Use this URL to cite or link to this record in EThOS:
Title: The integration of topology and entropy-based reliability into the optimal design of water distribution systems
Author: Saleh, Salah H. A.
Awarding Body: University of Strathclyde
Current Institution: University of Strathclyde
Date of Award: 2013
Availability of Full Text:
Access from EThOS:
Access from Institution:
Establishing a new water distribution system (WDS) essentially involves proper planning of system components to minimize cost, sizing of such components to operate under normal operating conditions and assessing system performance under abnormal operating conditions to minimize effects of components failure. This thesis investigates combinng these optimization aspects while specifically focusing on optimal planning of system components amongst all possible topologies, optimal design over all possible combinations of components sizes, and improving system performance over all possible levels of hydraulic reliability. In order to address these issues in full, a novel many-objective genetic algorithm approach to the simultaneous optimization of topology, design and reliability of the WDS has been developed in this research. The novelty and originality carried out in this research are presented next. A new multi-objective approach for coupled topology and pipe size optimization of WDS is developed. The approach is the first to exploit in full both feasible and infeasible parts of the entire solution space of topology and design within the search procedure. A new algorithm for topology confirmation is developed to identify and quantify infeasible topologies containing nodes and pipes isolated from supplying sources. The algorithm is coupled with the penalty-free strategy to enable both infeasible topologies and designs with insufficient pressures to contribute to achieving the least cost design of the WDS. Previously, solutions belonging to either infeasible topologies or having insufficient pressure were being discarded from the search process. The approach is computationally efficient and outperforms previous methods of coupled topology and design optimization of WDS. A new multi-objective approach for design and entropy-based reliability optimization of WDS is developed. The statistical entropy highly dependent on flow directions was used as a measure of the WDS reliability within the approach. The approach is the first to globally maximize entropy of the WDS. Previously, the search procedure was locally restricted to predefined sets of flow directions. To address this issue, a new algorithm that eliminates the need to specify limited sets of flow directions in advance of the search process is developed. The algorithm is integrated with the penalty-free strategy in order to fully exploit the entire solution space of entropy and design. The approach simultaneously combines local and global maximization of entropy with cost minimization over the entire solution space of pipe sizing. To reduce difficulties encountered in searching into the solution space of this many-objective problem, a new concept that computationally combines objectives of hydraulic infeasibility with global and local maximization of entropy into one objective is developed. The approach is applied to two benchmark networks yielding superior results in terms of global maximization of entropy and good balance between cost and entropy. The coupled topology and entropy-based design optimization of WDS are integrated into a penalty-free many-objective framework. It is the first to combine topology, pipe size and entropy-based reliability optimization of the WDS in a simultaneous way. The maximization of entropy in this combination is tackled in an entirley novel way by simultaneously exploring entropy belonging to multiple topologies and multiple entropies belonging to the same topology. To address exploring entropy of new topologies, a network complexity measure that accounts for number of pipes contained in a topology is introduced as an objective. The concept of handling this many-objective problem is extended to account for topologic infeasibility, hydraulic infeasibility and global and local maximization of entropy. Excellent results have been efficiently achieved regarding the provision of a variety of maximum entropy designs distributed to different toplogies and having good compromise between entropy and cost.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available