請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/45381
標題: | 自來水配水管網漏水現象之分析、預測與管理之研究 Analyzing, Predicting and Managing the Leakage Phenomena of Water Distribution Network |
作者: | Cheng-I Ho 何承嶧 |
指導教授: | 駱尚廉(Shang-Lien Lo) |
共同指導教授: | 林明德(Min-Der Lin) |
關鍵字: | 類神經網路,徑向基函數,地理資訊系統,管線汰換,漏水, Artificial neural network,Radial basis function network,Geographic information system,Pipe replacement,Leakage, |
出版年 : | 2009 |
學位: | 博士 |
摘要: | 本研究成功發展完成「自來水管線汰換ANGEL模組」,主要研究流程係蒐集案例區內自來水配水管網曾經發生漏水地點之資訊,運用模糊理論(fuzzy logic, FL)、地震因子類神經網路模式(seismic-based ANN model, SBAM)及地理資訊系統( geographic information system, GIS),據以推估不同地震震度(或規模)條件下之自來水配水管網漏水分布,並可藉由直讀漏水點位分布圖之疏密程度建立管線汰換序列,以加速決策分析時效。
根據文獻記載,本研究三個案例供水區將地震因子導入ANN乃是漏水相關研究的創舉;另為確保資料品質適合進行數值模擬,除採資訊化紀錄資料:如中央氣象局地震資料及台灣自來水公司「自來水修漏管理系統」管線漏水資料,並將無法量化或依法規要求需維持限定值之參數予以剔除,證實ANN的確適合做為推估漏水的研究工具,模擬過程亦比較ANN倒傳遞函數(backward propagation network, BPN)及徑向基函數(radial basis function network, RBFN)之模擬效能,結果為RBFN較BPN更適合建立管線汰換優化seismic-based ANN model(SBAM)。 本研究同時導入GIS據以建立三種不同型態漏水點位圖形,包括「歷年漏水點位分布圖 (site-based leakage point graph)」、「道路漏水點位分布圖 (road-based leakage point graph)」及「管線漏水點位分布圖(pipe-based leakage point graph)」,相較過去人工判讀紙圖上有關管線漏水的資料,藉由GIS的協助,可更有效達成快速翻閱及整理圖資,以進行後續管線汰換的排序作業。 過去管線汰換作業存在僅針對具有漏水記錄之管線才得以納入汰序的限制,對於「未具漏水紀錄」之管線則無法納入排序,因為在管線汰換實務作業上,並無資料可以協助不具漏水記錄的管線,該如何納入管線汰換?然而事實上管線「没有漏水記錄」並不代表「不會或未曾發生漏水」。因此本研究研發完成之SBAM已成功解決上述問題,並可有效降低人力及時間成本,對於推動台灣地區降低自來水漏水率計畫將有可預期的實務貢獻。 This work focused on developing an approach for prioritizing the order of pipe replacement of a water distribution system (WDS). The methodology was based on the integration of seismic-based ANN model (SBAM) and GIS-based system, fuzzy logic (FL), named “ANGEL” module, to assess water leakage. However, three scenarios in Taiwan were chosen as the case study because these districts were frequently hit by earthquakes which are the main leakage-caused factors. There’s something new in this research. That is, the influenced factor of “the number of maginitude-3+ earthquake” is emphasized. Especially, this study is the first attempt to manipulate earthquake data in the break-event ANN prediction model. Consequently, it was proven highly relative to pipe leakage. For the purpose of quality assurance and quality control, FL screens parameters out in consideration of non-quantifacation or the requirement to conform to regulatory standards. The qualified earthquake data obtained from “Taiwan Central Weather Bureau (TCWB)” and pipe data derived from “Taiwan Water Corporation Pipeline Leakage Repair Management System (TWC-PLRMS)” were classified to build SBAM which was analyzed by both backward propagation network (BPN) and radial basis function network (RBFN). However, a comparison of the accuracy and reliability of the prediction model between BPN and RBFN achieved that RBFN outperformed BPN. Additionally, GIS was used to display the visual effect of historical leakage points. Spatial distribution of the pipeline break event data was analyzed and visualized by GIS. As for the steps to develop visual effect, GIS was applied to draw as below: site-based leakage point graph, road-based leakage point graph, and pipe-based leakage point graph. Compared to the traditional processes for determining the priorities of pipeline replacement, the methodology developed is more effective and efficient. Finally, the methodology can overcome the difficulty of prioritizing pipeline replacement even in situations where the break event records are unavailable. In fact, it has made a valuable contribution to the practical tasks. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/45381 |
全文授權: | 有償授權 |
顯示於系所單位: | 環境工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-98-1.pdf 目前未授權公開取用 | 5.06 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。