A decision support tool for the design of integrated treatment systems for drinking water utilities

A decision support tool was developed for determining least-cost, integrated drinking water treatment plant design, or modification, with explicit consideration of influent variability and model parameter uncertainty. The decision support tool was applied to two design problems: design of an integrated treatment plant for particulate removal, and modification of an existing coagulation/flocculation process for enhanced arsenic removal. The tool provided the ability to evaluate the effect of influent variability and model parameter uncertainty on: (1) the design cost and regions of least-cost treatment configurations; and (2) the relative importance of the individual variable or uncertain parameters for reliably satisfying the appropriate effluent concentration constraint.; Least-cost designs were determined using a constrained non-linear programming approach; variability and uncertainty were included in the stochastic solution by Monte Carlo-type simulation. For any given influent condition, the least-cost treatment configuration was determined by selecting the lowest cost option from multiple potential configurations. Deterministic studies were performed to: (1) provide a basis of comparison for the stochastic results: and (2) illustrate the need for explicitly incorporating variability and uncertainty in the design process.; Water treatment plant design for particulate removal considered three treatment configurations based on four unit process, and included eight variable and uncertain parameters. The value of the stochastic solution was in reliably satisfying the effluent concentration constraint, not in the cost (cost increased with variability and uncertainty). Shifts in the least-cost configuration regions were dependent on the influent particle size distribution characteristics. For the smaller particle size range, variability and uncertainty had little affect on the least-cost configuration regions. For the larger particle size range, variability and uncertainty increased: (a) the conventional filtration region for larger averaged particle diameters, and (b) the contact filtration region for smaller averaged particle diameters.; Water treatment plant modification for compliance with a reduced arsenic MCL considered three treatment modifications based on two unit process, and included eighteen variable and uncertain parameters. Arsenic adsorption was described by a two-layer surface complexation model, and coupled with sedimentation and filtration models to describe removal. Predicted removals from the process model compared favorably to literature reports. The inclusion of variability and uncertainty enabled identification of regions where additional treatment was required, but there was no effect on the least-cost treatment modification, which always increased coagulant dose.