【摘要】：為實(shí)現(xiàn)對揮發(fā)性有機(jī)化合物(VOC)混合氣體檢測,本研究采用傳感器陣列和模式識別技術(shù)相結(jié)合的電子鼻方式對問題進(jìn)行探究。傳感器陣列是在實(shí)驗(yàn)室中自制的旁熱式金屬氧化物半導(dǎo)體傳感器組成的陣列,能對VOC混合氣體形成完備的響應(yīng)模式。傳感器陣列在不同VOC混合氣體響應(yīng)數(shù)據(jù)集均來源于實(shí)際實(shí)驗(yàn)測試。實(shí)驗(yàn)中為其問題的實(shí)驗(yàn)探究,自行搭建了傳感器測試系統(tǒng),其中的VOC混合氣體由乙醇,丙酮,甲醛和甲苯四種典型VOC氣體組成的目標(biāo)分析物,并且為了開發(fā)用于實(shí)際應(yīng)用的電子鼻,每種VOC及其組合的濃度在混合物中隨機(jī)分配。探究模式識別時,本文采取BP神經(jīng)網(wǎng)絡(luò)對傳感器陣列信號進(jìn)行分析和識別,完成對VOC混合氣體組分及濃度估計(jì)。BP神經(jīng)網(wǎng)絡(luò)在MATLAB中分析創(chuàng)建,首先須要做的是對數(shù)據(jù)歸一化預(yù)處理,以防止因數(shù)量級別而造成計(jì)量誤差,然后還探究了BP神經(jīng)網(wǎng)絡(luò)中隱含層神經(jīng)元個數(shù),激活函數(shù),性能目標(biāo)等結(jié)構(gòu)參數(shù)對網(wǎng)絡(luò)預(yù)測性能的影響,調(diào)試出適合本問題的最佳結(jié)構(gòu)參數(shù)。根據(jù)實(shí)驗(yàn)結(jié)果可知,BP神經(jīng)網(wǎng)絡(luò)輸出節(jié)點(diǎn)對目標(biāo)分析物中每種VOC的濃度可以給出連續(xù)預(yù)測,且在一定誤差范圍內(nèi),它是能夠完成對VOC混合氣體組分的定量分析。本文以提高系統(tǒng)預(yù)測精度為目的,對模式識別中方式進(jìn)行改進(jìn),思慮先由決策樹以VOC總量為標(biāo)準(zhǔn)對VOC混合物數(shù)據(jù)集進(jìn)行等級分類完成初步篩選,再由BP神經(jīng)網(wǎng)絡(luò)基于不同等級,調(diào)試出適當(dāng)?shù)慕Y(jié)構(gòu)參數(shù),并對等級內(nèi)的樣本針對性訓(xùn)練,完成濃度估計(jì)。實(shí)驗(yàn)結(jié)果顯示,改進(jìn)后的模式在每種VOC的濃度估計(jì)中的最大誤差約為2ppm,其精度優(yōu)于從單個BP神經(jīng)網(wǎng)絡(luò)獲得的結(jié)果。此外,當(dāng)預(yù)測的濃度高于20ppm時,相對誤差小于5%。這項(xiàng)研究表明:神經(jīng)網(wǎng)絡(luò)定量分析VOC混合物濃度方面的潛力。改進(jìn)后的模式能較準(zhǔn)確地完成VOC混合氣體組分的定量分析,成為開發(fā)針對VOC氣體進(jìn)行識別的電子鼻產(chǎn)品的基礎(chǔ)。
[Abstract]:In order to detect the mixed gases of volatile organic compounds (VOC), an electronic nose combined with sensor array and pattern recognition was used to study the problem. Sensor array is an array of side heat metal oxide semiconductor sensors made by ourselves in laboratory. It can form a complete response mode to VOC mixed gas. Sensor arrays in different VOC gas mixture response data sets are derived from the actual experimental test. In order to explore the problem in the experiment, a sensor testing system was set up, in which the VOC mixture gas was composed of four typical VOC gases, ethanol, acetone, formaldehyde and toluene. In order to develop an electronic nose for practical applications, the concentrations of each VOC and its combinations are randomly distributed in the mixture. In this paper, BP neural network is used to analyze and recognize the sensor array signals, and the mixture gas components and concentrations of VOC are estimated. BP neural network is established in MATLAB. The first thing we need to do is to preprocess the data normalization so as to prevent the metrological error caused by the quantity level. Then we also explore the number of neurons in the hidden layer and the activation function in the BP neural network. The effect of performance target and other structural parameters on the network prediction performance is studied and the optimal structure parameters suitable for this problem are debugged. According to the experimental results, the output node of BP neural network can give a continuous prediction of the concentration of each VOC in the target analyte, and within a certain error range, it can accomplish the quantitative analysis of the VOC mixture gas component. In order to improve the prediction accuracy of the system, the method of pattern recognition is improved in this paper. First, the decision tree classifies the VOC mixture data set according to the total amount of VOC, and then the BP neural network is based on different grades. The appropriate structural parameters were debugged and the samples in the grade were trained to estimate the concentration. The experimental results show that the maximum error of the improved model in each VOC concentration estimation is about 2 ppm.The accuracy of the improved model is better than that obtained from a single BP neural network. In addition, when the predicted concentration is higher than 20ppm, the relative error is less than 5. This study shows the potential of neural networks for quantitative analysis of VOC mixture concentrations. The improved model can accurately accomplish the quantitative analysis of VOC mixture gas components, which is the basis of developing electronic nose products for VOC gas recognition.
【學(xué)位授予單位】：寧波大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP212.9;TP183

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 Fengye Hu;Lu Wang;Shanshan Wang;Xiaolan Liu;Gengxin He;;A Human Body Posture Recognition Algorithm Based on BP Neural Network for Wireless Body Area Networks[J];中國通信;2016年08期

2 吳君章;趙盛翹;鄒小勇;賴燕華;;電子鼻在煙草行業(yè)中的研究與應(yīng)用進(jìn)展[J];分析測試學(xué)報(bào);2014年07期

3 路楊;李鵬珊;翟盼盼;;改進(jìn)BP神經(jīng)網(wǎng)絡(luò)在木構(gòu)古建筑中的壽命預(yù)測[J];計(jì)算機(jī)技術(shù)與發(fā)展;2014年05期

4 楊莉;;淺談目前的室內(nèi)空氣環(huán)境監(jiān)測[J];科技創(chuàng)新導(dǎo)報(bào);2013年34期

5 鄭秀亮;;VOC,不可忽視的環(huán)境污染因子[J];環(huán)境;2012年05期

6 王坤;;神經(jīng)網(wǎng)絡(luò)的特點(diǎn)及改進(jìn)方法綜述[J];科技廣場;2011年07期

7 馬宏偉;陳小通;祁昌禹;張紅霞;李工農(nóng);韓根亮;;氣體傳感器靜態(tài)測試系統(tǒng)電路設(shè)計(jì)[J];甘肅科學(xué)學(xué)報(bào);2010年04期

8 黃為勇;任子暉;童敏明;;多氣體的SVM數(shù)據(jù)融合定性識別方法[J];計(jì)算機(jī)工程與應(yīng)用;2009年09期

9 竺志大;王昌龍;;SnO_2基薄膜氣體傳感器制作與敏感性測試[J];機(jī)械制造;2008年12期

10 任先武;徐凌;周衛(wèi)宏;王元委;王振強(qiáng);;一種抗氫氣干擾的CO傳感器[J];傳感器與微系統(tǒng);2008年08期

相關(guān)博士學(xué)位論文 前2條

1 張紅梅;基于氣體傳感器陣列的幾種農(nóng)產(chǎn)品品質(zhì)檢測研究[D];浙江大學(xué);2007年

2 張覃軼;電子鼻：傳感器陣列、系統(tǒng)及應(yīng)用研究[D];華中科技大學(xué);2005年

相關(guān)碩士學(xué)位論文 前10條

1 談力;基于相似日選取的小波極限學(xué)習(xí)機(jī)短期負(fù)荷預(yù)測模型研究[D];南京理工大學(xué);2015年

2 柳潤琴;基于氣體傳感器陣列的有害氣體檢測系統(tǒng)的研究[D];寧波大學(xué);2014年

3 王愛霞;基于神經(jīng)網(wǎng)絡(luò)的微網(wǎng)逆變器控制策略研究[D];蘭州理工大學(xué);2014年

4 王瑋;惡臭氣體檢測裝置多傳感器信號分析與處理[D];河北工業(yè)大學(xué);2014年

5 朱濤;基于“微井”結(jié)構(gòu)的氣體傳感器的制備[D];電子科技大學(xué);2013年

6 李瑛;基于氣體傳感器陣列的人工神經(jīng)網(wǎng)絡(luò)算法的設(shè)計(jì)及C語言實(shí)現(xiàn)[D];電子科技大學(xué);2012年

7 劉雪瑩;基于神經(jīng)網(wǎng)絡(luò)的混合氣體檢測系統(tǒng)研究[D];中國科學(xué)技術(shù)大學(xué);2011年

8 顧磊磊;ZnO納米結(jié)構(gòu)氣敏傳感器[D];復(fù)旦大學(xué);2011年

9 李華曜;電子鼻硬件系統(tǒng)及其評價(jià)[D];華中科技大學(xué);2008年

10 許杰;低阻低溫的ZrO_2基底TMA傳感器的研究和制作[D];華東師范大學(xué);2007年

，

本文編號：2376616