Vibration Monitoring of Rolling Mill Auxiliary Drives | Bently Nevada — KEG TRK

The mill's main drive is the top-priority asset: it carries eddy-current probes, a 3500 protection system, and historian logging in System 1. But it's actually the auxiliary units — roller tables, shears, coilers, cooling and lubrication pump stations — that most often cause unplanned stops when monitoring is limited to visual rounds and occasional spot checks.

An auxiliary drive in a metallurgical complex works under harsh conditions: shock loads from sudden billet braking, vibration from neighboring stands, high temperature, oil mist, and metal dust. A bearing fault or a loosening foundation bolt develops faster here than on a "clean" pump or fan. The reliability engineer's job is to apply the same level of engineering control to these units as to a turbogenerator — while respecting budget and accessibility constraints.

Why rolling mill auxiliary drives are critical

A rolling mill is a synchronized line. Stopping one roller table ahead of the shears means the entire strand stops: the billet cools, the schedule slips, and the crew shifts into emergency mode. Based on experience at Kazakhstan and CIS plants, the cost of unplanned downtime in a metallurgical process line runs into tens or hundreds of thousands of dollars per day, depending on mill capacity and market conditions.

Auxiliary drives differ from the main drive in several ways:

• A diverse fleet — from 15–250 kW variable-speed motors to hydraulic shear drives.
• Uneven loading — starts with a full billet, reversals, brief overloads.
• Limited access — a sensor can't be mounted where red-hot metal passes.
• Varying criticality — a coiler ahead of shipping matters more than a secondary cooling water pump, but either can stop production.

This is why plants typically apply a two-tier strategy: critical units get continuous monitoring on a Bently Nevada 3500 rack with 3500/42M modules, while the rest are covered by compact Orbit DCM systems or route-based monitoring, with "problem" units later moved to permanent monitoring.

Typical units and monitoring points

Roller tables and billet conveyors

A roller table drive is a classic chain: electric motor, gearbox, chain or roll transmission. Common faults:

• gearbox and output shaft bearing wear;
• misalignment after a coupling change or gearbox repair;
• imbalance from scale buildup on the rolls;
• mechanical looseness in the frame supports.

Recommended points: horizontal and vertical vibration velocity on the gearbox bearing housing (Velomitor sensors), an accelerometer on the motor housing at the drive end. For roller tables running below 600 rpm, also monitor the low-frequency component — here the 3500/42M module, with configurable bands and alarms per ISO 10816, has an advantage over a "universal" threshold.

Shears and guillotine mechanisms

Shears generate a shock load on the drive with every cut. On an FFT spectrum this shows up as brief spikes and rising broadband noise. Continuous monitoring lets you separate the normal cutting impact from growing wear in the crank mechanism or a loosening blade mount.

On hydraulically driven shears, vibration is monitored on the pump unit and the pump drive motor — a typical scenario for an Orbit DCM with 4–8 channels.

Coilers and uncoilers

A coiler operates with a varying coil diameter: moment of inertia and drive load both change. Trend analysis in System 1 matters here — not a one-off threshold, but the rate of rise in 1x RPM and harmonics over weeks. A coiler support shaft bearing defect often shows up 4–8 weeks before failure, provided data is logged regularly.

Lubrication, cooling, and hydraulic pump stations

Pumps in a rolling mill's auxiliary circuits are high-speed machines well suited to vibration monitoring. Cavitation, impeller wear, misalignment after repair — all of this is visible in the FFT spectrum. An Orbit DCM is sufficient for lubricant and roll-cooling pumps; a module in the 3500 rack is warranted for the main drive's emergency cooling pumps.

Choosing a system: 3500/42M or Orbit DCM

Criterion	3500/42M in a 3500 rack	Orbit DCM
Number of channels	Scales with the rack	4–16 channels per cabinet
DCS integration	Relay outputs, Modbus, optional	Ethernet, relay modules
Data logging	System 1, trends, reports	Local + export to System 1
Cost per point	Higher, justified on critical units	Lower, for distributed drives
Trip protection	Yes, automatic shutdown	Yes, configurable alarms

A practical rule for a rolling mill shop: if stopping the unit halts the strand for more than 2 hours, install a 3500/42M. If the unit has redundancy or repairs can wait for a planned window, use an Orbit DCM with weekly trend reviews in System 1.

Implementation: from survey to alarms

1. Criticality — rank drives by their impact on downtime and safety.
2. Survey — route-based measurements, baseline spectra of "healthy" condition.
3. Design — sensor selection (Velomitor vs. accelerometer), cable routing, ATEX zones where needed.
4. Installation and commissioning — calibrate thresholds to API 670 / ISO 10816, not by eye.
5. Integration with maintenance — alarm → work order → fault confirmation → repair → trend closeout.

As the official Bently Nevada representative in Kazakhstan, KEG TRK handles the full cycle: from surveying auxiliary drives on an operating mill to commissioning the 3500 system and training shop personnel to read spectra.

Economics of the solution

Investment in monitoring 15–25 rolling mill auxiliary drives (a mixed 3500 + Orbit DCM configuration) typically pays back in 12–18 months if it prevents a single major unplanned strand stoppage. An added benefit is fewer unplanned gearbox repairs: replacing a bearing based on a trend is cheaper than replacing a gear set after a failure.

Frequency analysis: what to look for on auxiliary drives

A vibration engineer on a rolling mill needs to distinguish "normal heaviness" of a metallurgical drive from a developing fault. Basic signatures:

Imbalance — a dominant 1x RPM peak in the velocity spectrum, same phase on opposite bearings. On roller tables this is often linked to scale buildup; after roll cleaning, the System 1 baseline is updated.

Misalignment — 1x and 2x with different phase at the motor's DE and NDE; axial vibration above normal. A common cause is a coupling repair done without laser alignment.

Bearing defect — a series of BPFO/BPFI peaks in the envelope band; on a housing-mounted Velomitor this shows up as rising broadband noise in the 2–8 kHz range. More detail in bearing failure stages.

Gear defect — rising GMF (gear mesh frequency) and sidebands; see gearbox spectral analysis.

Mechanical looseness — multiple harmonics at 1x, 2x, 3x… with unstable amplitude; often follows vibration from feeding an oversized billet.

Sensor maintenance schedule in a metallurgical shop

Metal dust and scale are enemies of vibration sensors. Recommended schedule:

• weekly visual inspection of cable entries and connectors;
• monthly check of sensor mounting torque (without overtightening);
• quarterly cross-check against a reference SCOUT route measurement;
• after any drive repair — mandatory re-recording of the baseline.

Sensors near emulsion spill zones or water-cooled roller tables are fitted with protective shields; cable is routed in metal conduit from the measurement point to the Orbit DCM cable channel.

Sample configuration for a mid-size rolling mill

A typical KEG TRK project for a medium-capacity plate or section mill:

Unit	Quantity	Solution	Channels
Roller tables ahead of the stand	6	Orbit DCM	12
Shears	2	Orbit DCM	8
Coilers	3	3500/42M	12
Lubricant/cooling pumps	8	Orbit DCM	16
Main drive (already installed)	1	3500	—

Total: one 3500 rack with 42M modules on the coilers plus 2–3 Orbit DCM cabinets for the remaining groups. All trends converge in System 1 on the shop reliability engineer's workstation.

Frequently asked questions

Are eddy-current probes needed on auxiliary drives? For most 1500–3000 rpm auxiliary electric drives, a Velomitor and an accelerometer are sufficient. Proximity probes are justified on slow coiler shafts and units with axial impacts — see 3300 XL on large fans.

Can we start with route-based SCOUT surveys? Yes. The recommended path: 3 months of route measurements → criticality matrix → permanent monitoring on Class A assets. See route-based collection at a concentrator plant — the methodology also applies to metallurgy.

How does this connect to an existing CMMS? Orbit DCM and 3500 transmit alarm status via Modbus TCP or dry contacts; System 1 exports reports. Integration with 1C:TOIR or SAP PM is configured at the dispatch level.

Aligning with the preventive maintenance schedule

Vibration monitoring of auxiliary drives delivers maximum benefit when System 1 data drives the PM schedule, rather than sitting in an archive "for the record." A practical scheme for a rolling mill shop:

• a monthly report on the top 10 rising vibration trends, presented to the reliability team meeting;
• units flagged with a warning two months in a row get moved to "planned repair" instead of waiting for failure;
• after every planned mill stoppage — re-check baselines for all drives affected by the stop (vibration can shift due to line realignment).

At Kazakhstan plants, a typical effect is a reduction in unplanned auxiliary drive stoppages from 40% to 15% of total strand stoppages in the first year of the program, with disciplined alarm response.

Staff certification and knowledge transfer

A sustainable vibration monitoring program on a rolling mill requires at least one certified specialist per shift who can tell "shop noise" apart from a genuine Orbit DCM alarm. KEG TRK recommends a two-tier competency model: the shift operator confirms the alarm and notifies the mechanic; the vibration analyst performs spectral analysis and decides on repair. Without this separation, plants either ignore real alarms or stop the strand on false ones. Annual recertification with a review of 2–3 cases from the System 1 archive keeps the team's skills sharp despite staff turnover.

Summary

Rolling mill auxiliary drives are not "second class" compared to the main drive. They are, in fact, the units that most often break the production schedule. Bently Nevada systems — 3500/42M, Orbit DCM, Velomitor sensors, and the System 1 platform — give the reliability engineer the same tools used on a power-generation turbine unit, adapted to metallurgical realities.

To discuss a survey of your rolling mill's auxiliary drives, reach out via KEG TRK contacts.