Motor Fault Diagnosis Dashboard using Current Signature Analysis

Induction motors fail gradually — a cracked rotor bar, a pitted bearing or a shorted stator turn all leave traces in the motor's current long before the machine stops. Motor Current Signature Analysis (MCSA) reads those traces without vibration sensors or dismantling: faults modulate the air-gap flux and produce characteristic frequency sidebands around the supply frequency in the stator current spectrum. This project turns that industrial technique into a buildable student prototype. A Hall-effect current sensor samples the motor current, the microcontroller computes a real-time FFT, and a web dashboard shows the live spectrum, the time-domain waveform, key KPIs and a diagnosis log that flags developing faults. It is a rare combination of power engineering, embedded systems and signal processing in one project.

1. The ACS712 Hall-effect sensor in series with the motor phase produces a voltage proportional to the instantaneous current, biased to mid-rail for the ADC.
2. The microcontroller samples the sensor at approximately 2 kHz through an anti-alias RC filter.
3. Each 512 ms window is Hanning-windowed and transformed with a 1024-point FFT (design target) to produce the current spectrum.
4. The 50 Hz fundamental is located; sideband amplitudes at the rotor-bar (50 ± 2·s·f), bearing (BPFO) and 3rd-harmonic frequencies are extracted.
5. A rule-based classifier flags a fault only when a signature exceeds its threshold for three consecutive windows, suppressing transient false alarms.
6. Spectrum frames and fault status are published over Wi-Fi (MQTT); the dashboard renders the spectrum, KPIs, event log and the local 16x2 LCD mirrors RMS current and status.

Tech stack:

• Arduino Uno / STM32 (edge acquisition & FFT)
• ACS712-20A Hall current sensor
• 16x2 LCD (local status display)
• MQTT over Wi-Fi (spectrum streaming)
• HTML/CSS/JS live dashboard
• 12 V DC / small single-phase motor (test load)

• [Live FFT spectrum dashboard] The current spectrum from 0–200 Hz renders in real time with the 50 Hz fundamental marked and a configurable trip threshold line — fault sidebands are visible the moment they cross it.
• [Broken-rotor-bar detection] Sidebands at 50 ± 2·s·f are extracted and compared against the threshold, with the classifier requiring the signature to persist across three consecutive windows before flagging.
• [Bearing-defect monitoring] Characteristic BPFO harmonics are tracked in the spectrum, giving an early, quantified warning of bearing damage instead of a post-failure discovery.
• [Stator-winding fault indication] Elevated 3rd-harmonic content and phase-imbalance signatures are reported with a plain-language diagnosis on the dashboard.
• [Fault-injection test panel] Select healthy, rotor, bearing or stator conditions in the demo to show exactly how each signature appears in the spectrum — ideal for the viva demonstration.
• [Time-domain current view] A live 200 ms window of the phase current runs alongside the spectrum, so the link between waveform distortion and spectral content is visible.
• [Diagnosis event log] Every classification with timestamp is logged on the dashboard, producing ready-made evidence for the project report.

• Complete sensing prototype (sensor, controller, LCD, wiring)
• Edge firmware (ADC sampling, FFT, classifier, MQTT publishing)
• Live web dashboard (spectrum, time-domain view, KPIs, event log, fault-injection panel)
• Wiring diagram and calibration procedure
• Project report PDF (MCSA theory, signal path, methodology, results)
• PPT presentation for final review
• Viva Q&A preparation document (FFT, sidebands, slip, classifier design)

• Demonstrates MCSA on a single phase of a small test motor — a full three-phase industrial deployment would need three sensors and Park-vector analysis.
• Fault signatures are shown with injected/test conditions on the lab rig; the build does not claim detection rates measured on industrial motors.
• −42 dB threshold and 1024-point FFT are design targets for the delivered firmware, stated honestly and adjustable in code.
• Classification is rule-based, not machine-learned — it detects known signatures, not unknown fault types.
• The prototype monitors; it does not trip the motor — protective tripping is listed as future scope.