氣相層析自動化控制與積分演算法應用於空氣中有機污染物連續監測

NCU Institutional Repository > 理學院 > 化學研究所 > 博碩士論文 > Item 987654321/76689

請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/76689

題名:	氣相層析自動化控制與積分演算法應用於空氣中有機污染物連續監測
作者:	龎文韶;Pang, Wen-Shao
貢獻者:	化學學系
關鍵詞:	自動控制;層析積分演算法;訊號處理;層析峰模型;非甲烷總碳氫;Automation control;Chromatography peak detection algorithm;Data processing;Peak Model;Non-methane hydrocarbons(tNMHCs)
日期:	2018-07-26
上傳時間:	2018-08-31 11:33:23 (UTC+8)
出版者:	國立中央大學
摘要:	本研究目的是建立一套自動化軟體系統, 能夠針對空氣中揮發性有機化合物 (Volatile Organic Compounds, VOCs)進行連續監測。此系統包含兩大部分，分別為氣相層析儀的程序自動控制與層析訊號處理。在程序自動控制部分，系統包含了氣體閥門之控制、訊號擷取、溫度控制，此外亦建立了詢問型的程序控制架構，在此架構下，軟體可以進行四種不同的執行模式，分別為周界模式 (Ambient Mode)、標準氣體模式 (Cylinder Mode)、空白模式 (Blank Mode)、客製化模式 (Custom Mode)，模式間可以任意的切換，並且可以使這四種模式照使用者需求進行排列並執行序列模式 (Sequence Mode)，達到連續監測時的自動化校準與清理之目的。在層析訊號處理方面，本研究建立了兩套層析積分演算法，斜率法與模型法，前者主要以層析圖譜之微分值作為判斷層析峰位置與範圍之依據，並且針對層析峰之範圍進行面積積分，而對於共析層析峰則是以垂直切割 (Drop)的方式進行解析，後者模型法則以非線性曲線擬合法，並套用層析峰模型 (Peak Mode)以模擬層析峰形，最後根據所解析的層析峰進行面積積分。為驗證演算法之成效，本研究以環保署NIEA A505.12B方法之標準氣體進行測試，此氣體內含54種碳數介於C2 – C12 之VOCs。測試結果顯示在上述兩種演算法下有90%以上之物種其RSD小於1 %，而約90%之物種之R2值亦優於0.999，顯示本研究之演算法有良好的再線性與定量能力，並且具備接近商業化軟體之成效。我們亦將此軟體系統實際應用於非甲烷總碳氫分析(Total Non-methane Hydrocarbons, tNMHCs ) 連續監測上，使用本實驗室所設計開發之火焰離子 (FID) 偵測式總碳氫分析儀 (THC)。測試結果顯示該分析儀於濃度1.2 - 72 ppm間的甲烷與總碳氫 (Total Hydrocarbons, THC)之檢量線的R2皆大於0.995，此外在標準氣體的連續測試上亦成功驗證了此軟體控制系統在運作時的穩定性。 ;The objective of this research is to develop a software solution to automate a chromatographic system to monitor ambient volatile organic compounds (VOCs). This software solution consists of two components, the process automation and the data processing algorithm. In the process automation part, functionalities of gas-valve control, signal acquisition, and temperature control were designed. An inquiry-type architecture was constructed for process automation. Under this architecture, the software can perform four different execution modes and can arbitrarily switched between the four modes. These modes include Ambient Mode, Cylinder Mode, Blank Mode and Custom mode. Moreover, the modes can be arranged to execute in sequence according to the user’s needs, with an aim to perform both the concentration calibration and system cleaning in continuous cycles. For the data processing algorithm, two methods were developed to integrate chromatographic peaks, i.e., the slope method and the model method. The slope method uses differentials of the chromatographic signals as the basis for detecting the occurrence and the range of a chromatograph peak for area integration. In addition, co-eluted chromatographic peaks can be processed by the so called “drop-cut” method in the slope method. For the model method, it uses the nonlinear curve fitting method to simulate chromatographic peaks. Subsequently, peak area integration is performed on the modeled chromatographic peaks. To assess the effectiveness of the algorithm, the standard mixture of an EPA method NIEA E505.12B containing 54 VOCs from C2 – C12 was analyzed. Our results show that more than 90% of the species are better than 1% RSD, and about 90% of species show linearity (R2) greater than 0.999. As a result, both precision and accuracy can be demonstrated by our developed data processing algorithm which is close to the performance of a commercial software as the benchmark. When applied the software package to the analyzer of total hydrocarbons (THC) with flame ionization detection (FID) previously developed by our laboratory to analyze total non-methane hydrocarbons (tNMHCs). We found that the R2 values for methane and THC were greater than 0.995 between 1.2 - 72 ppm in concentration. Moreover, the analyzer equipped with the software was able to show long-term stability by continuous analyzing a standard gas.
顯示於類別:	[化學研究所] 博碩士論文

文件中的檔案:

檔案	描述	大小	格式	瀏覽次數
index.html		0Kb	HTML	222	檢視/開啟

在NCUIR中所有的資料項目都受到原著作權保護.

社群 sharing

資料載入中.....