Capacitor Bank Health Monitor with Dashboard

Power-factor-correction capacitor banks quietly degrade. Metallized-film capacitors lose capacitance as their self-healing events accumulate, cans run hotter as losses rise, and one failed stage silently shifts the whole bank below its design kVAr — until the utility's power-factor penalty arrives to announce it. Most plants discover this at billing time because nobody measures individual stages. This project makes bank health visible: a monitoring add-on that self-tests each stage, estimates effective capacitance from the stage current it actually draws, watches temperature per stage, and presents degradation trends and watch/fail alerts on a dashboard. The student learns why banks fail and how the failure announces itself electrically, long before the bill does.

1. The demo bank's six capacitor stages wire through contactors, with an ACS712-class current sensor and NTC thermistor on each stage.
2. At the scheduled self-test (design 06:00 daily, or on demand), the controller switches each stage in alone while other loads are steady.
3. The controller measures the stage's reactive current and the bus voltage, then estimates effective capacitance from C = I / (2πfV).
4. Each estimate is compared with the buyer-calibrated nameplate baseline; the percentage deviation is logged and trended per stage.
5. Deviations past −15% raise watch alerts; past −25% the stage is marked failed, its contactor is held off, and the bank's effective kVAr is recomputed.
6. Stage temperatures are sampled continuously; a stage hotter than the bank mean by the configured margin raises a thermal watch.
7. All data publishes over MQTT to the dashboard, which renders the deviation chart, stage cards, health index and the 24-hour alert log.

Tech stack:

• ESP32 (controller)
• ACS712-class current sensors (per stage)
• NTC thermistors (per-stage temperature)
• Capacitor bank demo panel with contactors
• Modbus / MQTT telemetry
• Web dashboard (HTML, CSS, JavaScript)
• Arduino IDE (C/C++ firmware)
• 16x2 LCD with buzzer alerts

• [Per-stage capacitance self-test] Each stage is switched in alone against the known bus voltage; effective capacitance follows from the measured stage current (C = I / 2πfV) — no LCR meter needed.
• [Degradation trending] Every self-test result is logged and plotted per stage, so gradual capacitance loss shows as a downward trend weeks before any threshold trips.
• [Watch / fail thresholds] A −15% deviation raises a watch alert (plan replacement); −25% marks the stage failed and the controller holds its contactor off.
• [Per-stage temperature watch] NTC sensors track each can's temperature; a stage running hot relative to the bank mean is flagged as an early degradation signal.
• [Bank health index] A 0–100 score combines stage deviations, temperatures and failed-stage count into one at-a-glance number.
• [Automatic failed-stage hold-off] Failed stages are excluded from the switching sequence and the dashboard notes the rebalanced effective kVAr.
• [Event log] Self-test results, contactor operations and alerts are timestamped for the report's measured-data appendix.
• [Buyer-run calibration] The student calibrates the estimator against nameplate capacitance at commissioning, so thresholds rest on verified baseline values.

• Demo capacitor bank panel: 6 capacitor stages, contactors, per-stage current sensors and NTCs
• ESP32 firmware with self-test sequencer, capacitance estimator and hold-off logic
• Web dashboard source (deviation chart, stage cards, health index, alert log)
• Nameplate-baseline calibration procedure with LCR cross-check sheet
• Wiring diagram and contactor driving notes
• Project report PDF (capacitor ageing theory, reactive power, estimation method)
• PPT presentation for final review
• Viva Q&A preparation document (capacitance from current, self-healing, PF correction, contactor switching)

• The estimator infers capacitance from current at line frequency; harmonic distortion on the bus affects the reading, which the report states openly.
• Estimation accuracy is a design target verified by the buyer against an LCR meter — the kit does not claim laboratory-grade capacitance measurement.
• The demo bank is low-voltage scale; a real LT/HT bank installation is the work of qualified personnel and is out of scope.
• This is a monitoring add-on, not certified protection equipment — it does not replace the bank's fuses, protection relays or maintenance schedule.
• Self-tests need a reasonably steady load during the test window; the firmware skips and retries if the bus is unstable.