| In recent years,there have been numerous bursting water pollution incidents in China.The frequency of incidents and the harm caused by them have increased year by year,which has brought hidden dangers to residents’ drinking water safety.Most of them are the problem of malodorous in water.The difficulty is not how to control and respond,but is to realize the objective and quantitative detection of odorous substances in water.And establish a new odorant evaluation system to monitor and predict odor online and target the pollutants in the first time.A technical route was developed by consulting the literature,which was,through a new sensory analysis mode,electronic nose technology,the method optimization and verification of electronic nose for water odor detection was studied.The establishment of water odor fingerprint method and the establishment of water quality evaluation method was researched at the same time.A new odor recognition and intensity rating evaluation system that was different from FPA method was established,which can avoid the uncertainty and instability caused by human subjective factors and realize the fast and online odor detection.The Portable Electronic Nose PEN3(Airsense Analytics,Schwerin,Germany)was used in this study.It is equipped with ten different heated metal oxide sensors.Through the optimization of headspace sampling conditions,a specific implementation method for improving the sensitivity of the electronic nose for the detection of water smell was proposed.The sampling volume was 200 mL and the water sample was heated in a constant temperature water bath at 65℃ for 10 minutes,while the air tightness of the device needed to be guaranteed in the whole process.After optimization,the electronic nose can distinguish the odorant and odorless water at the odor threshold concentration.The PCA method can distinguish between the two to 0.747.For the six common odorants,the total contribution of the PC A method was increased by 21.45%after optimization.For the odor concentration of a single odorant,the difference between the concentration clusters of the six odorants after optimization increased with the increase in concentration,and the low concentration clusters’ differences was enlarged.The sensitivity of the electronic nose is improved.Based on the optimized method of electronic nose for water odor detection technology,sensors with higher sensitivity were selected for the characteristic olfactory and the sensor signals were plotted on the radar chart to obtain the characteristic fingerprints of six common odorants,and the shape of the fingerprints will not change with the concentration of the substance.Each odorant had its own characteristic sensor response.For geosmin(GSM)and 2-methylisopropanol(2-MIB)represented by alcohol odorants,the characteristic sensor is No.6,8.For dimethyltrisulfide(DMTS)represented by sulfur-containing odorants,it’s No.6,7.For 2,6-nonadienal,β-cyclocitral and 2,4-heptadienal represented by aldehyde odorants,it’s No.6,8.And for β-cyclocitral and 2,4-heptadienal,the characteristic sensors were also include No.2.The multiple linear relationships between the odor intensity of six single odors and the signals of actual sensitive sensor and characteristic response sensor were established.By comparing the correlation coefficients of the two groups,the odor intensity of all odorants was more related to the signal of the characteristic response sensors except for DMTS,which showed the same correlation coefficients of the two groups.It showed that the electronic nose sensor had different characteristic responses to different odor substances and was reflected in the different response signals of each sensor.R2 of all models is greater than 0.95,which showed a linear dependence.The odor reconstruction analysis of six odorants and the prediction of their odor types were studied.The results showed that the odor types could be preliminarily predicted based on the characteristic responses of the sensors.Further calculations showed that the odors with Pearson correlation coefficients greater than 0.95 also showed a higher level in the actual compound odors.Actual water research and gas chromatography-mass spectrometry(GC-MS)detection found that the cucumber odor in the compound odor was easy to be masked,and the grassy and fishy odors had a synergistic effect.In the research of establishing water quality evaluation method based on electronic nose,we found that after the series treatment of the raw water in the water plant,suspended impurities,microorganisms,and bacteria in the water were removed.And the odor concentration of the water body was gradually reduced,which was consistent with the detection signal law of the electronic nose.Further research found that the electronic nose sensor signal has a positive correlation with TOC and turbidity.By using Minitab software to perform Partial least squares regression(PLS),the linear relationship between the electronic nose detection signal and TOC and turbidity was established.And the multivariate equation of the water quality prediction model is obtained.R2 of each water quality index were 0.831 and 0.966,respectively,showing a linear dependence,which made it possible to simultaneously reflect the multiple water quality through the electronic nose test,greatly saving the test costs. |
