Construction And Optimization Design Of Modular Electronic Nose System Platform

Electronic nose is an intelligent electronic bionic instrument system designed to simulate human odor recognition mechanism.It is an emerging detection technology developed in recent years.Electronic nose technology can make up for the subjective shortcomings of traditional senses.Compared with other traditional gas detection methods,electronic nose has the advantages of high efficiency,convenience,and high accuracy.It is widely used in food science,agriculture,medical and environmental fields.At present,electronic noses are mainly divided into commercial electronic noses and electronic nose systems that perform olfactory identification and classification studies on certain substances.Although these electronic noses have certain detection capabilities for specific substances,they are limited by the fact that sensor arrays cannot be replaced.Once faced with different tested samples will lead to electronic nose detective results are not accurate enough,resulting in the detection of the electronic nose's subject to limitations,the ability to detect is single.Therefore,the purpose of this paper is to develop a modular multi-mechanism electronic nose system platform equipped with sensor array replacement module,select corresponding sensor array module according to different samples,and optimize the air chamber and sensor array to compensate for existing electronic nose detection's limitation,according to different detection targets can be designed to form a highly targeted electronic nose system,so that the electronic nose system platform to meet the needs of different detection targets..All the development of the electronic nose system platform is divided into three parts: gas chamber fluid simulation,hardware circuit design and software system development.This thesis uses Airpak software to conduct gas fluid simulation analysis of the electronic nose chamber,and optimizes the sensor array and gas chamber based on the simulation results.And design plug-in sensor module,according to different objects to be measured to select the appropriate sensor array response module to make it wider detection range.The electronic nose system hardware circuit part uses the 32-bit microcontroller stm32f103 as the control core for the corresponding circuit design,mainly including the power supply circuit,the sensor drive circuit,the microcontroller circuit,the analog-to-digital conversion circuit,the serial communication circuit and the air circuit control circuit.The software system design uses keil5 software to design the microcontroller.The host computer is based on the LabVIEW development platform to build an intelligent analysis system.It sends commands to the microcontroller through the serial port to implement process control such as system sample injection and cleaning.Build pattern recognition algorithm library module,intelligent analysis system integrates multiple pattern recognition algorithms,optimizes and selects optimal algorithms and embeds database management system for multi-sensor output data features,and achieves classification and identification of multiple samples and good man-machine Interactive.The designed software and hardware modules are assembled and debugged to complete the final platform construction and basic performance testing.This thesis takes three groups of different flavors of coffee as research objects,carries out the odor information acquisition experiment and adopts the pattern recognition technology to classify the samples.The research results show that the qualitative analysis model of the support vector machine based on the PLS feature data set can distinguished between different flavors of coffees.At the same time,the electronic nose system platform is proved to have high accuracy and practicality,and based on the optimized sensor array identification accuracy rate of 95% which is higher than the non-optimized sensor array.