Configuration of urban water eutrophication abatement systems using a novel regression-based optimization scheme.

A novel regression-based target-oriented robust optimization (R-TORO) framework is introduced to obtain robust designs of surface water eutrophication abatement systems during the early stages of urban water planning. The proposed framework can process a manifold of data for estimating the relationships of system elements and the uncertainty distribution of imprecise parameters, a novel extension of the canonical TORO framework. A demonstrative case study about an early-stage urban water infrastructure planning for abating Phosphorous (P) accumulation in urban waters is employed to demonstrate the R-TORO framework. The following system targets are considered: target cost and target water cycle sustainability index (WCSI). The findings show that the abatement of P accumulation in soil contributes the largest to WCSI. Consequently, the R-TORO-generated system design suggests the implementation of impervious local roads connected to sewers with improved P removal. The simulated/estimated performance of the eutrophication abatement system design obtained using the R-TORO are as follows: import of water = 2.88 × 10 m/y; export of P = 4.96 × 10 g P/y; P concentration in recreational lake = 0.04 mg P/l; P concentration in canal system = 0.30 mg P/l; P accumulation in lake sediments = 0.08 g P/m,y; P accumulation in canal sediments = 1.18 g P/m,y; P accumulation in soil = 0.01 g P/m,y.