泄漏
还原(数学)
泄漏(经济)
遗传算法
航程(航空)
数学优化
节点(物理)
计算机科学
阶段(地层学)
数据挖掘
可靠性工程
算法
工程类
数学
几何学
经济
古生物学
航空航天工程
宏观经济学
环境工程
生物
结构工程
作者
Sophocles Sophocleous,Dragan Savić,Zoran Kapelan
标识
DOI:10.1061/(asce)wr.1943-5452.0001079
摘要
This research article presents a model-based framework for detecting and localizing leaks in water distribution networks (WDNs). The framework uses optimization and systematic search space reduction. The method employs two stages: (1) the search space reduction (SSR) stage and (2) the leakage detection and localization stage (LDL). During SSR, the number of decision variables is reduced along with the range of possible values, while trying to preserve the optimum solution. Then, at the LDL stage, the size and area of a leak are found. The leak localization method is formulated as an optimization problem, which identifies leakage node locations and their associated emitter coefficients. This is achieved such that the differences between the simulated and field-observed values for pressure head and flow are minimized. The optimization problem is solved by using a genetic algorithm. A model that has already been calibrated at least according to threshold standards is necessary for this methodology. Two case studies are discussed in this paper including a real WDN example with artificially generated data, which investigated the limits of this method. The second case study is a real water system in the United Kingdom, where the method was implemented to detect a leak event that actually happened. The results suggest that leaks that cause a hydraulic impact larger than the sensor data error can be detected and localized with this method. The real case outcome shows that the presented method can reduce the search area for finding the leak to within 10% of the WDN (by length). The method can also contribute to more timely detection and localization of leakage hotspots, thus reducing economic and environmental impacts. The optimization model for predicting leakage hotspots can be effective despite the recognized challenges of model calibration and physical measurement limitations from the pressure and flow field tests.
科研通智能强力驱动
Strongly Powered by AbleSci AI