Detection Of Component Gas Emitted From Coal-fired Power Plant Via Graphene-based Film Sensor Array

The flue gas emitted from coal-fired power plants and combustion processes,such as NO₂,SO₂ and CO,is one kind of heavy pollution to environment.As the advancement of science and technology,as well as more attention to atmospheric environment,the requirements of the detection and analysis of multi-component of flue gases are on great rise.But existing detection methods cannot meet them.Therefore,sensor array based on multiple sensor elements with excellent sensing performances is one of the best choices to be realized.The sensor array combining with patter recognition and data mining technology based on intelligent algorithm can realize mixed gas detection with high precision,which plays a significant role in the field of harmful multi-component gases detection.In this thesis,the detection system of multi-component flue gases emitted from coal-fired power plant is constructed by using gas sensor array based on nano-decoration graphene and combining with intelligent algorithm processing technology.Firstly,four kinds of metal oxide?MOx?modification reduced graphene oxide?rGO?are selected as sensing materials toward flue gas.The sensor array is fabricated via facile hydrothermal route and layer-by-layer?LbL?self-assembly method on the PCB substrate with interdigital electrodes.The microstructure of sensing films is characterized by using Scanning electron microscope?SEM?,X-ray diffraction?XRD?and Raman spectra.The properties of four kinds of MOx/rGO composite sensing films?SnO₂/rGO,TiO₂/rGO,CuO/rGO,ZnO/rGO?toward target gas NO₂,SO₂,CO are investigated at room temperature,including dynamic response,sensitivity,response/recovery time,repeatability,stability and selectivity characteristics.The testing results reveal that presented sensing films can response to sub-ppm level and ppb level gas at room temperature.Based on heterojuction interfaces,properties of physical and chemistry,the possible sensing mechanism for MOx/rGO response to target gas is discussed in detail.Furthermore,a novel sensor array with high-performance is constructed with four sensor elements.Resistance responses of the sensor array toward single component,two components and three components of flue gases are obtained,respectively.Cross sensitivity of sensor array toward NO₂,SO₂,CO is investigated.Finally,BP neural network,LS-SVM and RBF neural network model are designed to detect multi-component flue gases of coal-fired power plant.RBF neural network is the best pattern recognition of multi-component flue gases.The relative error of RBF network is less than 0.4%.The speed of training is very fast.The model of convergence effect is good and will be applied widely.In summary,this work innovatively achieves the recognition and prediction of flue gas component.Research work in this paper is helpful for the development of new intelligent system of coal-fired power plant flue gas detection,and has the important reality urgency and long-term social-economic benefits for real-time continuous emissions monitoring in environment and industrial toxic gas.