Research - Category IV: Water Development & Management Alternatives

Modeling and Optimization of Water Quality in a Large-Scale Regional Water Supply System
(Funded 2000-2001)

Principal Investigator: William W-G Yeh
Civil and Environmental Engineering Department
UC Los Angeles
(310) 825-2300
williamy@seas.ucla.edu

Executive Summary:
Statement of the Critical State Water Research Problem: In Southern California, population and economic growth are making increasing demands on the water supply. In the past, as demand for reliable developed, local and state water resources agencies regularly funded large projects that improved water supply and quality. Because of environmental considerations, this approach is no longer viable. During the last decade, regulatory decisions and environmental concerns have reduced the availability of imported supplies to Southern California from the State Water Project (SWP) and the Colorado River Aqueduct (CRA). In light of the shrinking water supply and increasing water demands, water resources agencies in Southern California are facing a challenge while continuing to meet the water demands for agricultural, municipal, and industrial water needs. In addition to the reduction in supply and the increase in demand, the costs of supplying imported water have risen substantially at the same time. When a regional water distribution system receives water from multiple sources of different quality, an important management objective is to optimize the operation of the water distribution system so that water is delivered to each demand location within the distribution system with the required quantity and acceptable quality. this requires proper control of water blending in the system, which serves to mix and convey the multiple sources of water supply.

Summary of Research Approach: The overriding objective of this proposal is to develop a multicommodity flow model that can be used to optimize water distribution in a regional water distribution system with blending requirements imposed as a set of constraints. Waters from different sources are considered as distinct commodities, which concurrently share a single water distribution system. The proposed approach uses volumetric water blend to represent water quality. This representation is much preferred by water resources agencies for their long-term planning. The multicommodity model will be embedded in an optimization model that determines the operation policy for a regional water distribution system. The optimization model will be solved by both the classical gradient-based algorithms as well as by non-gradient-based algorithms, such as the genetic algorithm.

Statement of Results, Benefits and/or Information Expected from the Project: The proposed research project will develop a new multiobjective optimization model that will incorporate the water blending requirements as constraints. However, the economic rational behind blending is not within the scope of the present proposal. The expected results will include a mathematical model that determines, for the decision-maker, the optimal water distribution policy--a policy that will optimize the specified objectives subject to the imposed set of constraints. The objectives to be considered include minimizing the total shortage in water supply, minimizing the deviation of reservoir storage from the preset targets, minimizing reservoir spills, and maximizing the resources usage. Weighting method will be used to resolve the multiple objectives. The proposed model can be used for sensitivity analysis to identify the bottlenecks in the system, that is, the constraint to which the optimal water supply policy is most sensitive. If the bottleneck involves physical constraints, this information will assist in capacity expansion planning. Arrangements have been made with Metropolitan Water District of Southern California for collaborative participation to maximize the immediate applicability of the proposed methodology.