A semi-pressure-driven approach to reliability assessment of water distribution networks

A new methodology has been developed, using the EPANET Toolkit functions and the C++ programming language, for the optimal locations of isolation valves in water distribution systems. The new methodology is based upon interfacing a new semi-pressure-driven approach (SPDA) to network hydraulics for reliability/availability assessment in terms of supply shortfalls, an optimization procedure based upon simulated annealing, and the EPANET hydraulic model.; Hydraulics of water distribution systems can be solved using two approaches. In the demand-driven analysis primacy is given to demands and pressures are calculated regardless of their validity. The head-driven approach recognizes a relationship between nodal pressures and nodal demands. The general consensus is that the demand-driven approach works well under normal operating conditions while the head-driven approach is more suitable for simulating partially-failed conditions.; SPDA uses the demand-driven results as a starting point to find supply shortfalls in an iterative manner. In addition to its use as a descriptive tool, the new methodology also has been used for reliability assessment considering probabilistic aspects of water pipe breaks in distribution networks. The final phase of this research coupled the new methodology with the simulated annealing search to determine the optimal valve scheme of a network in order to maximize reliability/availability.; A network valve scheme is recognized for realistic isolation of pipe breaks in this new method. Hydraulic performance is measured deterministically as the ratio of the available supply to the original demand at any node and is called available demand fraction. We illustrated the use of SPDA as a descriptive tool for predicting hydraulic performance considering three what-if scenarios in an example network. The method has also been successfully applied for reliability assessment using three life distribution models for pipe breaks.; Our optimization efforts to find the optimal valve scheme in an example network justified the need for valve optimization and revealed that significant cost savings can be achieved with a less-than-fully-valved network while sacrificing little from reliability. The example application also supported our hypothesis that smaller diameter pipes will more likely have isolation valves than larger diameter pipes in a reliability based optimization framework.