Overview
Timeframe: 2024-2025
Category: Manufacturing Automation / Controls Engineering
Automated machine faults flowed straight to dispatch tickets so techs were paged within seconds instead of waiting for operator reports.
The Problem
Manual dispatching delayed responses and produced inconsistent downtime data. We needed low-latency alerts without touching the MES backend.
Constraints
- No throughput loss in live production
- Must survive PLC resets and network hiccups
- Minimal changes to existing enterprise systems
Approach
I built the end-to-end fault pipeline:
- Mapped PLC states to actionable maintenance, quality, and material events
- Added PLC and HMI logic to encode context and suppress duplicates
- Created a lightweight API service to generate dispatches with retries and watchdogs
- Validated alarms in simulation first, then under live production load
Tools & Technologies
PLC Programming
HMI
REST APIs
SQL
Ladder Logic
Structured Text
Impact
- Cut time from fault to notification and improved MTTR
- Captured standardized downtime data for RCA
- Removed dependency on operator availability
Key Learnings
- Reliability patterns (retries, watchdogs, debouncing) matter as much as the logic.
- Right-sized fault classification keeps trust high and alerts actionable.