The Data Analysis And Treatment Of Malodorous Gas Monitoring System

All along,the rapid development of the world economy,always accompanied by environmental pollution problems,so the problem of malodorous pollution gradually attention.The initial malodorous pollution is mainly characterized by human sense of smell and subjective sensation.For a single malodorous gas,the smell of a single malodorous gas can be obtained by the smell of a single odor of different concentrations of malodorous gas.But the relationship between the malodorous value and its concentration is very complex for the compound malodorous gas of unknown composition,and the relationship between the concentration and the malodor value can not be obtained by artificial sniffing.For the unknown malodorous gas,only by artificial sniffing to get its odorous value,the detection of malodorous gas is always a lot of manpower and material resources.The continuous improvement of the level of science and technology,artificial sniffing method has been gradually replaced by electronic nose odor method.Because the traditional method of artificial sniffing and chemical composition analysis can not detect malodor in real time continuously,it is very important to establish the relationship between malodor and artificial sniffing in real-time monitoring system based on multi-gas sensor.This paper presents a method based on sniffing of complex malodorous gas electronic nose detection.The core of the method is to use the sensor array for the composite odor gas multi-sampling,the formation of response fingerprints and artificial sniffing results contrast calibration,the measurement of malodorous gases.As the gas sensor and the characteristics of ordinary sensors there is a big difference,more prominent is a serious lag.Due to the presence of hysteresis,in the case of on-line detection,most of the sensor response can not reach the steady-state value.These hysteresis include both linear and non-linear.For the linear lag,a first order linear hysteresis model is established,and the best estimate of the steady-state value is realized by using the intermediate data of the transient process.The experimental results show that the model has high accuracy.Aiming at the nonlinear lag,this paper also proposes an algorithm based on shift regression,which can effectively eliminate the nonlinear hysteresis and has high adaptability.In addition,the shift regression algorithm is also optimized.This method is not only suitable for electronic nose system,but also for a variety of gas sensor signal detection applications.In this paper,the correctness of the gas sensor model is verified by a large number of experimental analysis and processing.At the same time,the optimization of the shift regression algorithm is described in detail.Finally,the database of malodorous fingerprint recognition is established by experiments of monogas gas and compound malodorous gas.Completed the work of malodorous gas detection.