運用開徑式傅立葉轉換紅外光光譜儀定位大氣環境中之逸散源

Using OP-FTIR to Localize Emission Sources in Ambient Environment

陳靜慧(Ching-Hui Chen)

指導教授 ： 吳章甫

國立臺灣大學/公共衛生學院/職業醫學與工業衛生研究所/碩士(2008年)

https://doi.org/10.6342/NTU.2008.02504

摘要

本研究之目的在於利用光學遙測偵測儀器＂開放光徑式傅立葉轉換紅外光光譜儀（Open Path Fourier Transform Infrared Spectroscopy, OP-FTIR）＂收集濃度資料，再利用OP-FTIR測線上最高濃度區域結合氣象資料藉以推估逸散源之位置。本研究在一空曠的平地模擬可能發生之汙染源逸散狀況，量測氣狀汙染物濃度資料並配合風向資料藉以找出逸散源之位置。雖然近年來已有類似之研究運用一維重建結合風向資料成功的定位出大氣環境中之污染逸散源，然而，有別於以往的是，本研究將同時定位兩種不同之逸散源，N2O 和SF6。而在運用一維重建方法上除了應用高斯分布來當成基底方程式以外，還多採用了beta 和weibull basis function。並發展出不同的PIC處理過程來使得重建結果更具代表性。為了研究這些不同的基底方程式以及不同的PIC 處理步驟，我們進行了模擬研究來了解它們的應用與限制。此外，考慮到一維重建的不確定性，我們利用蒙地卡羅模擬方法來描述不確定性。
本研究利用兩種方法來取得測線上之最高濃度位置 (Peak Locations)。首先運用一維重建來取得濃度分佈資料。將OP-FTIR所偵測到之Path Integrated Concentration （PIC）資料結合最佳化演算法如Smooth Basis Functions Minimization (SBFM)方法可得到逸散源在測線上一維的濃度分佈情形。第二種方法為算出所有PIC 在各反射鏡之間的區域濃度 (PICsegment)，將測線上擁有最高PICsegment之區段當作所有可能之Peak Locations。這個新方法利用簡單的數學運算來取得最高濃度之區段，且不需複雜的最佳化演算過程就可以提供有關最高濃度區域之相關資訊。
模擬分析的結果顯示出normal basis function 應用在濃度分布其在整條測線上之下限不為0時有較好之重建結果；此外，在PIC處理方法上，若給予第一段PIC一個彈性空間而不去考慮它也可以得到較好之重建結果。而實驗的結果顯示出一維重建結合風向資料所找到之釋放源位置距離真實位置，對於N2O最近可以到4.3公尺，而對於SF6而言，最近可以到3.7公尺。第二種方法可以在尋找釋放源上提供初步但是具有意義的資訊。它所找到之釋放源位置距離真實位置，對N2O而言與SF6而言均為7.7公尺。這兩種方法所找出的逸散源估計位置均非常接近真實釋放源之位置。以經過標準化之估計位置出現頻率來看定位之結果，這兩種定位方法在出現頻率在前50%以上之估計區域都涵蓋了真實釋放源之位置。

關鍵字

定位 ； 空氣汙染 ； 逸散源 ； 開放光徑式傅立葉轉換紅外光光譜儀 ； 最佳化演算 ； 蒙地卡羅

並列摘要

An experimental field trial for tracing back two gaseous emission sources upwind was conducted to simulate the real situations of occurring leakages in the industrial estates. There are four innovative aspects in this study. First, two emission sources, N2O and SF6, were localized simultaneously in this study. Secondly, we investigated three different methods to process the PIC data and applied three basis functions, which are normal, beta and weibull basis functions, to reconstruct 1-D profiles with a set of PIC data. Thirdly, the uncertainties of 1-D profile reconstructions were considered in this study. Fourthly, we introduced a new approach by applying Monte Carlo method to predict the peak locations along the measurement line without reconstruction procedures.
Combining the peak locations of air pollutants concentrations along a measurement line downwind of monitoring area with wind directions can trace back the upwind emission sources. Two approaches are introduced to find the peak locations in this thesis. First approach is to reconstruct the 1-D profiles along the measurement line to obtain the reconstructed peak locations. Second is to find the possible peak locations within the region of highest segmented PIC by Monte Carlo method along the measurement line. In order to investigate the applications and limitations of the 1-D profile reconstructions, the simulation analysis was also conducted.
In simulation analysis, when the plumes are distributed broadly, and not congregated near the origin of the whole dispersion plume, RECON3bm, a PIC processing method which proceeded with the original generated PIC without the information of PIC1, provided the better results than other processing methods. Due to the flexibility of the lower limit in normal basis function, when the beginning of the testing plume is not at the origin, applying a normal basis function can provide better results than the others. In the field study, the first localization approach can successfully localize both of N2O and SF6. The smallest distances between real source locations were 4.3 meters for N2O and 3.7 meters for SF6. The second approach can provide the preliminary but significant indications of tracing back the emission sources. The distances from this approach were 7.7 meters for N2O and SF6. These two localization approaches can provide good estimations of source locations. The real source locations were properly covered by the estimated area whose standardized frequencies are higher than 50%.

並列關鍵字

source localization ； air pollutant ； emission sources ； OP-FTIR ； optimization algorithm ； Monte Carlo

參考文獻

5. 環境檢測方法 -空氣及物理. 行政院環保署環境檢驗所.

4. Clean Air Act. In.: U.S. EPA; 1990.

7. Compendium Method TO-16: Long-Path Open-Path Fourier Transform Infrared Monitoring of Atmospheric Gases, 2nd edition. Edited by U.S. E: U.S.,EPA; 1999.

13. Hashmonay RA, Yost MG: Innovative Approach for Estimating Fugitive Gaseous Fluxes Using Computed Tomography and Remote Optical Sensing Techniques. Journal of the Air & Waste Management Association 1999, 49(8):966-972.

14. Hashmonay RA, Natschke DF, Wagoner K, Harris DB, Thompson EL, Yost MG: Field Evaluation of a Method for Estimating Gaseous Fluxes from Area Sources Using Open Path Fourier Transform Infrared. Environmental Science & Technology 2001, 35(11):2309-2313.