Research And Application On Resilience Index Scheme For Water Distribution And Sewer Network Systems

The fast development of urbanization causes the phenomenon that "Construction is more crucial than management" in the domestic water untilities during recent years. In doing so, the water distribution and sewer network systems(WDSNSs), due to lack of valid planning and scientific management, are operated at an inefficient state. It will make an extra cost to maintain the system, and hardly offers a sufficient service. In addition, economical development brought the severe environmental problems. The sustain WDSNSs is a great concern increasingly. Hence, in this paper, the resilience theory was introduced. The evaluation indices of resilience in WDSNSs were formulated. The multiple objectives model of optimal design was employed for economy, resilience and sustainability in WDSNSs.When water distribution systems(WDSs) run at a low pressure state, the error of the hydraulic computation based on demand driven model will occur. In this paper, the model based on the pressure driven demand method was developed, which is incorporated with the global gradient algorithm(GGA) to improve the EPANET program and extend the application programming interfaces(APIs). The dynamic wave routing method is reviewed for sewer network systems(SNSs). The APIs of SWMM is extended. The meta-heuristic optimization algorithms are reviewed and the classic NSGA-II and new developed BORG are used to solve the optimization problems of WDSNSs.On the basis of the resilience theory, the concept of resilience in the WDSNS area is proposed. The three new resilience indices are formulated, i.e., available power index(API), pipe hydraulic resilience index(PHRI), diameter-sensitive flow entropy(DSFE). Hanoi benchmark network is used, and the model of optimization design is built in WDSs for the cost-effectiveness tradeoff using resilience as an objective. Demand fluctuation and pipe break uncertainties are taken into account. The Latin hyper cube sampling method is used to quantify the demand uncertainty. The pipe break uncertainty is solved by the probability method in which all combinations of the possible pipe failure are traversed. The assessment framework of uncertainty perturbation is formulated to evaluate the capacity of resilience for tolerating disturbation by using failure scenario count, failure node count and failure degree fraction. The results show that the performance of resisting the perturbation of PHRI and API are better than that of DSFE. Moreover, the DSFE index has more complicated objective space and needs more evolution generations to achieve convergence.The concept of resilience is introduced into the SNSs. The surplus capacity of conveying is used to establish the resilience index of SNSs. The optimal design model of cost-resilience objectives is developed to optimize diameter and slope. The FSS benchmark network is used to obtain the cost and resilience tradeoff. The variation of sewerage amount maybe obey the normal, uniform, log-normal, and Weibull probability density distributions(PDFs). The Chi-square goodness-of-fit method is employed to investigate propagation of uncertainty through the SNSs routing. The results show that when the uncertainty of sewerage follows the normal, uniform and log-uniform distributions, the statistics of water depth with respect to normality outperform that of Weibull distribution. The pairwise correlations amongst average failure pipe count, average overload pipe count, total overflow, average water depth of manholes, sewer impacted factor, sewer network system resilience and cost are investigated under the conditions of different distributions laws of sewerage uncertainty. It is demonstrated that the more severely the uncertainty perturbs, the more significantly pipe are impacted and the correlations between indices are stronger. The correlation of node flows is ascertained. It is showed that the stronger the correlations among the nodal flows, the larger the extent of variation of water depth is, but the average of water depth is consistent.The practical optimization concept is introduced. The parallel computing, new modeling measure and mutual mode intervention technologies are used to optimize design for the real water distribution networks. The message passing interface(MPI) of parallel computation technology is used to parallelize the BORG multi-objective optimization algorithm. The distributed computing cluster is constructed. Using the decomposing modeling method is to reduce the complexity of decision variable space and objective space. The intervention strategy of mutual mode is added. These technologies all achieve the aims that accelerate the convergence of optimization process and reduce the computational burden. The D-Town water distribution network and W-Town sewer network are used as the real-world case studies. The triple objectives optimization model regarding cost, resilience and carbon footprint is established. The design solution is picked up from an optimal solution set to validate the obtained results by using the weighted objective method. The current optimization technology is proved to be feasible to be utilized in the town scale network design.