Introduction: Why early fault detection matters
Online condition monitoring of induction motors aims to prevent unplanned outages and reduce maintenance expenditure. In most industrial settings, motor failures do not occur instantaneously; they typically originate from a minor defect that propagates over time until it becomes a functional failure. This progression is commonly described by the –
curve (Potential Failure to Functional Failure), which highlights a time window in which developing faults can be detected through observable changes in operational signals such as stator current, vibration, temperature, or lubricant condition.
Detecting faults early within this window reduces secondary damage, enables planned interventions, and lowers repair costs. This guide outlines practical online diagnostic methods for common induction motor faults, focusing on identifying characteristic signatures in measured signals and translating them into actionable maintenance decisions.
1. Fundamentals of condition monitoring and modern maintenance strategies
Effective online monitoring should be embedded within a clear maintenance philosophy. Industrial maintenance has evolved from reactive approaches toward data-driven strategies that improve reliability and asset utilization. The most widely used strategies include:
- Corrective maintenance (run-to-failure)
Maintenance actions occur only after a breakdown. While simple to implement, this approach often causes unplanned downtime, higher emergency repair costs, and elevated safety and collateral-damage risk. It is typically appropriate only for low-criticality assets. - Time-based maintenance (TBM)
Maintenance is scheduled at fixed intervals (e.g., everymonths or after
operating hours). TBM improves planning but does not reflect the actual health of the asset. It may lead to premature replacement of healthy components or early-life failures due to installation errors or defective parts.
- Condition-based maintenance (CBM)
Maintenance is triggered by measured condition indicators. Sensors continuously or periodically track variables such as vibration, temperature, and current signatures. CBM addresses the question: “What is happening to the machine right now?” and supports planned interventions based on evidence rather than averages. - Predictive maintenance (PdM)
PdM extends CBM by estimating the remaining useful life and forecasting the likely time to failure using historical and real-time data. This supports just-in-time interventions, optimized spare-part inventory, and reduced operational risk.
The diagnostic methods described below form the technical foundation for CBM and PdM programs.
2. Online detection of common faults via signal-based analysis
Electrical and mechanical faults in induction motors leave identifiable patterns—often called signatures—in operational signals, particularly stator current and vibration. Frequency-domain analysis is widely used because many faults generate deterministic spectral components whose amplitudes increase as the defect progresses. Trending these components over time is often more informative than a single measurement snapshot.
2.1 Broken rotor bars
Cracks or breaks in squirrel-cage rotor bars are commonly driven by thermal cycling from frequent starts, sustained overloads, load fluctuations, or manufacturing imperfections. The fault redistributes rotor currents into adjacent bars, increasing localized heating and accelerating progressive rotor damage.
Primary detection method: Motor Current Signature Analysis (MCSA).
Typical signature: characteristic sidebands around the supply frequency (e.g., around ) in the stator current spectrum.
2.2 Air-gap eccentricity
Air-gap eccentricity occurs when the rotor-to-stator radial clearance becomes non-uniform. It may be static (the minimum air gap remains fixed) or dynamic (the minimum air-gap location rotates with the rotor). Eccentricity produces unbalanced magnetic pull, torque ripple, elevated vibration, and increased bearing stress, and in severe cases may lead to rotor–stator contact.
Primary detection method: spectral analysis of stator current (often complemented by vibration).
Typical signature: characteristic frequency components whose amplitudes should be trended over time; small baseline eccentricity may exist due to manufacturing tolerances, so growth is often the key indicator.
2.3 Mass imbalance
Mass imbalance arises when the rotor’s center of mass does not align with its rotational axis, whether due to the rotor, coupling, fan, or driven load.
Primary detection method: vibration analysis (primary), with current-based indicators as supporting evidence.
Typical signature: a pronounced increase at the rotational frequency component () in the vibration spectrum.
2.4 Rolling-element bearing faults
Bearing defects may occur on the outer race, inner race, rolling elements, or cage. Localized damage produces impulsive vibration as rolling elements pass over the defect.
Primary detection method: vibration spectrum analysis (often using envelope/demodulation methods in practice).
Typical signature: components at characteristic defect frequencies (e.g., and
), derived from shaft speed and bearing geometry.
2.5 Stator winding short circuits
Stator faults are typically driven by insulation degradation. While phase-to-phase or phase-to-ground faults often trigger protection systems, turn-to-turn faults may begin subtly and escalate due to localized heating.
Primary detection method: MCSA.
Typical signature: growth in specific current spectral components that may indicate early-stage turn-to-turn faults.
2.6 Shaft misalignment
Misalignment occurs when the motor shaft and driven shaft are not collinear (parallel/offset or angular). It increases cyclic stresses on the coupling, shaft, bearings, and the driven machine.
Primary detection method: vibration spectrum analysis.
Typical signature: increased amplitude at running speed, sometimes accompanied by
and
components in severe cases.
3. Why combined current and vibration analysis improves diagnosis
A single measurement domain can leave important fault modes undetected or create ambiguity. A combined approach—simultaneous analysis of stator current and vibration—offers:
- Broader fault coverage, capturing both electrical- and mechanical-dominant defects
- Higher diagnostic confidence through cross-verification, particularly for faults that influence both electromagnetic and mechanical behavior
A modern online motor condition monitoring program should therefore support synchronized acquisition and analysis across these domains.
4. Practical implementation example: WiseGrid Energy MCM1 condition monitoring system
MCM1 is an advanced system for condition monitoring of electric machines, designed to assess machine health through simultaneous analysis of current and vibration signals.
4.1 Measurement capabilities
- Inputs: three current channels, three voltage channels, and two vibration sensor inputs, enabling concurrent electrical and mechanical assessment.
- Measurement range: direct measurement up to
and
. For medium- and high-voltage machines, the system operates using instrument transformer secondary signals (CT/PT).
- Accuracy and sampling: approximately
measurement accuracy with sampling rates up to
, enabling spectral analysis from
to
.
- Field usability: battery-powered operation facilitates on-site measurements and deployments.
4.2 Fault detection and analytics scope
MCM1 supports identification of a broad set of fault modes, including:
- Rotor and stator-related defects such as broken rotor bars, air-gap eccentricity, and stator winding faults (including turn-to-turn short circuits)
- Mechanical issues such as mass imbalance, misalignment, looseness, and bearing faults
- Power analysis functions, including measurement of total harmonic distortion
, power factor, voltage/current sequence components, and active/reactive power
4.3 Intelligence, reporting, and data management
- Embedded diagnostic algorithms trained on extensive datasets to report fault likelihood as a percentage.
- Built-in bearing library to facilitate calculation of characteristic bearing defect frequencies.
- Vibration evaluation aligned with standards such as ISO
, supporting classification into condition zones (e.g., normal to critical).
- Reporting and storage: automatic PDF report generation,
internal storage, and optional cloud connectivity to enable long-term trending and overcome local storage constraints.
- Data export: wireless access via the device’s built-in Wi‑Fi using a web browser (no additional software required), with USB-based export also available.
4.4 Applications
While commonly used for induction motors, the platform is applicable to other rotating assets such as generators (hydro, gas, etc.), fans, pumps, and conveyors across industries including petrochemical plants, mining, and power generation.
Closing remarks
Online condition monitoring enables earlier identification of developing faults in induction motors by tracking characteristic signatures in electrical and mechanical signals. When integrated into CBM and PdM strategies, these techniques reduce unplanned downtime, improve safety, and lower lifecycle maintenance costs. The most robust outcomes are achieved when current- and vibration-based diagnostics are combined and trended over time to support evidence-based intervention planning, which is already available in MCM1 device.
Youtube link of the webinar: https://www.youtube.com/watch?v=IX7A1xc0Sf4
