Real-Time Monitoring Ozone by an Intelligent Sensor Terminal With Low Cost

Ozone (O-3) is common in industrial process, and the operating condition of equipments can be estimated by monitoring the concentration of O-3. In this article, the core innovation lies in the fabrication of an O-3 sensor based on InAs nanowires (NWs) grown by metal organic chemical vapor deposition (MOCVD), which features sensitive and selective to O-3 at room temperature. To further enhance the application effectiveness of the sensor, a low-cost terminal-to-cloud intelligent O-3 sensor has been developed. An adaptive Kalman filtering algorithm is proposed and implemented at the terminal, resulting in significant improvements in both the signal-to-noise ratio and real-time response. Machine learning method is employed to predict O-3 concentration, and the particle swarm optimization (PSO) algorithm is also used for the artificial neural network (ANN) parameters to predict the O-3 concentration by processing the output signals in cloud. Experimental results show that the error between the predicted and real concentration of O-3 is less than 2%. Thus, the intelligent sensor terminal has potential to be deployed on industrial equipments for gas leakage fault detection with low cost.