SES-005 Smart Energy Systems

Legacy sensor network → predictive maintenance platform. Not a single sensor replaced.

14 weeks implementation·93% cost savings·€86,400/year saved

IoT Sensor Network

RS-485 · MQTT · InfluxDB · TensorFlow Lite · Predictive Maintenance

The Situation

A facility with 47 vibration and temperature sensors — all 15 years old, RS-485 bus, no cloud connectivity, no real-time monitoring. Plant manager read values manually twice daily. New sensor network quote: €1,240,000.

Approach: RS-485 → MQTT → Cloud

All 47 sensors were working perfectly. The hardware was fine — what was outdated was the connectivity layer. We built a custom RS-485 to MQTT gateway that reads all sensors and forwards data in real time to a cloud platform.

RS-485 Bus 47 sensors, Modbus RTU protocol, 15 years of operation

Custom Gateway ESP32-based, reads all 47 sensors, sends via MQTT every 10 seconds

Messaging Broker RabbitMQ (AMQP, MQTT via plugin) on-premise, buffering for offline scenarios

Time Series DB InfluxDB — 18 months of historical data imported

ML Anomaly Detection TensorFlow Lite — 94% accuracy in failure prediction

React Dashboard Real-time OEE, alert management, maintenance planner

Kosteneinsparung

Direktvergleich

New sensor network

€1,240,000

Solvetronix

€87,000

Sie sparen

€1,153,000

Einsparung

93%

ROI typisch: 12 months

Additional annual savings through predictive maintenance

The ML model predicts failures 3–7 days in advance. In the first year: 3 major failures prevented. Annual savings: €86,400 in avoided downtime and emergency repairs.

Legacy sensor network → predictive maintenance platform. Not a single sensor replaced.

Approach: RS-485 → MQTT → Cloud

Legacy sensor network?