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

Ozone (O3) is common in industrial process, and the operating condition of equipments can be estimated by monitoring the concentration of O3. In this article, the core innovation lies in the fabrication of an O3 sensor based on InAs nanowires (NWs) grown by metal organic chemical vapor deposition (MOCVD), which features sensitive and selective to O3 at room temperature. To further enhance the application effectiveness of the sensor, a low-cost terminal-to-cloud intelligent O3 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 O3 concentration, and the particle swarm optimization (PSO) algorithm is also used for the artificial neural network (ANN) parameters to predict the O3 concentration by processing the output signals in cloud. Experimental results show that the error between the predicted and real concentration of O3 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.