- 學位論文
建立室內空氣生物氣膠濃度之可行性探討-以辦公大樓與醫療場所為例
Feasibility Study on the Development of an Empirical Prediction Model of Indoor Bio-aerosol Concentration - In Office Buildings and Hospitals
摘要
台灣地處高溫高濕的特殊地理環境中，相較於其他歐美國家，提供了微生物滋生之溫床。相關病態大樓症候群 (Sick building syndrome, SBS) 及過敏 ( Allergy )疾病患者逐年增加，間接反應室內環境對人體健康危害。目前生物氣膠檢測程序繁瑣、耗時，且檢測成本高。故本研究發展利用室內空氣生物氣膠濃度推估模式取代需花費時間與金錢之傳統生物氣膠檢測方法。本研究以辦公大樓與醫療場所作為研究之室內空氣生物氣膠濃度預測場所，本研究將各項檢測結果與文獻蒐集之室內空氣品質調查資料，以複迴歸中F檢定分析室內空氣生物氣膠濃度與室內外空氣各測項之相關性，並利用分析所得之相關因子建立室內空氣生物氣膠濃度推估模式。室內空氣生物氣膠推估模式之準確度以平均絕對百分誤差 ( Mean Absolute Percentage Error, MAPE) 表示之。辦公大樓室內空氣細菌濃度之模式推估結果顯示，於不同區域範圍模式推估濃度值與實測濃度值之MAPE為16~105 %；採用單一獨棟建築物之複迴歸線性模式，其MAPE驗證結果較佳，MAPE為16 % (2~36 %)；室內空氣真菌濃度之模式推估結果顯示，於不同區域範圍模式推估濃度值與實測濃度值之MAPE為2~108 %，以單一獨棟建築物之複迴歸線性模式，其MAPE驗證結果較佳，MAPE為2 % (0~5 %)。此外，第1類場所之醫療場所室內空氣細菌濃度之模式推估結果顯示，於不同區域範圍模式推估濃度值與實測濃度值之MAPE為29~59 %；以單一獨棟建築物（複迴歸線性模式與線性規劃模式無建立單一獨棟建築物之推估模式）之指數模式推估結果，其準確度較佳，MAPE為29 % (0~84 %)；室內空氣真菌濃度之模式推估結果顯示，於不同區域範圍模式推估濃度值與實測濃度值之MAPE為1~69 %；以單一獨棟建築物之複迴歸線性模式，其準確度結果較佳，MAPE為1 % (0~3 %)。研究結果顯示，辦公大樓與醫療場所之室內空氣細菌及真菌濃度模式，準確度皆以單一獨棟建築物較佳。 本研究結果可提供辦公大樓與醫療場所，進行場所自主管理時之即時室內空氣細菌及真菌濃度預測，並作為空氣品質改善前後成效評估之工具。
並列摘要
Taiwan is located in a special geographical environment with its humidity and temperature not only higher than those in most American-European countries but also more compatible with the growth of microorganisms. This particular environmental feature has triggered a steady increase in the number of the patients suffering from SBS (Sick Building Syndrome) related diseases and allergies every year in Taiwan and it indicates directly the damages of indoor environment upon human health. At present, measurement of bio-aerosol can be complicated and both time- and cost-consuming. Our study accordingly strives to develop an empirical prediction model to measure indoor bio-aerosol concentrations as a replacement of the traditional bio-aerosol concentration test for saving both time and cost. The study uses portable direct-reading devices to measure indoor air quality of office buildings and hospitals and analyze the collected data. Multiple Regression analysis is adopted to examine the indoor and outdoor measurements as well as the indoor/outdoor ratio so as to identify the variables influencing indoor bio-aerosol concentrations. The variables are then used to develop the empirical model for predicting indoor bio-aerosol concentrations. Accuracy of the model is represented by MAPE (Mean Absolute Percentage Error), and the prediction results indicate that the developed empirical model fails to achieve a satisfactory accuracy in calculating the concentration of indoor bio-aerosol based on the combined data of office buildings and hospitals. In the case of office buildings, the MAPE of the empirical model falls in the range between 58~62 % of Indoor Bacteria Concentration; the MAPE appeared to be 16 % (2~36 %) of the Multiple Regression analysis empirical model of Indoor Bacteria Concentration in a single office building; Moreover, the MAPE of the empirical model falls in the range between 2~208 % of Indoor Fungi Concentration in office buildings. The MAPE appeared to be 2 % (0~5 %) of the Multiple Regression analysis empirical model of Indoor Fungi Concentration in a single office building. In the case of hospitals, the MAPE of the empirical model falls in the range between 29~59 % of Indoor Bacteria Concentration; the MAPE appeared to be 29 % (0~84 %) of the exponent empirical model of Indoor Bacteria Concentration in a single hospital. The MAPE of the empirical model falls in the range between 1~69 % of Indoor Fungi Concentration in hospitals; the MAPE appeared to be 1 % (0~3 %) of the Multiple Regression analysis empirical model of Indoor Fungi Concentration in a single hospital. Based on the analysis results with the application of the empirical prediction model, the degree of accuracy in calculating the concentration of indoor air bacteria of in a single office building and hospital is better than all the hospitals of Taiwan and all the offices of Taipei city and county. The empirical prediction model proposed by in this research can be applied to perform instant calculation of the indoor bio-aerosol concentration in office buildings and hospitals and to serve as an assessment tool to confirm the improvement in indoor air quality.