Optimal choice of wastewater treatment train by multi-objective optimization

Optimal choice of wastewater treatment train by multi-objective optimization

0.00 Avg rating0 Votes
Article ID: iaor20081914
Country: United Kingdom
Volume: 39
Issue: 2
Start Page Number: 125
End Page Number: 145
Publication Date: Mar 2007
Journal: Engineering Optimization
Authors: , ,
Keywords: heuristics: genetic algorithms, programming: markov decision, geography & environment
Abstract:

The typical domestic wastewater treatment train consists of some combination of unit operations for preliminary, primary, secondary, tertiary, and advanced treatment, and residual management, with many options being available for each type of unit operation. The challenge is to select treatment trains for which the extent and reliability of treatment are high, whereas the capital, operation and maintenance (O&M) costs of the treatment and land area requirement are low. This proposition has been formulated as a multi-objective optimization problem, and solved using the evolutionary/genetic optimization technique. The inputs required are the capital costs, O&M costs, land area requirements, and reliabilities of the unit operations of various types. In addition, overall environmental cost (E) corresponding to various treatment trains is input as a normalized parameter, which can take values in the range 0–100, with E being 100 corresponding to the ‘no treatment’ option. In other cases, E is a function of both treatment train efficiency and reliability. The problem was solved to determine the Pareto optimal, i.e. ‘no worse’ than each other, set of solutions under three conditions, viz. when E was not constrained, and for E<75, and E<50. Correctness of the algorithm was probed through a threefold analysis, (1) by solving a simplified two-objective problem, (2) by demonstrating the efficiency of the algorithm in picking up ‘sure-optimal’ solutions, i.e. solutions deliberately made optimal through manipulation of input data, and (3) by demonstrating that the set of optimal solutions remains approximately the same irrespective of the variations in the initial population size chosen for the genetic operations.

Reviews

Required fields are marked *. Your email address will not be published.