|Ph.D Student||Avni Noa|
|Subject||Management of Regional Water Systems under Demand|
|Department||Department of Civil and Environmental Engineering||Supervisors||Professor Barak Fishbain|
|Professor Emeritus Uri Shamir|
|Full Thesis text|
The objective is to develop a model for management of regional water supply systems (WSSs), fed from natural sources and desalination plants, under demand uncertainty. Consumer demands are among the uncertain quantities that affect planning and operation of regional WSSs, characterized by high heterogeneity due to the existence of different consumer types (urban, industry, and agriculture). Consequently, the management policy of a WSS has to consider the spatial and temporal demand uncertainty. A two-stage approach has been developed to meet the research objective: (i) the regional spatial and temporal demand variability is analyzed and its effect on the WSS operations is elucidated; and (ii) a model for regional WSS operations under demand uncertainty is developed and demonstrated.
Implementable decisions are derived from an ensemble of Implicit Stochastic Programming (ISP) solutions. In ISP, a deterministic optimization model, is applied to an ensemble of possible demand realizations. ISP has the advantage of rapid implementation, and facilitates visualization of the range of possible operating policies. However, this process yields a set of separate optimal solutions, and not the required single implementable decision. To determine an implementable operating plan, the set of ISP solutions are aggregated by applying an aggregation rule, with a user-defined parameter, over the range of resulting decision variables. The rule proposed herein uses a p-percentile coverage over the set of the monthly desalination quantities in the set of ISP solutions.
The entire process was applied to a small demo WSS and to the central part of the Israeli National WSS (INWSS). Water demand uncertainty was embedded by applying pattern analysis to a set of real water consumption data, and then generating an ensemble of demand scenarios using a Fourier-transform algorithm. The results show a trade-off between system reliability and operational cost, the total desalination production, and excess production of unused desalinated water, thus providing the regional WSS managers with a tool to support decisions in management of regional WSS fed from natural and desalination sources in the face of demand uncertainty