水廠原水水質指數之建立

Abstract

原水水質之良瓢與後續淨水處理單元及效率有極大之關聯性，因此，要評鑑淨水場之處理效能(performance)，必須先考量原水之水質狀況，方能合理鑑別水廠硬體在設計、操作上是否滿足有效處理此原水水質之要求? 另一方面承擔原水水質較差之淨水場，若要將此原水處理到符合飲用水水質標準，則所付出之貢獻是較大的，因此原水水質指數之建立，可以作為評估淨水場水質處理負荷的一客觀指標，據以合理計算淨水場之貢獻度。 本研究係藉由蒐集國、內外現有水質指數發展狀況，瞭解水質指數建置程序，並依據國內水源管理現況及淨水場原水等特性，選定水質指數計算模式，再經由檢討、制訂「評價因子」、「權重」及「點值」等3項變數，以建立台灣淨水場原水水質指數，達到評估水質優劣之目的。研究過程採用「加權平均法」數學模式，並採計4項評價因子(大腸桿菌群、氨氮、總有機碳及重金屬等4項)，利用「德爾菲法」專家問卷調查方式決定權重(分別為0.24、0.21、0.32及0.23)，並檢討訂定「點值(Ii)」，分別以「自訂級距」及「距標倍數」2種方法來計算台北自來水事業處所轄4個淨水場 (有10種原水)，及台灣自來水公司所轄338個淨水場(有357種原水)之指數。以96年平均原水水質數據計算之結果顯示，原水水質指數(I) 為100佔有之比例太高及同分問題，無法有效區分水質；原水水質指數(Ⅱ)之結果，可有效區分淨水場原水水質等級，可作為評估原水水質優劣之一有效方法。

The raw water quality will highly affect the performance of following treatment units. To evaluate the state of plant performance, the first step is, therefore, necessary to realize the conditions of raw water quality. In case a plant with poorer raw water quality than others need to do more works and take more stress-loading in order to comply with the requirements of drinking water standard. Establishment of raw water quality index for water treatment plants can be, therefore, adopted to as a effective tool to evaluate the treatment performance for water treatment plants and identify the contributions by water treatment plants. The approach of this study was applying the mathematical model using averages weighting method with four parameters and the Delphi Technique (Delphi survey) to decide the weights of these four parameters and corresponding points (Ii). The results of the weights for four parameters such as total coliform, ammonia, TOC and heavy metal are 0.24, 0.21, 0.32 and 0.23, respectively. Then, according to “self-made class interval” and “length multiple” methods, the raw water quality indexes of ten raw waters from four water plants of Taipei Water Department and 357 raw waters from 338 water treatment plants belong to Taiwan Water Corporation were calculated. The points were determined based on the yearly raw water quality data in 2007. The results show that the approach by using the self-made class interval method failed to distinguish the water quality because of too many plants earned 100 points as well as having the equivalent score. In the mean time, the approach by using the length multiple method can effectively distinguish the state of raw water quality for most of plants.