- 學位論文
應用灰色理論之GM（1,1）模型 預測社區污水處理廠之出流水水質
THE PREDICTION OF COMMUNITY EFFLUENT WATER QUALITY BASED ON GM（1,1）OF GREY SYSTEM THEORY
摘要
從內政部營建署〝挑戰二ΟΟ八污水下水道建設實施計劃〞中，已勾畫出國家未來污水下水道的藍圖，預計未來污水下水道的普及率將由九十一年底的10.1﹪提昇至九十六年的20.3﹪，為了防止生活污水及工業廢水未經處理逕排放至河川（承受水體），將污染生態環境，嚴重影響國人的生活品質，故污水處理廠更扮演著舉足輕重的角色。 本研究利用灰色理論之GM（1,1）模型，以實際的社區污水處理廠操作數據，在系統因子時間序列與交互作用關係下，對其出流水水質之BOD、COD及SS進行模擬預測，再以實際值及模型預測值進行殘差分析，以評估模型之預測效能。 本論文研究結果發現，對某社區污水廠之出流水水質能做準確預測，本實驗建模預測的準確度，在COD的預測方面，平均預測值約為56.8220，平均殘差約為6.9595％。在BOD的預測方面，平均預測值約為15.2157，平均殘差為17.7555％。在SS的預測方面，平均預測值約為7.7713，平均殘差約為2.5696％。效果表現頗佳，將可提供操作管理人員重要之參考，以預先採取適當的因應措施，以免承受水體遭受過大的污染。
並列摘要
In the “Challenge 2008 Sanitary Sewer Construction & Implementation Plan” of the Construction and Planning Agency, Ministry of the Interior, it maps out the blueprint of the future sanitary sewer. It is predicable that the ratio of prevalence of the future sanitary sewer will be increased from 10.1% by end of 2002 to 20.3% in 2007. In order to prevent the direct discharge of the unpurified household sewage and industrial wastewater to the river (receiving water body), which run risk to pollute the ecological environment and affect our living quality, a wastewater treatment plant plays an important role. This research utilizes the GM (1.1) model of the grey system theory and use the actual operation data of a community wastewater treatment plant to conduct simulation prediction on the BOD, COD and SS of its effluent water quality based on the factorial time sequence and interaction relationship. Then the actual value and model prediction value are utilized to conduct the residual analysis in order to evaluate the prediction function of the model. The result of this thesis research finds that accurate prediction can be conducted on the effluent quality on a certain community wastewater treatment plant. For the accuracy of this experimental model prediction and in respect of COD, the average prediction value is 56.8220 and the residual mean-value is 6.9595%. In respect of BOD, the average prediction value is 15.2157 and the residual mean-value is 17.7555%. In respect of SS prediction, the average prediction value is 7.7713 and the residual mean-value is 2.5696%. The effect performance is quite good that can be provided to the operators as an important reference, so that advanced and appropriate measures can be adopted in order to avoid too much pollution on the receiving body of water.