Hazardous Gas Monitoring System for Power Cable Tunnel Based on PID and MLPNN

With the increasing demand for safety assurance in underground power cable tunnels, an online Photoionization Detector (PID) gas monitoring and fault warning system for cables was designed. A gas data acquisition hardware platform for cable tunnels was built using PID modules, ZigBee communication networks, and GD32 microchip. The data collected by the platform were transmitted to a data server using a 5G communication module. A monitoring center was developed to monitor Volatile Organic Compounds (VOC) in real time. The system processed signal noise caused by complex power tunnel environments using a signal noise, eliminating noise interference and sensor voltage mutations. The Multi-layer Perception Neural Network (MLPNN) was built to solve the cross-sensitivity problem commonly found in gas sensors and improve data accuracy. An improved particle swarm optimization algorithm was introduced to accelerate network convergence and achieve early warning of VOC concentration anomalies in power tunnels.