Metals & Mining
When the roaster stops, the whole plant stops.
Smelters, steel mills, and mining operations run 24/7 on motor-driven equipment that often sits where mounted sensors are hard to install, maintain, or trust. Extreme heat. Dust. Underground shafts. When these assets fail undetected, the cost is measured in hours of lost production, not maintenance invoices.
Trusted by metals and mining operators
24/7 production. Zero tolerance for unplanned stops.
In metals processing, equipment is chained. One failure does not take down one machine. It takes down the process.
01 Trigger
One motor degrades
A mill drive, smelter pump, or roaster motor starts to fail. Vibration routes do not catch it. SCADA shows green.
02 Cascade
The line stops
Equipment is chained. Downstream assembly idles within hours. Production schedules collapse.
03 Emergency
Repair, not maintenance
Crews scramble. Spares get airfreighted. Energy continues to bleed on degraded assets.
$2M / hour
Industry estimate for an automotive steel mill with downstream assembly idle. Tier 1 copper mines run near $480K per hour.
Sensors do not survive
Heat, dust, vibration, and access constraints make mounted sensors a maintenance liability of their own.
Most assets run dark
Vibration routes cover the biggest machines. The rest is a dark fleet: critical, but condition-blind.
Energy waste hides
Up to 60% of opex. Degraded motors bleed energy for weeks before anyone sees it on the bill.
Monitor from the MCC. Not the process floor.
No sensor survives 1,000°C at the strip mill face. But the motor current at the MCC carries the signature of mechanical degradation in the driven equipment. SAM4 reads that signal. No sensors on the asset. Brief MCC outage to install.
No sensors in the harsh environment
SAM4 installs at the motor control cabinet, away from heat, dust, and vibration. Current transformers clip around the phase wires. No equipment on the asset, no cabling to the process floor, no sensor replacement cycle.
Extend monitoring across more of the motor-driven fleet
One technology covers pumps, fans, conveyors, gearboxes, crushers, and agitators across many AC-motor-driven configurations. Fit depends on motor type, drive topology, duty cycle, and the failure modes you need to detect. Monitor more of the fleet from the MCC room, not individual assets.
Brief MCC outage per install
Installation usually takes under 60 minutes per asset, including a brief de-energisation. No process-floor sensor installation. Site-specific electrical permits and cabinet procedures still apply. At Nyrstar Budel, the full fleet was instrumented without a plant-wide production stop.
Detects electrical and mechanical faults
ESA reads bearing degradation, mechanical imbalance, belt wear, gearbox faults, motor winding issues, and grid-level electrical faults. Vibration analysis misses electrical faults entirely. At Nyrstar, SAM4 detected a shaved cable on a 10kV rectifier transformer that would have caused over €1 million in equipment damage.
“Thanks to SAM4, we can monitor assets that would otherwise remain unreachable.”
Two operating environments. One way to monitor through them.
Metals and mining push assets to the edge of what mounted sensors can survive. The MCC sits in a benign environment, and ESA reads fault-relevant evidence from there.
Metals
Process floor, heat, dustStrip mill drives
Continuous operation under 1,000°C ambient. Mounted sensors degrade in months.
Climate-controlled cabinet. No replacement cycle in the heat. Same fault library as any motor.
Blast furnace cooling and slag pumps
Bearing housings buried behind chemical injection lines. Hard to inspect, harder to instrument.
Motor, pump, and bearing-frequency content captured at the supply. Process containment stays intact.
Roasters, mill drives, process motors
Continuous duty in dust. Mounted sensors degrade silently.
The signal travels the supply cable that is already there. No new wiring across the dust floor.
Mining
Underground or remote, depth, distanceDewatering pumps
Submerged at the lowest point of the operation. No diver, no confined-space entry.
The pump stays in the sump. Surface electrics carry the signal up.
Underground ventilation fans
Remote shaft headings. Below ground, away from network coverage.
Data comes up the supply cable, not down a comms run. SAM4 stays at surface.
Long conveyors and primary crushers
Drive heads spread across kilometres of pit-to-port. Crusher motors in dust.
One installation per drive head. Coverage scales with cabinet count, not conveyor length.
The pattern is the same in both: the asset is hard to reach, the cabinet is easier to access, and SAM4 reads the electrical evidence from there.
The equipment that keeps your plant running
SAM4 monitors any AC motor-driven rotating equipment from the motor control cabinet. In metals and mining, these are the asset types where unplanned failure costs most. See all monitored assets
Pumps
Centrifugal pumps in water, chemicals, oil & gas, and process industries.
Fans & blowers
Ventilation fans, cooling fans, process blowers, and aerators.
Conveyors
Belt conveyors, screw conveyors, and chain drives.
Transmissions
Belt drives and coupling systems between motor and driven asset.
MV motors
Medium and high voltage motors in critical processes.
Not every asset is a SAM4 asset.
Strongest fit
AC motor-driven asset, signal accessible at the MCC or VFD, failure modes that produce measurable changes in current, voltage, torque, speed, or load.
Needs engineering review
Non-standard motor configurations, complex drive topologies, unstable loads, or unusual duty cycles. We flag these in asset-fit review.
Some assets are better covered by vibration on already-instrumented critical equipment, process instrumentation, OEM monitoring, or inspection. Asset-fit review tells you which is which before you commit.
Named customers. Quantified outcomes.
Loose cardan joint caught the same day SAM4 was installed
Read how SAM4 caught a loose cardan shaft coupling within hours of installation on this long-term steel customer's runout table roll.
Early warning on failing cardan shaft coupling in runout table roller
Read how SAM4 caught a failing cardan shaft coupling in a runout table roll for this steel manufacturer.
Bearing failure avoided in motor driving critical runout table roll
Read how SAM4 caught motor bearing failure in a runout table roll 7 months in advance for this steel manufacturer.
9x common faults detected in steel manufacturing
In this case study, we explain how SAM4 detects common faults in steel mills using actual results from anonymized SAM4 data, to help engineers evaluate SAM4’s
Condition monitoring for shot blasting machines: a case study
Read how SAM4 caught a loose blast wheel belt guard before it caused damage to this steel manufacturer's shot blasting machine.
Early warning of failing heated godet roll prevents $90k in production loss
SAM4 Health detects motor overloading in a heated godet roll, enabling timely maintenance. This prevents downtime, saving $70,000 in lost production and
16x early alerts to degrading rollers save €650k in lost production
By consistently detecting developing faults in its roller conveyors well ahead of failure, this hot strip mill is keeping production on track while spending
How ArcelorMittal prevented 31 hours of downtime by detecting 27 failures ahead of time
How ArcelorMittal prevented 31 hours of downtime by detecting 27 failures ahead of time
How zinc smelter Nyrstar got 800% ROI in 11 months
Nyrstar is one of Europe’s leading zinc smelters, refining over 300,000 tons of high-purity zinc every year. Operating continuously, the plant relies on
From email alert to planned repair
Detection is only useful if it changes what your team does next. Here is how reliability teams at smelters and steel mills use SAM4 to stop firefighting.
Alert
Email lands with fault classification
Asset name, fault type, severity, recommended action. The reliability engineer sees what is wrong before walking to the floor.
Validate
Engineer reviews and decides
The engineer checks SAM4 trend data and decides whether to act now or schedule for the next planned shutdown. No more opening up equipment to confirm a hunch.
Plan
Planned repair, not emergency callout
Parts are ordered. The crew is scheduled. The repair happens during a planned window. No weekend callouts. No overtime premiums. No cascade.
Improve
Feedback loop sharpens detection
Maintenance outcome gets recorded and fed back to Samotics. The model trains on your assets, your operating conditions. Detection performance improves with every closed work order.
At Nyrstar, that shift from reactive to planned maintenance drove 800% ROI in 11 months.
“The invisible pain points become visible.”
Patrick Emmers
Nyrstar
No sensors on the process floor
Clip on at the MCC
Current transformers clip around the phase wires inside the motor control cabinet. Brief MCC outage; under 60 minutes per motor in typical configurations. Site-specific electrical permits and cabinet procedures still apply.
Connect via cellular or Ethernet
Data transmits over 4G/5G or local network. Cellular deployments can avoid firewall and VPN changes. Store-and-forward buffering handles connectivity gaps in underground or remote areas. SAM4 is monitoring-only and does not send control commands.
Monitor from anywhere
The SAM4 dashboard shows asset health, fault alerts, and energy performance across the fleet. Alerts route to email, your CMMS, or your control room. Works alongside existing vibration programs.
The questions metals engineers actually ask
So was Nyrstar's team. They received an alert, did not believe it, and opened the casing anyway. The fault was exactly what SAM4 predicted. That single event changed the culture. We recommend starting with a small fleet of critical assets. Let the system prove itself with real detections on your equipment, in your environment. Scepticism disappears when the maintenance team sees it work.
Yes. SAM4 is highly compatible with variable frequency drives. VFDs are increasingly common in metals and mining for energy savings and process control. SAM4 monitors VFD-driven motors from the MCC using the same clip-on installation. The analysis accounts for variable speed operation.
SAM4 hardware sits in the MCC room, not on the process floor. The motor control cabinet is climate-controlled or at least protected from the worst of the heat and dust. No sensor needs to survive 1,000°C or metallic dust exposure. The measurement happens at the electrical supply, not the mechanical asset.
No. SAM4 complements vibration, it does not replace it. On assets where vibration sensors are already installed and working, ESA adds electrical fault detection and energy data that vibration cannot see. On the rest of your fleet, the assets that are too hot, too remote, or too low-priority for vibration, SAM4 is the primary condition monitoring source. ArcelorMittal described it as a good alternative to vibration measurements for hard-to-reach assets.
None. SAM4 builds its own baseline from live electrical measurements starting at installation. Within weeks, it has enough data to detect anomalies. Nyrstar had their first detection within the initial deployment period. No historical data export or SCADA integration is required to begin.
Yes, if the motor is powered from a surface or accessible MCC. Underground ventilation fans and dewatering pumps are powered from motor control cabinets that are often at the surface or in accessible electrical rooms. SAM4 installs there. The asset can be 500 metres underground. As long as the motor current is measurable at the MCC, the asset is monitorable.
Nyrstar achieved 800% ROI in 11 months. At $480,000 per hour of unplanned downtime in a Tier 1 copper mine, a single avoided production stop can pay for a multi-year deployment. The payback depends on your downtime cost, fleet size, and current failure rate. We provide a plant assessment that models the expected return for your specific operation.
SAM4 is monitoring-only. It reads electrical signals passively. It has no connection to control systems and cannot send commands to any asset. Data transmits over cellular or a dedicated network, completely separate from the operational network. No IT/OT convergence is required.
Stop firefighting. Start predicting.
Start with high-consequence conveyors, fans, pumps, roasters, mills, and auxiliary drives where access, heat, dust, or distance limits current monitoring coverage.