Multi-criteria Calibration of Water-supply Networks under Unsteady Flow

To evaluate a pipe network performance, both steady and unsteady flow conditions are required to be accurately analyzed in the system. For this purpose, mathematical models are developed and utilized. Calibration of parameters with dynamic effects like the steady and unsteady friction loss factors as well as pipe wave speed is crucial to the simulation models. A common approach for calibration of pipe networks under unsteady flow is the Inverse Transient Analysis (ITA) method. This study introduces a model for hydraulic analysis of unsteady flow in pressurized pipes using the numerical Method of Characteristics (MOC). To each calibration parameter a correction factor is specified. Then, a nonlinear optimization problem is developed with the aim of minimizing the discrepancy between the calculated and measured transient pressure heads at the measurement sites. The decision variables are the correction factors multiplied by the pipe wave speeds, steady friction factors and unsteady friction loss components. In this Article, the proposed model applied to an experimental system set up in the Technical University of Lisbon. By defining the genetic algorithm to minimize the objective function, after 30 generations error rate was reduced from 140 m to 92.15 meters. The results, increasing the accuracy of modeling showed as much as 33%.