- 學位論文
應用AERMOD進行臭氣擴散之研究-以北台灣某製藥廠為例
Odor Load Investigation for a Pharmaceutical Plant of North Taiwan by AERMOD
摘要
本研究運用AERMOD(American Meteorological Society / Environmental Protection Agency Regulatory Dispersion Model)模式以北台灣某製藥廠為例，對其所排放惡臭進行評估，並模擬臭氣排放路徑確認影響地區，先後利用OP-FTIR(Open Path Fourier Transform Infrared) 與GC/MS (Gas Chromatography / Mass Spectrometry detector)來監測所排放出污染物濃度。本研究於2006年以OP-FTIR選取兩次不同時間進行長期污染物監測；2007年則以GC/MS將現地採樣結果進行分析。將兩年度臭氣擴散情形以AERMOD進行模擬，並將實測與模擬結果由皮爾森公式，求出相關係數r值分別為0.9297與0.8392，兩者誤差尚於合理範圍內，是故對於臭氣濃度之預測具有良好成效。 於模式當中，臭氣衰退半衰期亦為一重要參數。因此本研究以2006年採得樣品臭氣濃度與距離進行迴歸分析，其中R 值範圍為0.6249-0.8886。將半衰期輸入模式中，能使模擬結果更貼近實際污染情況。再者擴散源所含之污染物種，會因個別濃度與閾值而影響臭氣濃度。本研究將污染物經OSL(Odor Simulation Level)轉換後，找出最具影響力之三項物種，經迴歸後可得一迴歸方程式，式中之係數本研究稱為臭氣權重係數，可作為影響臭氣濃度之判斷指標。將方程式預測與實測值以皮爾森公式運算，得出兩者相關係數r值為0.8518，故為一良好且可信之數據。綜合AERMOD、OP-FTIR及GC/MS，不僅能找出致臭物，估算受污染範圍，亦能供政府單位與製造業方面管理階層作為參考，進而防制空氣污染物擴散。
並列摘要
This research was using AERMOD (American Meteorological Society / Environmental Protection Agency Regulatory Dispersion Model) to investigate odor load for a pharmaceutical plant of North Taiwan. It was conducted to correlate the odor index and possible pollutants from a pharmaceutical plant based on the odor threshold with OP-FTIR (Open Path Fourier Transform Infrared) and GC/MS (Gas Chromatography / Mass Spectrometry detector) technique, and to model the results using AERMOD at the long-term. Two different observations pollutants were monitored with OP-FTIR in 2006. In 2007, 10 groups of pollutants were sampled on-site around the pharmaceutical plant in order to analyse them by GC/MS. And then, the dispersion route of the odorant was modeled in 2006 and 2007 to identify the influenced area. The results of regression analysis in 2006 and 2007 showed that r value is 0.9297 and 0.8392. The error of predicted result is in the rational range, so the established model is suitable to predict the odor index. In the AERMOD model, the half-life of odor also plays an important parameter. In order to correspond the simulation result to the realistic odor emissive situation, this research will discuss the half-life of odor. According to the odor index and distance of regression analysis, the range of R was between 0.6249-0.8886, and the wind speed was 4.60 (m/s) on average that day, and the half-life was 56.86-141.50 seconds. With the input of half-life information to the model, it can make the prediction much closer to actual polluted condition. Moreover, it can determine the odor index influenced by the proportion of species of pollutant source. When the relationship between them is studied, the value divided from the concentration of pollutant species and threshold then can be transformed into the OSL (Odor Simulation Level). To find three species with the strongest influence, they can be the equation after the linear return. The coefficient in the equation from three species in this research is called the weight coefficient of odor index. The larger coefficient can present which species influence odor index more. According to the specific pollutants and equations mentioned above, the odor index can be predicted. In this study, the correlation between the results of prediction and actual data after the regresstion analysis can determe that r is 0.8518 which is good for prediction the odor index. Moreover, AERMOD was used to model the odorant dispersion route in order to identify influenced areas. The influenced contour for odor index was simulated by AERMOD and has a good correlation observed between the AERMOD predictions and the actual observed values. From the combination of OP-FTIR/GC-MS/AERMOD, this research not only identifies odorant and predicts pollutant areas quickly, but also offers advices to the government and the manufacturers for improving air pollutant emission.