Development Of Multi-sensor System For Odor Detection

As a form of air pollution,odorous gas directly harms people’s health and mental state through the human respiratory system.The treatment of odor pollution has attracted the attention of the whole society.At present,the general odor detection method is subjective detection by the traditional artificial smell identification,and the objective evaluation of the odor intensity level is not achieved by the odor detecting instrument.With the improvement of scientific law enforcement requirements in environmental monitoring,as well as the need for rapid and networked testing instruments,higher requirements are imposed on odor detection.Therefore,it is necessary to develop an objective detection instrument that meets the human sensory evaluation and to realize the platformization of the detection data.In this paper,the development trend of odorous gas detection methods and electronic nose systems at home and abroad is analyzed,and a odor detection system based on sensor arrays combined with the detection requirements of odorous gase is developed.By analyzing the detection principle and performance characteristics of different types of gas sensors,the electrochemical sensor,photoionization sensor and metal oxide sensor are selected to form the sensor array,and based on the system requirements,the development of power supply circuit and drive circuit and signal conditioning circuit is completed.A metal oxide sensor chamber is designed according to the characteristics of gas flow direction and sensor injection mode.The pumping serial connection gas path is adopted,and the gas circuit switching and gas injection are realized by the solenoid valve and the diaphragm pump.According to the characteristics of the sensor array output data and the requirements of the lower computer control system,the data packet format in data communication is designed.The server software obtains the response data of the sensor array by establishing a TCP/IP long connection with the 4G transparent transmission module,and the related data is stored in the MySQL database to achieve real-time upload of test data.The client software is responsible for the display,saving,processing,and user control of test data.Aiming at the common cross-sensitivity problem of gas sensors and the requirement of sensory intensity evaluation in odor detection,a odor intensity model algorithm suitable for multi-sensor systems is proposed.The odor model is divided into two parts: concentration analysis and intensity calculation.The differential algorithm is used to extract the eigenvalues of the sensor array.The information fusion of multi-sensor eigenvalues is completed by partial least squares method.Based on this,the regression analysis algorithm of sample gas concentration is designed.Through the analysis of the odor characteristics and the common odor intensity model,the malodor intensity model analysis method combining sensory evaluation with the odor concentration-strength relationship model is determined.Finally,the performance test and experimental analysis of the developed multi-sensor system are carried out.The experimental results show that the maximum standard deviation of the output signals of each sensor in the sensor array is 0.0031 V,and the maximum relative standard deviation is 1.10%,which has good stability and repeatability.Taking the mixed gas of hydrogen sulfide and carbon disulfide as the sample gas,the algorithm model of the relationship between the eigenvalues of the sensor array and the odor intensity level is established.The predicted value of the detection system is compared with the artificial sniffer identification result,the correlation coefficient is 0.986,which has higher prediction accuracy and meets the requirements of the detection system.