Pipe networks for both water and oil distribution are prone to the formation of restrictions and, if not managed, possible obstructions. These reduce the efficiency of pipe systems and, in turn, cause negative economic impacts, temporary losses of service, and environmental risks. The present work focuses on a noninvasive methodology for the detection of restrictions in pipe networks. Restrictions are identified by minimizing, via genetic algorithms, a function that represents the discrepancy between on-field measured data and those simulated numerically. Measured data consist of a limited set of steady-state pressure heads and flow rates, which are the most commonly available information for pipe networks. The outcome of the technique is the “equivalent residual diameter” of each pipe in the network. This parameter allows the company managing the pipe network to identify the pipe segments where restrictions are most likely to be present and require further investigations. The approach is numerically validated for 15 different scenarios, considering five different sets of available measures and three different restriction conditions, in a mixed branched-looped network with complex topology for crude-oil transportation. The results show that the presence of restrictions is clearly identified and their magnitude is generally assessed with an accuracy of 5%.

中文翻译：

---

使用不完整和嘈杂的流量和压力稳态测量来检测管网中的限制和障碍物

用于输水和输油的管网容易形成限制，如果不加以管理，可能会形成障碍。这些会降低管道系统的效率，进而导致负面的经济影响、暂时的服务损失和环境风险。目前的工作重点是一种用于检测管网限制的非侵入性方法。通过遗传算法最小化表示现场测量数据与数值模拟数据之间差异的函数来识别限制。测量数据包括一组有限的稳态压头和流速，这是管网最常用的信息。该技术的结果是网络中每条管道的“等效剩余直径”。此参数允许管理管网的公司识别最有可能存在限制并需要进一步调查的管段。该方法在 15 种不同情况下进行了数值验证，考虑了五组不同的可用措施和三种不同的限制条件，在具有复杂原油运输拓扑的混合分支回路网络中。结果表明，限制的存在被清楚地识别出来，其大小的评估精度通常为 5 在具有复杂拓扑结构的混合分支回路网络中，用于原油运输。结果表明，限制的存在被清楚地识别出来，其大小的评估精度通常为 5 在具有复杂拓扑结构的混合分支回路网络中，用于原油运输。结果表明，限制的存在被清楚地识别出来，其大小的评估精度通常为 5% .