Concentration prediction of binary-component gases based on SnO2 sensor array with machine learning algorithms

Abstract

Ammonia (NH₃) and formaldehyde (HCHO) are common and highly toxic indoor gases, which are released into the environment through furniture, decorative materials, etc. When the human body is in the environment of ammonia and formaldehyde for a long time, it will cause irreversible harm to the human body. Therefore, it is of great significance for human health to detect low concentration NH₃ and HCHO mixtures efficiently. The gas sensor based on SnO₂ (tin oxide) has the characteristics of fast response, fast recovery and good selectivity, so it has a broad application prospect in detecting indoor toxic gases. In this paper, tin oxide and copper-doped tin oxide gas-sensing materials synthesized by bio-template and hydrothermal methods are introduced for self-made gas-sensing sensor arrays. The sensor array combines SSA-BPNN (Sparrow search algorithm optimized Back-propagation neural network) algorithm to predict and analyze indoor toxic gas concentration. The elemental composition of SnO₂ nanomaterials was characterized and analyzed by XRD (X-ray diffraction). The gas sensing characteristics of the sensor array were tested. The sensor array was combined with a neural network algorithm to successfully predict the concentration information of mixed toxic gases.