This article is based on the Whitepaper ‘Condition Monitoring of Rotating Electrical Machines 2025, by Ron Scollay, Engineer, Machinemonitor.

There is a particular kind of frustration familiar to engineers who manage rotating machines at scale. You know the asset is deteriorating. You can see it in the vibration signature, hear it in the harmonics, and read it in the current distortion. And yet the machine keeps running — until the day it doesn't.

Condition monitoring exists precisely to close that gap between what engineers know and what organisations act on. The question is not whether it works. The question is why so many operators are still not doing it systematically, and what it costs them when they don't.

The problem starts at installation

Many failure pathways in rotating machines are established before the equipment ever runs under load. Concentricity, magnetic centre alignment, foundation integrity, fastening torque and distortion. These are not secondary concerns. They are the baseline from which every subsequent degradation mechanism compounds.

An eccentricity introduced at installation doesn't stay static. It creates uneven air gap flux distribution, which produces current distortion, accelerating localised thermal stress in the winding, and shortening insulation life. The causal chain is well understood. The problem is that without baseline vibration and current data taken at commissioning, there is no reference point against which developing anomalies can be trended.

Condition monitoring begins not at the first sign of trouble, but at day one. Establishing correct installation, magnetic centre, shaft alignment, foundation flatness, soft-foot, and capturing a clean baseline is the foundation on which everything that follows depends.

Condition-Based vs. Time-Based: The maintenance argument

The traditional model, scheduled overhaul at fixed intervals, was designed for an era when real-time asset data was either unavailable or prohibitively expensive to collect. That era is over.

The argument for condition-based maintenance (CBM) rests on a straightforward observation: identical machines in identical service accumulate damage at different rates depending on loading history, thermal cycling, starting frequency, and alignment drift. A time-based schedule cannot account for these differences. It either over-maintains assets that have life remaining or, more dangerously, under-maintains assets that are degrading faster than the schedule assumes.

Vibration amplitude and motor current signature analysis (MCSA) are the primary tools for detecting mechanical and electrical anomalies in service. Bearing defect frequencies, rotor bar cracking, stator winding asymmetry, and misalignment each produce characteristic signatures that online monitoring can detect weeks or months before a failure event. Complementary parameters, temperature, partial discharge, and oil analysis extend that visibility further, particularly for machines where insulation condition or lubrication integrity are the critical failure drivers.

The technology is mature. The interpretation frameworks exist. What CBM requires is defined alarm thresholds, trending analysis to distinguish normal variation from genuine deterioration, trained personnel who can translate data into decisions, and integration with maintenance management systems so that action actually follows insight.

The total cost of ownership calculation

The commercial case for condition monitoring is often made in terms of avoided failures. It is better made in terms of production economics.

Consider a mineral processing operation with a main feed conveyor driven by a 220 kW, 415 V induction motor. Nameplate value is modest. But at a production output of $45,000 per hour, an unplanned outage requiring a four-hour motor changeout costs $180,000 in lost production alone, before factoring in emergency parts procurement, unplanned labour, and the downstream disruption that distinguishes an emergency response from a planned intervention.

This single event dwarfs the cost of condition monitoring for that asset across several years of operation. That is the total cost of ownership (TCO) calculation done honestly, one that accounts for energy efficiency losses from misaligned or deteriorating machines, shortened asset life from unmanaged fault progression, and the full cost of failure events, including production impact.

Equivalent Annual Cost (EAC) analysis provides a complementary tool for evaluating monitoring investments: the annualised cost of monitoring technology and labour set against the annualised expected cost of failure, including production loss, repair, and safety consequences. When that comparison is on the table, the investment case for systematic condition monitoring rarely requires further justification.

Process Control, VSDs, and the fatigue question

The shift toward variable speed drives (VSDs) in rotating machine applications has introduced a new set of condition monitoring considerations that are not always well understood at the maintenance planning level.

VSDs reduce mechanical stress from direct-on-line starting and allow precise control of the machine duty point, resulting in genuine efficiency gains. But they also introduce harmonic content into the supply, and in some configurations create shaft voltage and bearing current issues that accelerate bearing degradation in ways that traditional vibration trending may not immediately detect.

There is, however, a compensating advantage: modern drives provide accessible data, current, torque, speed, that can itself serve as a diagnostic source when correctly interpreted. VSDs, used well, enhance condition monitoring capability rather than simply complicating it.

The fatigue implications of frequent stopping and starting, thermal cycling, and torque transients need to be explicitly factored into maintenance strategy, particularly for machines operating at variable loads across a wide speed range. Condition monitoring in VSD-driven applications requires adaptation of both the measurement approach and the interpretation framework, not avoidance of monitoring, but more informed monitoring.

Training is not optional

Condition monitoring technology is only as effective as the people who deploy and interpret it. This point is consistently underweighted in the investment case.

Maintenance and reliability personnel require genuine competence in the failure modes of the machines they are monitoring, correct installation and calibration of measurement equipment, interpretation of vibration spectra and current signature data, and, critically, recognition of the limits of their own expertise and when specialist analysis is required.

An alarm without the contextual understanding to interpret it is not condition monitoring. It is noise. An organisation that invests in sensor infrastructure without investing in the human capability to act on what those sensors reveal has not implemented condition monitoring. It has installed expensive equipment that generates reports nobody acts on.

Investment in training also drives a broader cultural shift: from reactive and time-based thinking toward predictive, condition-driven decision-making. That shift compounds the benefits of the technology over time.

What good looks like

The organisations that manage rotating machine fleets most effectively share a few consistent characteristics.

They understand their reliability and safety objectives at the asset level, not just the fleet level, and they have prioritised maintenance strategies accordingly. The availability, maintainability, and reliability ratios that operations target (in power generation, commonly around 90:7:3) are only achievable when maintenance intervention is driven by asset condition, not calendar.

They align electrical machine maintenance with driven equipment overhaul windows, treating planned downtime on a ball mill reline or mechanical overhaul as the trigger for inspection of the driving motor, opportunistic maintenance that maximises the value of each planned outage rather than treating assets as independent.

They have done the TCO and EAC calculations properly, including production loss exposure in failure scenarios. And they have built the organisational competence needed to ensure condition monitoring data drives maintenance decisions.

The gap between organisations that do this well and those that don't is not technology. The technology is available to everyone. The gap is in the systematic commitment to applying it.

The engineering and commercial case for condition-based maintenance of rotating machines and HV assets is a central theme at the 24th Annual Machines and HV Assets Conference. Join Australia's most experienced HV and rotating machine engineers for two days of applied technical exchange.