The Challenge
The client operated 18 stamping and press machines across two production halls. Shift supervisors tracked production on paper-based logs and a basic Excel spreadsheet updated at end-of-shift. Real OEE was unknown — estimates ranged from 55% to 70% depending on who you asked.
Unplanned breakdowns were the biggest pain point. On average, 4–6 machines failed unexpectedly per month, each requiring 2–8 hours of downtime for emergency repair. The maintenance team operated entirely in reactive mode. There was no visibility into machine health, and bearing failures — the most common fault — were only caught after audible damage had already occurred.
Management wanted a single screen showing live production status for all machines, and a way to predict failures before they caused downtime.
Our Approach
MordeTech deployed a phased IIoT connectivity project over 18 weeks:
- Phase 1 (weeks 1–6): Installed Siemens SIMATIC ET 200SP edge I/O units on all 18 machines, connecting spindle speed, cycle count, power draw, and fault registers via OPC-UA to a local MQTT broker
- Phase 2 (weeks 7–12): Deployed MordeTech OEE Pro dashboard on a local server with cloud backup — showing real-time availability, performance, and quality per machine, per shift, per day
- Phase 3 (weeks 13–18): Installed triaxial MEMS vibration sensors on 12 priority machines (those with highest breakdown history), feeding an ML model that detects bearing wear signatures up to 5 days before failure
"For the first time, I can see from my phone at 11pm what's happening on the floor. Last month the system warned us about a bearing on Press #7. We changed it on Sunday — no downtime. That one alert alone saved us a full day's production."
— Plant Manager, Mid-size Press Shop, Aurangabad (name withheld at client request)
Technical Implementation
- 18× Siemens SIMATIC ET 200SP edge I/O modules — OPC-UA server on each machine controller
- Eclipse Mosquitto MQTT broker on local industrial PC — aggregates all machine data streams
- Node-RED flows for data normalisation, OEE calculation, and alert routing
- MordeTech OEE Pro — React dashboard with Grafana backend, accessible from browser and mobile
- 12× triaxial MEMS vibration sensors (SparkFun LSM6DS3) on main bearing housings
- Edge ML model (isolation forest + FFT frequency analysis) for bearing wear detection
- WhatsApp + email alerts for predictive maintenance warnings, auto-routed to maintenance team
- Shift-level report PDFs auto-generated and emailed to management every 8 hours
Results After 6 Months
- OEE improved from 61% to 83% — a 22-point gain across all 18 machines
- Unplanned breakdowns dropped from 4–6 per month to zero in months 4, 5, and 6
- Predictive alerts issued on 7 bearing failures — all actioned during scheduled maintenance windows
- Average shift efficiency reports now completed in real-time vs. 2-day lag previously
- Energy consumption per part reduced 18% through idle-power management via OPC-UA control
- Maintenance team shifted from reactive to proactive — planned maintenance up from 34% to 91%
- Full ROI achieved within 14 months of deployment