摘要
近年來因全球水資源短缺，新興水源的需求備受關注，而國內於廢污水再生利用相關研究成果極為豐碩，又因政策推動使水再生利用上下游產業鏈更加完善，但台灣水價偏低，造成廢污水再生利用經濟誘因以及相關實廠經驗與資料不足，使預算編列與實際成本估算，常難以掌握成本動態資訊，因此造成實務應用評估上的困擾，故本研究利用目前國內之水回收再生利用處理廠的各項成本資料，依水源水質以及處理流程分類，進行Cobb-Douglas、Translog成本函數分析，再依統計檢定驗證其模式之正確性，建立成本函數推估之方法，利於將其應用於工程推估模式。 依據分析結果顯示以六家不同都市污水廠為水源放大至54筆數據，產水規模1000-50000 CMD，逆滲透(RO)薄膜-A程序可以Cobb-Douglas 成本函數表示，微濾薄膜(MF)-B程序與砂濾-C程序部分，Cobb-Douglas、Translog成本函數皆可應用於兩個處理流程之成本推估模式。工業廢水部分，依二級放流水水質區分甲、乙，分別以三家不同之工業區污水處理廠為水源，個別放大至43、73筆數據，甲水質部份三種處理程序皆無法以Cobb-Douglas、Translog成本函數進行成本推估；乙水質部份，Cobb-Douglas 成本函數可表示為微濾薄膜(MF)-B程序的成本函數，但在逆滲透(RO)薄膜-A程序、砂濾-C程序，Cobb-Douglas、Translog成本函數經統計檢定結果皆不符合迴歸方程式之假設。工業區污水廠因成本資料不足、成本函數本身存在隱函數，建議後續依據更新資料持續進行推估模式修正，評估此推估模式之適用性。 Cobb-Douglas 成本函數以均化成本為依變數，發現當產水量增為2倍，單位產水成本減少至0.8-0.9倍左右，以此判斷具經濟規模，且由分析結果顯示，不應只藉由判定係數作為迴歸方程式選擇之依據，應再檢視迴歸方程式的配適度、參數檢定以及殘差分析，確立迴歸方程式之合理性。
並列摘要
Water resources issue has been greatly concerned due to global water shortage nowadays. In Taiwan, studies of wastewater reuse have been greatly developed to be alternative water resources in past decades. However, in default of the economic incentives, practical experiences and relative information of wastewater reuse, it was difficult to establish the integrated budget estimation. In this study water sources and treatment processes were classified, and Cobb-Douglas and Translog cost function analyses were performed by regression model fitting in order to establish an adequate model for industrial applications. For secondary domestic wastewater, Cobb-Douglas cost function could be applied in process A (reverse osmosis system), B (microfiltration system) and C (sand filtration), and Translog cost function could be used in process B and C. For secondary industrial wastewater, first and second class of wastewater were defined according to the water quality. First class Cobb-Douglas and Translog cost function could not represent the model in process A, B and C. Second class, Cobb-Douglas cost function could represent the model in process B, but neither of Cobb-Douglas nor Translog cost function could represent models in process A and C. Average cost is the response variable Cobb-Douglas cost function. When two times of water production, the average cost reduced about 10-20%, demonstrating the cost function has economies of scale. The results of analyses indicated that besides of R square value, the goodness of fitting, regression coefficient and residual analyze also need to be investigated in order to check the model adequacy.