Warehouse Fire Detection System Based on Multi-Sensor Information Fusion.

Area of Science:

• Engineering
• Computer Science
• Safety Science

Background:

• Traditional fire detection systems suffer from false negatives, false positives, and delayed responses.
• Warehouse environments present unique challenges for effective fire detection.

Purpose of the Study:

• To develop an intelligent warehouse fire detection scheme using multi-sensor information fusion.
• To improve the accuracy and speed of fire detection while minimizing false alarms.

Main Methods:

• Construction of a ZigBee wireless sensor network integrating temperature, CO concentration, and smoke sensors.
• Implementation of an improved Grubbs criterion for outlier elimination and median for data processing.
• Development of a BP neural network optimized by the cosine decreasing inertia weight particle swarm optimization (CIW-PSO) algorithm.
• Integration of a fuzzy logic reasoning mechanism for decision-making.

Main Results:

• The proposed system achieved early fire warning in a 50 m × 100 m warehouse.
• Demonstrated a 42% reduction in false alarm rate compared to single-sensor systems.
• Achieved a 35% reduction in response time.

Conclusions:

• The multi-sensor fusion approach offers an efficient and reliable intelligent solution for warehouse fire safety.
• The CIW-PSO optimized BP neural network and fuzzy logic enhance detection accuracy and reduce misjudgments.
• The system effectively addresses limitations of traditional fire detection methods.