk drives capital to private sector: analyst” (source: live-verified) describes capital rotating toward private-sector exploration and development as geopolitical risk reshapes financing. More exploration capital means more drill programs, more temporary dewatering, and more sumps and pump stations commissioned quickly and handed to operations with limited instrumentation review. The failure mode worth naming here is not the tax treatment of exploration spending itself. It is what accelerated commissioning does to the level loop in a dewatering circuit: a slow-responding or slowly drifting level sensor that no high/low alarm will ever catch.

The mechanism: drift is not a step change

A level transmitter on a pump sump or thickener underflow reports a continuous variable. Its failure modes are not limited to dead, stuck, or grossly wrong. The dominant dangerous failure mode in dewatering service is slow drift: the indicated level diverges from true level at a rate of millimetres to centimetres per week. Causes are ordinary. A differential-pressure (DP) leg partially fills with solids. A guided-wave radar probe builds a dielectric coating from process scale. A capacitance probe’s reference shifts with moisture ingress. A float switch’s pivot stiffens with fines.

None of these produce a step change. They produce a monotonic offset that grows below the resolution of an operator watching a trend screen and below the trip point of a fixed high or low alarm. The loop still reads plausibly. The pump still runs. The sump still appears to be cycling within its normal band.

Why conventional alarms cannot see it

A high-high level switch is a discrete device. It asserts at one point. If the analog transmitter has drifted 400 mm low, the switch still asserts at its own setpoint, so the circuit appears protected. What has actually happened is that the analog loop’s indicated level is now wrong, and every control action derived from it — pump start, pump stop, duty/standby changeover, valve modulation on a thickener underflow control valve — is executing against a false measurement.

The consequence in a dewatering circuit is a sump that runs higher than indicated. Freeboard is consumed. In the worst case the pump loses suction, cavitates, and the circuit trips; in the better case the pump runs continuously against a closed or throttled discharge, and the mechanical seal and bearing housing absorb the consequence. On a tailings return line, sustained high level can back up a thickener and load the underflow pump beyond its design duty.

IEC 61511 does not treat this as an alarm-tuning problem. It treats it as a diagnostic coverage problem. The standard’s safety lifecycle requires that the achieved risk reduction of a safety instrumented function be supported by proof testing and by diagnostic coverage of dangerous detected and dangerous undetected failures. A level transmitter whose drift is undetected between proof tests contributes to the dangerous undetected failure rate of the function. If the loop is credited in a safety instrumented function, that contribution must be quantified in the SIL verification calculation, not assumed away.

Diagnostic coverage in practice

Diagnostic coverage for an analog level loop is built from comparison, not from the primary reading. Practical measures, in rough order of cost:

• **Redundant or diverse measurement.** A second transmitter on a different technology — radar alongside DP, or DP alongside a capacitance probe — allows deviation comparison. A sustained deviation beyond a defined threshold is a detected failure. This is the highest-coverage option and the one most often omitted on exploration-era dewatering skids.
• **Range and rate-of-change plausibility checks.** A level that has not moved in 72 hours on a sump with continuous inflow is suspect. So is a level that moves faster than the physical fill rate allows. These are logic checks in the PLC or RTU, not sensor features.
• **Comparison against a physically coupled variable.** Pump run-hours versus volume moved, or sump level versus flow balance across the circuit, provides an inferential check. If the pump has moved 4,000 m³ this week and the sump level trend is flat, the level reading is wrong.
• **Proof testing on a defined interval.** A proof test that only confirms the transmitter reads something is not a proof test. It must compare the indicated value against a traceable reference — a calibrated gauge glass, a temporary reference transmitter, or a verified manual dip — across at least three points spanning the operating range. Record the as-found deviation. A drift trend across successive proof tests is the earliest reliable indicator that the loop is degrading.

What to do with the proof-test record

The as-found deviation from each proof test is the data that matters. A loop that reads within 10 mm at commissioning and 250 mm low eighteen months later has a drift rate. That rate, not the alarm setpoint, determines the proof-test interval needed to keep the dangerous undetected failure contribution within the SIL verification assumption. If the interval is set by a calendar default rather than by observed drift, the assumption is unverified.

This is also where the capital-cycle argument connects back to the MINING.COM piece. When exploration and development capital moves quickly, dewatering assets are often commissioned with a single level transmitter, no deviation comparison, and a proof-test interval inherited from a template. The tax treatment of the exploration spend is a financing question. The instrument loop it leaves behind is a functional-safety question, and it is answered by diagnostic coverage and proof-test evidence, not by the presence of a high-level switch.

Getting started

Pick one dewatering sump on site. Pull the last twelve months of level trend and pump run-hours. Check whether the level reading is consistent with the volume the pump has moved. If it is not, you have a candidate drift. Then check the proof-test record for that transmitter: does it record as-found deviation against a traceable reference, or does it record only that the loop responded? If it is the latter, the loop’s dangerous undetected failure rate is unquantified, and any safety instrumented function crediting it is resting on an assumption nobody has tested. Fix the proof-test method before adding instrumentation.

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.