Nonlinear Programming Design of Wastewater Treatment Plant

An efficient optimization method is selected and adapted to the least-cost design of a wastewater treatment plant including six units (primary settler, aerator, final settler, thickener, anaerobic digester, and vacuum filter). A previous companion paper presented a complete mathematical model describing the operation of the plant. The model, which turns out to be highly nonlinear, is reformulated as a geometric program, in order to take advantage of this quite flexible transcription mode. The generalized reduced gradient is then used and adapted for seeking the least-cost design of the plant. This method appears very efficient with regard to both computational results and cost minimization. The last part of the paper emphasizes and illustrates the recurrent use of the method in generating optimal solutions submitted to various practical and operational constraints, thus providing the designer with a efficient and wieldy tool for preliminary design.