以雙測線之一維輻射光徑煙流分布重建法定位污染源

Source Localization with One-dimensional Radial Plume Mapping Technique Using Two Monitoring Lines

林士鈞(Shih-Chun Lin)

指導教授 ： 吳章甫

國立臺灣大學/公共衛生學院/環境衛生研究所/碩士(2010年)

https://doi.org/10.6342/NTU.2010.01361

摘要

本研究之目的是使用光學遙測偵測儀器－開徑式傅立葉轉換紅外光光譜儀（Open Path Fourier Transform Infrared Spectroscopy, OP-FTIR）結合一維輻射光徑煙流分布重建法 (1D-RPM) 來推估地面污染源之位置（污染源高度接近於地面）。本研究使用Smooth Basis Function Minimization (SBFM)重建法來重建下風處濃度最高的位置，再結合濃度最高的位置與對應的風向資料回推污染源之位置。此研究的獨特性為藉由建立兩條互相垂直的測線取代先前研究中所使用的單一下風處測線，來減少推估污染源的不確定性及縮短推估釋放源的位置與真實位置的距離。
在真實的情況下，污染源的位置是未知的，所以在電腦模擬部分，我們模擬了在200公尺乘以200公尺的區域中所有可能的污染源位置，來比較使用兩條1D-RPM測線與先前研究的差異。我們以估計位置出現頻率來評估定位之結果，使用兩條1D-RPM測線除了在推估污染源的準確性有所提升以外，可準確定位污染源的範圍也大幅增加。此研究也發現下風處的三道光徑中，濃度最高位置的比例不同可能會影響推估的準確性。
在戶外實地實驗中，使用OP-FTIR結合兩條1D-RPM測線來定位人為釋放的污染源（追蹤氣體）。實驗的結果與電腦模擬相似，使用兩條1D-RPM測線縮短了推估釋放源位置與真實位置的距離，回推的不確定性也較低。推估釋放源位置與真實位置的距離小於40公尺。

關鍵字

污染物定位 ； 開徑式傅立葉轉換紅外光光譜儀 ； 一維輻射光徑煙流分布重建法 ； 光學遙測 ； 逸散源 ； 空氣污染

並列摘要

This paper reports the computational simulation and field experimental results by using the OP-FTIR spectroscopy with the 1D-RPM technology to locate the ground-level emission sources. The objective of this study is to improve the 1D-RPM technology by adding one additional monitoring line to reduce the uncertainty of source location results and the distance between the estimated source and the real source location.
The smooth basis function minimization (SBFM) algorithm was used to reconstruct the peak location of the plume. Combining the peak location of downwind concentrations along the monitoring line and the wind data gives the estimated source locations.
In the computational simulation study, we propose using the concept of “error map” to indicate the performance and uncertainty of the two-line 1D-RPM. The error map shows that using two-line 1D-RPM can locate the sources more accurately, and cover wider area than using one monitoring line only. We also found the locating result varied as the proportion of the reconstructed peak locations.
In the field experiment, the distance between the region having highest density of estimated of source location and the real source location (DIST) and the uncertainty area are smaller by using two-line 1D-RPM than original one line 1D-RPM. The DIST is smaller than 40 meter (20 percent of the size of the experimental domain) by using two-line 1D-RPM.

並列關鍵字

source localization ； OP-FTIR ； 1D-RPM ； optical remote sensing ； emission ； air pollutant

參考文獻

2. Chen, C.L. Source location and characterization of volatile organic compound

5. Zhao, W. Source identification of volatile organic compounds in Houston, Texas;

Environmental Science and Technology. 2004. 38(5): p. 1338.

Source Localization: Evaluation of Reconstruction Algorithms; National Taiwan University, Master thesis. 2008.

7. Thomson, L.C. An improved algorithm for locating a gas source using inverse