Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/33210
標題: 揚塵量(PM10)預測之研究 ─ 以線西測站為例 ─
A Study on the Suspended Fine Particulate (PM10) of Fugitive Dust Forecasts - A Case Study of Siansi Monitoring Station -
作者: 古緯中
Gu, Wei-Jhong
關鍵字: 倒傳遞類神經網路;Back-propagation neural network;揚塵量;PM10;最佳輸入變數組合;Fugitive dust;Particulate matter;The optimum input variables
出版社: 水土保持學系所
引用: 1.王偉(1995)「人工神經網絡原理-入門與應用」,北京航空航天大學出版社。 2.王炎秋(2006)「應用類神經網路及時間序列預測台南地區空氣品質- PM10」,立德管理學院資源環境研究所碩士論文。 3.江世民(1999)「台北地區臭氧及懸浮微粒型態與其他污染物及氣象條件之分析比較」,第十六屆空氣污染控制技術研討會。 4.朱錦紅、柯文雄、許晃雄(2002)「臺灣降水變化研究回顧:氣候分區、降水與東亞大氣環流之關聯」,全球變遷通訊雜誌,第三十六期。 5.江勝偉(2007)「氣象因子與大氣懸浮微粒含碳量變異關聯性研究」,國立成功大學環境工程學系碩士論文。 6.林斌龍、李崇德(1990)「台北地區PM10 濃度與氣象因子之多變量統計分析」,第七屆空氣污染控制技術研討會。 7.林昇甫、洪成安(2002)「神經網路入門與圖樣辨識」,全華科技圖書股份有限公司,修訂二版。 8.林逸塵(2002)「類神經網路應用於空氣品質預測之研究」,國立中山大學環境工程研究所碩士論文。 9.周鵬程(2008)「類神經網路入門-活用Matlab」,全華圖書股份有限公司,修訂三版。 10.林孟希(2010)「大氣粒狀物粒徑分佈影響因子分析研究」,國立高雄第一科技大學環境與安全衛生工程系碩士論文。 11.高全興(1997)「類神經網路於空氣品質短期預測之研究」,國立雲林科技大學環境與安全工程研究所碩士論文。 12.莊秉潔(1993)「地表層擴散係數、風速、雨量及混合層高對台北都會早上八時懸浮微粒濃度之影響」,中國環境工程學刊,第三卷,第二期。 13.莊秉潔、許家彰、陳建隆(2001)「懸浮微粒濃度與微氣象因子之關係探討台北案例研究」,第十八屆空氣污染控制技術研討會。 14.陳佩儀(2005)「亞洲沙塵暴影響南台灣雲林地區粒子特性與PAHs分佈特徵之研究」,國立雲林科技大學環境與安全衛生工程系碩士論文。 15.郭石柱(2009)「揚塵懸浮微粒PM10與氣象因子相關性分析-以台東縣關山鎮空氣品質監測站為例」,國立中興大學水土保持學系所碩士論文。 16.游繁結(1989)「台灣海岸的防風定砂」,水土保持專輯,0230。 17.程萬里、黃怡嘉(2000)「中部地區臭氧及懸浮微粒與氣象因子之相關研究」,第十七屆空氣污染控制技術研討會。 18.靳蕃、範俊波、譚永東(1992)「神經網路與神經計算機」,儒林出版社。 19.詹俊南(1996)「台灣地區PM10汙染特性分析」,國立台灣大學環境工程研究所碩士論文。 20.葉怡成(2002)「應用類神經網路」,儒林圖書公司,修訂三版。 21.葉怡成(2003)「類神經網路模式應用與實作」,儒林圖書公司,修訂八版。 22.鄭淑燕、望熙榮、鄭曼婷(1991)「台中地區空氣品質分析」,第八屆空氣污染控制技術研討會。 23.談珮華、周佳、梁靜宜、吳柏霖(2008)「台北都會區的假期效應」,大氣科學,第三十六期第三號,第197~215頁。 24.蔡盛行(2011)「倒傳遞類神經網路於懸浮微粒量預測與防制工法之研究」,國立嘉義大學土木與水資源工程學系研究所碩士論文。 25.蘇志朋(2010)「以倒傳遞類神經網路及多元線性迴歸模擬建築工地對台中市粒狀污染物之影響」,朝陽科技大學環境工程與管理系碩士論文。 26.行政院環境保護署空氣品質監測網:http://taqm.epa.gov.tw 27.行政院環境保護署中國大陸沙塵監測網:http://dust.epa.gov.tw 28.Acuna, G., H. Jorquera and R. Perez. (1996) “Neural network model for maximum ozone concentration prediction.” Lecture Notes in Computer Science, No.1112, pp.263-268. 29.Blanco Humberto and Rattan Lal. (2008) “Principles of Soil Conservation and Management.” Springer Science: 55-80. 30.Chu, S. H. (1997) “Meteorological Conditions Conducive to Regional High Particulate Matter Episodes.” Air and Waste Management Association’s 90th Annual Meeting and Exhibition, June 8-13, Toronto, Ontario, Canada 97-MP112.06. 31.Comrie, A.C. (1997) “Comparing neural networks and regression models for ozone forecasting.” Journal of Air and Waste Management, Vol.47, pp.653-663. 32.Huang, M., G.B. Peng, J.S. Zhang and S.H. Zhang. (2006) “Application of artificial neural networks to the prediction of dust storms in Northwest China.” Global and Planetary Change, Volume 52, Issue 1-4, 216-224. 33.Jorquera, H., R. Perez, A. Cipriano, A. Espejo, M. V. Letelier, and G. Acuna. (1998) “Forecasting Ozone Daily Maximum Levels at Santiago, Chile.” Atmospheric Environment, 32(20): 3415-3424. 34.Lam, K.C. and S. Cheng. (1998) “A synoptic climatological approach to forecast concentrations of sulfur dioxide and nitrogen oxides in Hong Kong.” Environmental Pollution, Vol.101, pp.l83-191. 35.Lam, K.C. and S. Cheng. (2000) “Synoptic typing and its application to the assessment of climatic impact on concentrations of sulfur dioxide and nitrogen oxides in Hong Kong.” Atmospheric Environment, Vol.34, pp.585-594. 36.Monn, C.H., O. Braendli, G. Schaeppi, CH. Schindler, U. Ackermann, PH. Leuenberger and T. Sapldia. (1995) “Particulate matter < 10 μm (PM10) and total suspended particulates (TSP) in urban, rural and alpine air in Switzerland.” Atmospheric Environment, Vol.30, Issue 9, 1529. 37.Nash, J.E. and J.V. Stucliffe (1970) “River flow forecasting through conceptual models part I - A discussion of principles.” Journal of Hydrology, 10(3), 282-290 38.Perez, P., A. Trier and J. Reyes. (2000) “Prediction of PM2.5 Concentrations Several Hours in Advance Using Neural Networks in Santiago,Chile.” Atm. 39.Perez, P. and J. Reyes. (2002) “Prediction of Maximum of 24-h Average of PM10 Concentration 30 h in Advance in Santiago, Chile.” Atmospheric Environment 36. 4555-4561. 40.Prybutok, V. R., J. Yi and D. Mitchell. (2000) “Comparison of neural network models with ARIMA and regression models for prediction of Houston''s daily maximum ozone concentrations.” European Journal of Operational Research, Vol.122, Issue 1, pp.31-40. 41.Sohn, Sang Hyum, Oh, Sea Cheon, Jo, Byung Wan and Yeo Yeong-Koo. (2000) “Prediction of Ozone Formation Based on Neural Network.” Journal of Envirorunental Engineering, pp688-696. 42.Smith, Steve, F. Trevor Stribley, Peter Milligan and Ben Barratt. (2001) “Factors influencing measurements of PM10 during 1995-1997 in London.” Atmospheric Environment, Vol.35, 4651-4662. 43.Slini, T., A. Kaprara, K. Karatzas and N. Moussiopoulos. (2006) “PM10 forecasting for Thessaloniki, Greece.” Environmental Modelling and Software, 21, 559-565. 44.Tsuang, B.J. and C.P. Chao. (1999) “Application of circuit model for Taipei City PM10 simulation.” Atmospheric Environment, Vol.33, 1789-1. 45.Tsai, Y.I. and M.T. Cheng. (2004) “Characterization of chemical species inatmospheric aerosols in a metropolitan basin.” Chemosphere 54: 1171 -1181. 46.Van der Wal, J.T., and L.H. J.M. Janssen. (1996) “How contribute emission of PM10 and meteorology to concentrations of fine particles in the Netherlands.” Journal of Aerosol cience 27(Suppl. 1), S681-S682. 47.Van der Wal, J.T. and L.H. J.M. Janssen. (2000) “Analysis of spatialand temporal variations of PM10 concentrations in Netherlands using kalman filtering.” Atmospheric Environment 34, 3675-3687.
摘要: 
本研究以倒傳遞類神經網路進行揚塵量PM10濃度之預測,為提升其預測力,除了以彰化縣線西空氣品質監測站之氣象參數(溫度、降雨量、相對濕度、風向及風速)與環境監測參數(PM10及PM2.5)作為輸入變數之外,並增加梧棲氣象站之氣象參數(全天空輻射、日照時數、氣壓及雲量)作為輔助線西空品站缺乏之輸入變數。輸入變數於網路之前,先進行正規化及數據篩選,並為瞭解時間序列不同對預測結果之差異,本研究分為不分季模式(預測全年、夏季與冬季)、夏季模式及冬季模式,經由訓練與測試得知,分季後之預測能力皆有所提升,而本研究模式之預測能力依序為夏季模式>不分季模式>冬季模式,進一步探討模式之最佳輸入變數組合,結果顯示藉由篩選變數之動作,除了可減少網路訓練時間及有效提升實際值與預測值之相關性外,亦可避免僅由相關性分析作為判斷篩選變數之基準,造成誤刪變數之情形發生,而篩選後之最佳輸入變數皆包括前一日PM10、相對濕度、風速、日照時數、降雨量及氣壓等六項。為判別輸入變數對於模式之影響力,於各模式進行敏感度分析,其結果顯示PM10普遍受到前一日之延續作用影響,亦即為前一日PM10濃度影響模式甚鉅,其中又以不分季與冬季模式較為顯著。三組模式可接受之誤差百分比範圍約為30 %,顯示本研究之成果具有一定代表性,可提供後續研擬防範PM10措施之相關依據。

This study used Back-Propagation Neural Network to predict the concentration of PM10 of fugitive dust. For improving the prediction, meteorological parameters (temperature, rainfall, relative humidity, wind direction and wind velocity ) and environmental monitoring parameters (PM10 and PM2.5) from the Siansi air quality station were taken as input variables, and also used meteorological parameters (surface radiation, atmospheric pressure, cloud) from Wuqi weather station to be supporting input variables. Before setting parameters in the model, the data be normalized and screened. In order to understand the effect of different time series on prediction, this study developed models including the whole-year model (predicting the entire year, summer and winter), the summer model and the winter model. Through training and testing, the results showed that models which developed from different seasons could improve the prediction, and the predictions are as follow: the summer model > the whole-year model > the winter model. Further investigating optimum input variables in five models, showed that after sieving variables, it could reduce training times, improve the correlation between predictions and actual observed values, and it could also avoid variables being deleted by mistake. The optimum input variables by screening included PM10 of the previous day, relative humidity, wind velocity, sunshine hours, rainfall and atmospheric pressure. To compare the effect of input variables, this study analyzed the sensitivity of variables. The results showed that the concentration of PM10 was affected by PM10 of the previous day greatly, and the situation was obvious in the whole-year model and the winter model. The acceptable range of percentage error was 30% in models, and it showed that the performance of this study was representative and could provide basis for follow-up studies.
URI: http://hdl.handle.net/11455/33210
其他識別: U0005-2308201210370700
Appears in Collections:水土保持學系

Show full item record
 
TAIR Related Article

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.