A Proposed Flow Path Model for Scheduling Optimization of a Water Distribution System

http://content.asce.org/conferences/ewri2008/

 

論文發表人: 鄭蔚辰  (加州大學洛杉磯分校土木及環境工程研究所博士班)

 

本研究提出一適用於求解水資源管理系統之通用性最佳化模型：流路法。此模型可描述並求解現有最佳化模型所無法處理之水資源管理問題。一般來說，水資源管理之模型描述水資源問題的兩個方面：(a) 水的運送及交易介於送水者與受水者間之關係; (b) 在送水系統中，水的位置及運送路線。目前世界上所使用之最佳化模型多以網流為架構，因此只能描述上述兩方面中之一項，而無法在不簡化問題的情形下同時而完整地描述水資源管理之問題，此外因受限於網流架構及特性，許多水資源的管理規則及用水人問的協議難以被現有模型所考量。因此本研究針對水資源系統之特性提出一更為通用之最佳化模型：流路法。流路法之優點如下：(1) 是一通用之水資源最佳化模型，可同時描述並最佳化水資源的分配、運送、排程、交易、交換及調用；(2) 同時描述上述(a)、(b)兩問題時無需簡化系統；(3) 可做為及時操作之決策輔助工具。此研究應用流路法於南加州之水系統，並得到極佳的結果。

 

A flow path model is developed for scheduling optimization of a water distribution system. The model simultaneously describes a water distribution system in two parts: (1) the water delivery relationships between suppliers and receivers and (2) the physical water delivery system. In the first part, the model considers waters from different suppliers as multiple commodities. This helps the model to clearly describe water deliveries by identifying the relationships between suppliers and receivers. The physical part characterizes a physical water distribution network by all possible flow paths. The advantages of the proposed model are that: (1) it is a general methodology to optimize water distribution, delivery scheduling, water trade, water transfer, and water exchange; (2) no simplifications are made for either the physical system or the delivery relationship; and (3) it can be used as a tool for real-time decision making for scheduling optimization. The model optimizes not only the suppliers to each receiver but also their associated flow paths for supplying water. This characteristic leads to the optimum solution that contains the optimal scheduling results and detailed information of water distribution in the physical system. That is, the water right owner, water quantity and its associated flow path of each delivery action are represented explicitly in the results rather than merely an optimized total flow quantity in each link. The proposed model is first tested on two hypothetical water distribution cases. The results show that the flow path model has the ability to optimize both the quantity of each water delivery as well as the flow path. The model is being applied to the water distribution system of the Metropolitan Water District of Southern California (MWD), which supplies water to 18 million people in Southern California.