The Research On The Pattern Recognition Of Gas Detection Based On The Saw Sensor Array

With the rapid development of science and technology, changing material and rising living standards, industrial and living exhaust gases increase hazards of the life and growth of plants and animals, people’s working and living environment quality is getting worse. Creating a good, healthy, safe and non-toxic living environment for plants and animals to survive and grow is very necessary. So to study one appropriate gas detection method and design a complete gas classification method are particularly important. Only by detecting and identifying the type and concentration of the gas, can you take timely and effective measures to deal with toxic and harmful gases, and can you provide a healthy and safe environment for animals and plants.In this thesis, it designs SAW sensor to detect toxic gases, which is based on the mechanism of surface acoustic wave. Furtherly, it develops a gas detection system based on surface acoustic wave sensor array, which is mainly used for identification and detection of the mustard simulant(2-CEES), the nerve agent simulant(DMMP) and general laboratory common gas, such as C7H8,C2H5 OH,DMF,C2H4Cl2.This thesis mainly introduces the basic principle of surface acoustic wave and the basic framework of the gas detection system, demarcates the training sample space and studies the principles and basic steps of the gas pattern recognition algorithms. The contents of this research include the following aspects:① Design the SAW gas sensor array: introduce the basic principles of the sensor the function structure, design the SAW gas sensor array’s the physical and function structure.② Research on gas pattern recognition system: study the basic structure of the pattern recognition system and main component, data processing method of the gas frequency signal, and some important gas detection system’s pattern recognition algorithm.③ Design the gas detection system based on surface acoustic wave gas sensor array: introduce the overall framework of the gas detection system, test and select sensitive units, collect data, and extract data preprocessing and data eigenvalues.④ Analysis of the detection result of gas detection system: predict and analysis pattern recognition result, analysis some external factor’s impact of the data collection process, such as the gas flow rate, gas type, chamber temperature, the reference cell, etc. and find the right solution.Using the hand-held gas detection device to test 2-CEES, DMMP, DMF, toluene, ethanol, dichloroethane, etc, the recognition rate of the gas in different concentration reaches 83.33%, the recognition rate between the organic agents and the general gas reaches 100%, and the concentration of identification error also is decreased by 30%.