SWCNTs-based MEMS gas sensor array and its pattern recognition based on deep belief networks of gases detection in oil-immersed transformers

MEMS gas sensor arrays and specially designed pattern recognition systems are the main research directions in the field of modern sensing technology in the engineering, especially in the smart sensing and monitoring of faults in large power equipment such as oil-immersed transformers. In this paper, the MEMS sensor array composed by eight SWCNTs-based (pure, OH functionalized, COOH functionalized, NH2 functionalized by ethylenediamine, NH2 functionalized by aniline, Ni-coated, Pd-doped, ZnO-doped) sensing units was palced in the fault characteristic gases (H2, CO, and C2H2) of oil-immersed transformers, and their gas-sensing characteristics were tested in single and mixed gas atmosphere. Combined with the DBN-DNN pattern recognition method, the qualitative identification and quantitative analysis of the sensor array in a mixed gas atmosphere was realized, and the accuracy and reliability of the results are higher than the traditional BPNN model.