Shipbuilding & Manufacturing
Shipyard welding equipment and crane condition monitoring
Operating states and types of welding equipment and a crane, estimated from attached sensor nodes alone, without any operating-signal link.

Project overview
- Site
- Shipyard welding hall and large crane machinery room
- Target
- Welding equipment gear drives, crane machinery-room rotating equipment
- Data
- Vibration, sound, magnetic field, temperature, humidity, pressure, air quality
- Objective
- Operating states and baselines without an operating-signal link
- Client
- Korean shipbuilder E
- Duration
- PoC
Challenge
Welding equipment and large cranes change operating state with work schedules and loads. Without an operating-signal link, nothing records when each machine ran or how, which makes a normal-operation baseline hard to set.
XyloZero deployment
Several XyloZero sensor nodes per machine collected vibration, sound, magnetic and environmental data, separating signals seen by one node from signals seen by several.
- Several sensor nodes per machine on welding gear drives and in the crane machinery room
- Collection by mounting alone, with no operating-signal link
- Vibration, sound, magnetic field and environmental data sent to the server
Analysis
- Running status determined from vibration and cross-checked with sound and magnetic field
- Operating types classified by spectrum and vibration level; welding-current intervals separated by magnetic field
- A sensor-language model fed with sensor data to identify operating types
Results
- Operating state and work schedule estimated per machine without operating signals
- Operating baselines per machine and work period
- Off-hours vibration and device fault intervals identified
- Baselines to be confirmed against operating logs and maintenance history for predictive maintenance