A multi-step approach is proposed for the optimal design of the D-Town pipe network provided within the framework of the Battle of the Water Networks II (BWN-II). The approach is based on three steps: the first two steps concern identification of optimal solutions which make it possible to respect the water utility requirements under normal operation of the system and consist of 1) system analysis aimed at identifying feasible solutions of the problem through engineering judgement and 2) their optimization by using a multi-objective genetic algorithm; the third step concerns the adjustment of the optimal solutions obtained after step 2 of the procedure in order to respect the water utility requirements under power failure operation as well. The analysis manually developed within step 1 of the procedure made it possible to identify some good first attempt solutions and a reduced set of decision variables to be used within the automatic optimization process (step 2): in this way the optimization space was significantly reduced and the multi-objective optimization was properly “guided” through a not exceedingly large research space. The optimal solutions were then manually adjusted (step 3) in order to cope with a two-hour interval of power failure, which could take place at every time of the simulation period, by inserting a suitable number of sufficiently powerful diesel generators in the pumping stations. The final solution provided is characterized by a zero value of the water age objective function and was selected among the solutions of the Pareto front in order to achieve a good balance between total annual costs and total annual Green House Gas emission.
A multi-step approach for optimal design of the D-town pipe network model
ALVISI, Stefano;CREACO, Enrico Fortunato;FRANCHINI, Marco
2012
Abstract
A multi-step approach is proposed for the optimal design of the D-Town pipe network provided within the framework of the Battle of the Water Networks II (BWN-II). The approach is based on three steps: the first two steps concern identification of optimal solutions which make it possible to respect the water utility requirements under normal operation of the system and consist of 1) system analysis aimed at identifying feasible solutions of the problem through engineering judgement and 2) their optimization by using a multi-objective genetic algorithm; the third step concerns the adjustment of the optimal solutions obtained after step 2 of the procedure in order to respect the water utility requirements under power failure operation as well. The analysis manually developed within step 1 of the procedure made it possible to identify some good first attempt solutions and a reduced set of decision variables to be used within the automatic optimization process (step 2): in this way the optimization space was significantly reduced and the multi-objective optimization was properly “guided” through a not exceedingly large research space. The optimal solutions were then manually adjusted (step 3) in order to cope with a two-hour interval of power failure, which could take place at every time of the simulation period, by inserting a suitable number of sufficiently powerful diesel generators in the pumping stations. The final solution provided is characterized by a zero value of the water age objective function and was selected among the solutions of the Pareto front in order to achieve a good balance between total annual costs and total annual Green House Gas emission.I documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
