Developing a Reliability-Based Waste Load Allocation Strategy for River-Reservoir Systems

Abstract

Enhanced socioeconomic criteria and temporal changes in the topology of the system often require waste load reallocation (WLRA) in a river-reservoir system to sustain long-term water quality standards. In addition to climate and hydrological changes, hydrologic fragmentation and dam construction may significantly affect the waste-accepting capacity of the water body through changes in its physical, chemical, and even biological characteristics. Deterministic waste load allocation optimization designs are often bounded with a set of rigid constraints. These constraints do not allow any flexibility to account for uncertainties and the possibility of system failure. This paper presents a reliability-based waste load reallocation model in a complex river-reservoir system. We have linked a physical and surrogate simulation model with the Particle Swarm Optimization algorithm to present an efficient methodology for reallocating waste loads in a river-reservoir system with reliability constraints. Reliability requirements are addressed by different sets of constraints in three different formulations for the entire planning horizon. The problem, as defined, contains real and integer variables, and is formulated as a mixed-integer nonlinear programming problem. It finds the maximum values of monthly waste loads that may be discharged into the river-reservoir system under predefined reliability constraints. The surrogate model itself is refined using an online dynamic routine which makes it suitable for planning waste load allocation in multiperiod and high-dimensional system optimization under reliability-based water quality constraints. The proposed model is applied to the Karkheh river-reservoir system to illustrate its performance under various reliabilities.