AIoT Risk System
Maritime Battery Thermal-Runaway Monitoring
An AIoT edge-cloud monitoring project for EV battery thermal-runaway risk during maritime transport.
Problem
Battery thermal runaway at sea requires decisions under noisy sensor signals, network loss, physical vessel constraints, and high safety stakes.
Workflow
- 01Define multi-source sensor fusion rules using temperature-rise rate, CO/CO2 concentration, and humidity.
- 02Discretize continuous signals into safe, overheated, and burning business states.
- 03Map states to silent monitoring, automatic vehicle evacuation, and high-pressure water-cannon response.
- 04Design armored pressure-resistant wireless nodes and offline edge operation for public-sea network loss.
Evidence
Patent authorization
National invention patent ZL 2024 1 0314143.3, certificate no. 8295894, for a battery protection processing system and method.
UN ITC 2026
Advanced on site in Rotterdam to the October 2026 ITC finals, with global finalist selection 10 / 44.
Tsinghua and UNITAR awards
Tsinghua engineering-management best result / gold; UNITAR SDG Open Hack CSG venue second prize and X-lab second prize.
Prototype integration
Sensor nodes were connected to the China Unicom Gewu IoT platform and validated in a COSCO Shipping Hainan ferry sea-trial context.
Boundary
- This is a monitoring and decision-support system, not a claim of full vessel automation.
- Public proof summarizes architecture and awards without exposing partner-sensitive operational material.
- The system uses AIoT state logic and edge-cloud constraints; it is not positioned as pure model research.
Role Mapping
- AIoT product: connects sensors, edge behavior, cloud supervision, and safety response.
- Industrial AI / solution product: translates physical constraints into state machines and operating procedures.
- Data product: turns continuous noisy signals into actionable risk levels.