Shop-floor process automation with IoT line monitoring and OEE dashboard for an Indian SME manufacturer

Context

The client runs twelve production lines and was logging line state hourly on paper. Downtime cause-codes were captured retrospectively by supervisors, and Overall Equipment Effectiveness (OEE) was reported monthly by spreadsheet.

There was no real-time view of what was running, what had stopped, or why.

Challenge

Two constraints framed the problem:

1. Off-the-shelf MES platforms were roughly ten times the available budget — and most assumed greenfield instrumentation that did not match the existing floor.
2. The existing PLCs had to stay. They were paid for, working, and replacing them was not on the table.

The build had to retrofit, not replace.

Approach

We added a thin sensing layer on top of the existing infrastructure:

• ESP32-based controllers read state signals from existing PLC outputs — no PLC re-programming required.
• MQTT broker on the local network handles ingest. Network failures buffer locally and replay on reconnect.
• Node.js ingest service writes to a time-series store optimised for downsampled queries.
• Supervisor dashboard in React shows current line state, downtime alerts, shift-end OEE rollup and trailing seven-day trends.
• Cause-coding moved from paper to a tablet at each line — supervisors tag downtime reasons in the moment, not after the fact.

Outcome

The first week of live data surfaced a recurring micro-stop on one line that nobody had previously associated with a specific cause — it was eating about seven percentage points of OEE on that line alone. Fixing it required no capital, only the visibility to know it was happening.

All twelve lines were live inside two months. Monthly OEE reports are now generated from the system rather than reconstructed from logbooks.