Design Of Bionic Chambers For Hand-held Electronic Noses

As a novel modern detection device,electronic nose(e-nose)has learned a lot from the mammalian olfactory system and has the advantages of portability and rapidity.However,for the current e-nose systems,there are still some shortcomings in the design of gas chambers,such as large volume and lack of structure optimization,resulting in poor data collection quality and slow cleaning speed.In view of these problems,this thesis takes liquor identification as the application background,and focuses on the design of hand-held electronic nose and its bionic chamber.The main research work is as follows:(1)A hand-held electronic nose device for liquor identification was designed.The hardware part includes ARM(Advanced RISC Machine)system board,air pump,solenoid valve,LCD(Liquid Crystal Display)screen,sensor array,peripheral drive,signal filtering/amplification,AD(Analog to Digital)conversion,air path and the bionic chamber.The software part was designed with QT framework based on Linux system.After the system is turned on,only the relevant parameters(including time,material,and sampling frequency,etc.)need to be set,then the air pump and solenoid valve can be automatically controlled to complete the tasks of sampling,cleaning and identification for the liquor samples.(2)Bionic chambers for handheld electronic noses were designed.Inspired by the structures of turbinate and nasal cavity of the mammalian animals,a variety of bionic chamber schemes were designed.The flow characteristics of the designed bionic chambers and the traditional square chamber were simulated using a fluid dynamics software.The results show that the designed bionic chambers can not only improve the quality of e-nose sampling data,but also raise the sampling speed.(3)The liquor online identification experiment has verified the superiority of the designed bionic chamber.The algorithms related to data preprocessing,feature extraction and dimensionality reduction,and support vector machine(SVM)were integrated into the system board to achieve the online identification of liquor.Through the liquor identification experiment,the sampling and cleaning time of the e-nose with the eight gas chamber structures were analyzed to verify the accuracy of the simulation results.Six kinds of Chinese liquors were sampled by using both the optimal bionic chamber and the square chamber,and the sampled data were identified with the SVM algorithm.The results show that the sampling time of e-nose with the bionic chamber was much shorter than that with the square chamber,and the classification accuracy is higher.