Electrical Signature Analysis (ESA): What It Is and How It Works

What Is Electrical Signature Analysis?

Electrical Signature Analysis is the practice of treating the current and voltage signals feeding an electric motor as a diagnostic channel. Every mechanical or electrical anomaly inside the motor, from a cracked rotor bar to a degrading bearing race, slightly distorts how the motor draws power. ESA captures that distortion, decomposes it into its frequency components, and compares the result against known fault frequency patterns to identify which component is failing and how severely.

Because the measurement point is the power cable rather than the motor casing, ESA works on motors in hazardous locations, inside enclosures, and on machines running under full production load, conditions where installing accelerometers or thermocouples would be impractical or impossible.

How ESA Works

Signal Capture

A current transformer (CT) clips around the supply cable to the motor and outputs a proportional analog current signal. A voltage transducer connects in parallel across two phases. Both channels are sampled simultaneously at rates typically between 10 kHz and 50 kHz by a data acquisition unit. Sampling rate matters: the Nyquist theorem requires the sampling frequency to be at least twice the highest fault frequency of interest, and many motor faults produce harmonics well above 1 kHz.

FFT and Spectral Analysis

The acquired time-domain signal is converted into a frequency-domain spectrum using the Fast Fourier Transform. The result is a plot of amplitude versus frequency. A healthy motor shows a dominant peak at the supply fundamental frequency (50 Hz or 60 Hz) with small harmonic content. A motor with a developing fault shows additional spectral components at fault-specific frequencies, called sidebands, clustered around the fundamental and its harmonics.

The frequency resolution of the FFT determines how closely spaced two sidebands can be before they merge. Resolution (Hz) equals the sampling rate divided by the number of samples collected. To resolve sidebands separated by 0.1 Hz, for example, a 10-second measurement at 10 kHz sampling yields 100,000 samples and a resolution of 0.1 Hz.

Fault Frequency Calculation

Each fault type produces sidebands at predictable frequencies derived from motor geometry and operating slip. The most important fault frequencies are:

• Broken rotor bars: sidebands at f_s(1 +/- 2ks), where f_s is supply frequency, s is per-unit slip, and k = 1, 2, 3...
• Bearing outer race (BPFO): f_s +/- n * f_BPFO, where f_BPFO = (N_b/2) * f_r * (1 - d/D * cos alpha)
• Static eccentricity: sidebands at f_s +/- n * f_r, where f_r is rotor mechanical frequency
• Dynamic eccentricity: sidebands at f_s +/- n * (f_r +/- f_s)

These formulas mean fault sidebands shift with operating speed. A motor running at higher load has lower slip, so broken rotor bar sidebands move closer to the fundamental. Accurate diagnosis requires knowing the exact operating slip at the time of measurement.

Worked Numerical Example: Broken Rotor Bar Detection

Consider a 4-pole, 60 Hz induction motor running at 1,746 rpm under load.

Synchronous speed: (120 * 60) / 4 = 1,800 rpm

Per-unit slip: s = (1,800 - 1,746) / 1,800 = 0.030

Lower sideband (k=1): f_s(1 - 2s) = 60 * (1 - 0.06) = 56.4 Hz

Upper sideband (k=1): f_s(1 + 2s) = 60 * (1 + 0.06) = 63.6 Hz

If the FFT spectrum shows peaks at 56.4 Hz and 63.6 Hz with amplitudes more than 40 dB below the 60 Hz fundamental, the motor is healthy. Amplitudes rising to within 25 to 35 dB of the fundamental indicate one or two broken rotor bars. Amplitudes within 20 dB signal an advanced fault requiring prompt action. This threshold guidance aligns with the practice described in IEEE Standard 1776 for motor electrical testing.

ESA Variants and Techniques

Motor Current Signature Analysis (MCSA)

Machine condition monitoring using MCSA analyzes only the stator current waveform. It is the most widely implemented ESA variant because current measurement is straightforward and the technique is mature. MCSA is best suited to detecting rotor faults (broken bars, rotor eccentricity) and load-side mechanical faults that modulate the motor's electromagnetic torque.

MCSA does have limits: it is less sensitive to stator winding faults and insulation degradation, which do not strongly modulate the current spectrum under early fault conditions.

Extended Park's Vector Approach (EPVA)

EPVA transforms three-phase current measurements into a two-dimensional representation (the Park's vector, also called the current space vector). In a healthy motor, this vector traces a perfect circle. Faults cause elliptical distortion, asymmetry, or harmonic rings. EPVA is particularly sensitive to stator inter-turn short circuits and supply voltage unbalance that MCSA may underestimate.

The Park's vector modulus is computed as:

i_d = (2/3)[i_a - (1/2)i_b - (1/2)i_c]

i_q = (2/3)[(sqrt(3)/2)i_b - (sqrt(3)/2)i_c]

The FFT of i_d² + i_q² reveals fault harmonics that appear at twice the supply frequency and its multiples for stator faults, distinguishing them cleanly from rotor fault signatures.

Instantaneous Power Analysis

Instantaneous power p(t) = v(t) * i(t) combines voltage and current into a single signal. Faults that modulate both signals simultaneously appear more strongly in the power spectrum than in either signal alone, improving signal-to-noise ratio. Instantaneous power analysis is especially effective at detecting load-side mechanical faults, such as gearbox tooth defects and pump cavitation, that create periodic torque ripple transmitted back through the electrical signal.

Failure Modes ESA Detects

Broken Rotor Bars

A broken rotor bar in an induction motor disrupts the symmetric distribution of rotor currents, creating a backward-rotating magnetic field component that induces sidebands at (1 +/- 2ks)f_s in the stator current. A single broken bar in a 400 kW pump motor may not immediately cause visible performance degradation, but the thermal cycling caused by the asymmetric current flow accelerates damage to adjacent bars. Without ESA, the fault typically progresses to catastrophic failure, destroying the rotor cage and requiring a full rewind or motor replacement costing five to fifteen times the price of a planned rotor repair.

Bearing Faults

Predictive maintenance programs targeting bearings traditionally rely on vibration analysis. ESA provides a complementary view: as bearing race defects worsen, the resulting load modulation produces current sidebands at the bearing defect frequencies (BPFO, BPFI, BSF, FTF). ESA detects these sidebands through the supply cable without accelerometers, making it valuable for motors inside enclosures, submerged pump motors, and other inaccessible installations.

Air-Gap Eccentricity

Eccentricity occurs when the rotor center does not coincide with the stator bore center. Static eccentricity (fixed offset) creates a steady magnetic pull on one side of the rotor; dynamic eccentricity (offset that rotates with the rotor) causes the magnetic pull to rotate. Both create specific sideband patterns in the current spectrum. Mixed eccentricity, the most common real-world condition, shows sidebands at f_s +/- n(f_r +/- kf_s). Uncorrected eccentricity leads to stator-rotor rub, which destroys both the stator winding and rotor surface.

Insulation Degradation

Inter-turn faults in the stator winding increase current in the faulted coil and produce asymmetry in the three-phase current balance. Early-stage insulation faults are detectable through voltage unbalance monitoring and EPVA before they develop into full phase-to-phase or phase-to-ground faults that trip the motor and potentially damage associated drives and switchgear.

Load-Side Mechanical Faults

Gearbox tooth defects, pump impeller damage, compressor valve failures, and fan blade imbalance all produce periodic torque variations that appear in the motor current spectrum. ESA can diagnose driven equipment faults without sensors on the driven machine, which is particularly valuable in multi-stage gearboxes where sensor placement is impractical.

ESA vs. Vibration Analysis

Best-practice motor health programs use both techniques together. ESA covers rotor and electrical faults from the control room or MCC; vibration analysis covers mechanical faults at the motor casing and driven equipment. Together they close the diagnostic gaps left by either method alone.

Where ESA Is Applied

Induction Motors in Manufacturing

Induction motors driving pumps, compressors, conveyors, fans, and machine tool spindles are the primary ESA targets in manufacturing plants. A 250 kW cooling water pump motor that fails unexpectedly can halt an entire production line; ESA provides early warning 4 to 12 weeks before failure in most documented cases, allowing a scheduled replacement during planned downtime.

High-Voltage Motors in Process Industries

Large high-voltage motors in chemical plants, refineries, and pulp mills are expensive to rewind and have long lead times for replacement rotors. ESA monitors these assets continuously from the switchgear room without requiring personnel to enter high-voltage areas or shutting down the process for manual testing.

Submersible Pump Motors

Submersible motors in water treatment, wastewater, and mining dewatering applications cannot be fitted with conventional vibration sensors without pulling the pump from the well or sump. ESA applied at the surface cable connection monitors bearing condition, rotor integrity, and winding health continuously without any physical intervention.

Motor Control Center Integration

Modern asset condition monitoring platforms install current and voltage measurement modules directly inside the MCC, enabling continuous ESA on every monitored motor from a single connection point. This approach scales motor monitoring across a facility without deploying individual sensors at each machine.

Relevant Standards

• IEEE Standard 1776: recommended practice for thermal evaluation of unsealed or sealed insulation systems for AC electric motors rated 15,000 V and below; references electrical testing including current signature analysis.
• IEC 60034-4: methods for determining synchronous machine quantities, including electrical measurement techniques applicable to motor testing.
• NEMA MG 1: motors and generators standards covering performance testing that electrical signature measurements support.
• ISO 13373-1 and -9: general requirements for vibration monitoring of rotating machines; ESA is used alongside these vibration standards in comprehensive motor health programs.

ESA in a Condition-Based Maintenance Program

Condition-based maintenance using ESA follows a structured workflow: continuous data collection, automated spectral analysis against fault frequency libraries, severity trending over time, and alert generation when sideband amplitudes cross defined thresholds. The goal is to identify the P-F interval (the period between detectable fault onset and functional failure) and schedule intervention before failure while minimizing unnecessary maintenance on healthy motors.

Continuous ESA through permanently installed sensors delivers several advantages over periodic manual measurements. Faults that develop rapidly between quarterly surveys, such as a rotor bar fractured by a mechanical impact or a sudden voltage unbalance event, are captured within hours rather than missed for months. Trending data also reveals whether a fault is stable or accelerating, which determines urgency and repair planning.

The Bottom Line

Electrical Signature Analysis gives maintenance and reliability engineers a non-intrusive window into motor health that no other single technique matches. By capturing fault signatures in the current and voltage waveforms at the supply cable, ESA detects broken rotor bars, bearing defects, eccentricity, and insulation faults without stopping the machine, entering hazardous areas, or installing sensors directly on the motor body.

ESA is not a replacement for vibration analysis; it is a complement to it. A program that combines both techniques covers the full spectrum of motor failure modes, from mechanical imbalance and misalignment to electrical winding faults and rotor damage, providing the advance warning that transforms emergency motor failures into planned, budgeted repairs.

The practical outcome is straightforward: motors that would have failed catastrophically during production instead get repaired during scheduled downtime, reducing unplanned stops, extending asset life, and lowering the total cost of motor ownership across the fleet.

Frequently Asked Questions