Ilemobade, A.A.Stephenson, D.2016-11-042016-11-042006Ilemobade, A.A. and Stephenson, D. Application of a constrained non-linear hydraulic gradient design tool to water reticulation network upgrade. Urban Water Journal 2006 3(4), pp. 199-214. DOI: http://dx.doi.org/10.1080/157306206010602131573-062X (Print)1744-9006 (Online)http://hdl.handle.net/10539/21408Southern Africa has embarked on substantial expansion of its water supply network in order to ensure safe, reliable, convenient and sufficient water for everyone. To achieve this, new systems are being built and many existing systems are being upgraded. The upgrade of many existing systems is required for two reasons: some currently functional systems may run dry if subjected to additional demands as these systems were not initially designed to cater for such demand, and some systems are currently non- or sub-functional as they were ill-designed and/or ill-implemented from inception. Many of the systems that require upgrade are underdesigned due to a lack of skill, tools and/or knowledge of designers, or from other extraneous factors (e.g., illegal connections and sabotage). It is hardly surprising therefore that the failures of water projects in developing countries are recorded to be as high as 80%. Ill-designed systems increase operation and maintenance costs significantly. In especially Southern Africa, designers require simple, yet rigorously tested tools to facilitate sustainable, yet cost-effective network designs. Presented in this paper is a simple, yet robust constrained non-linear hydraulic gradient network reticulation design tool. The design tool is calibrated using the New York City water supply problem that has served as a benchmark problem for other models and then applied to the Selebi – Phikwe (SP) water reticulation network (WRN) in Botswana, which was designed based on engineering judgement. The optimization algorithm employed in the design tool is based on the concept that a hypothetical hydraulic gradient for a hydraulically balanced WRN exists that, when achieved iteratively, produces optimal pipe sizes and an optimal flow relation between each pipe. The unique problems and challenges of the SP WRN (pressure deficiencies in sections of the existing network and the proposed addition of three new residential developments) required determining the most appropriate peak and night flow operating scenarios, and optimal pipe sizes for the proposed expansion of the network. Optimization by trial and error had been previously employed in the design of the SP WRN—a common practice amongst water system designers, and the results are compared with those generated with the design tool. The design tool achieved a 62% reduction in total pipe cost from that obtained by trial and error for the SP WRN problem. At the same time, the design tool gives comparable pipe costs to those published in literature for the New York City water supply tunnels problem.enThis is an Accepted Manuscript of an article published by Taylor & Francis in “Urban Water Journal” on 22 December 2006, available online: http://www.tandfonline.com/ [http://dx.doi.org/10.1080/15730620601060213].Water networksExpansionCost savingsWater distributionWater reticulation networkOptimal flowHydraulic gradientPreprint