High-Performance Multi-Component Gas Detection in Based on Sensor Array and Neural Network Model

Abstract

The gas detection is crucial important for environmental monitoring, biological and chemical terrorism, mining safety and disease diagnosis. In order to determine the multi-component gas including carbon monoxide, hydrogen and methane in the mining, sensor array combining with effective measurement model are used in this paper to decouple its cross-sensitivity and broad-spectrum analysis. Multivariate nonlinear regression and neural network are employed to establish the measurement model. The experimental results show the effectiveness of the two methods, and artificial neural network in prediction accuracy is better than the multivariate nonlinear regression method, which yields a higher measuring accuracy for constructing high-performance electronic nose in mining applications.