A Novel Water Pipeline Asset Management Scheme Using Hydraulic Monitoring Data
A Novel Water Pipeline Asset Management Scheme Using Hydraulic Monitoring Data
复制标题
利用水力监测数据的新型输水管道资产管理方案
DOI：
10.1061/9780784482506.020
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Momeni, Ahmad
中科院分区：
文献类型：
--
作者：
Piratla, Kalyan R.;Momeni, Ahmad
美国和其他几个国家的许多水管系统处于恶化状态，需要立即干预。这种干预措施面临的一个关键挑战是缺乏经济可靠的检查工具来评估管道资产的状况并确定其修复的优先次序。尽管在过去几年中开发和演示的创新检查技术很少，但由于检查成本高，这些技术仅用于数量有限的看似失败的资产。嵌入式传感和人工智能技术领域提供了独特的机会，可以根据水管系统的水力监测数据预测资产状况。本文提出了一种新的资产管理方案的水管道系统，管道流量和压力数据流通过SCADA系统被利用来推断不确定的资产参数，如减少管道直径和粗糙度值。为了证明所提出的方案，一个著名的配水网络在本文中使用的压力和流量数据监测在三个不同的位置，每个可以告知粗糙度值的所有管道系统中。首先，在一定的合理范围内随机降低选定的水网络中的所有管道的粗糙度系数，以表征现实世界的系统行为。随后，合成监测数据的管道压力和流量的生成使用水力模拟200种情况下，不同的节点需求。最后，开发了一种基于逆向工程方法的优化算法，利用合成监测数据预测管道粗糙度系数。所有200种情况下的建模和合成流量（和压力）数据之间的最小平方误差被认为是优化过程中的目标。MATLAB编程接口与EPANET水力建模工具结合使用。拟议的计划和本文件所载的结果将为资产管理提供一个有前途和强有力的办法，其中全系统的监测数据可以告知资产的状况，并随后支持优先修复。
Many water pipeline systems in the United States and several other countries are in a deteriorated state needing immediate intervention. One of the critical challenges to such interventions is the dearth of economical and reliable inspection tools to assess the condition of the pipeline assets and prioritize their rehabilitation. Although few innovative inspection techniques have been developed and demonstrated in the last few years, they are reserved to be used on a limited number of seemingly failing assets due to high inspection costs. The fields of embedded sensing and artificial intelligence techniques offer unique opportunities to predict asset conditions based on hydraulic monitoring data from water pipeline systems. This paper proposes a novel asset-management scheme for water pipeline systems where pipeline flow and pressure data streamed in through the SCADA systems are leveraged to deduce uncertain asset parameters such as reduced pipe diameters and roughness values. To demonstrate the proposed scheme, a well-known water distribution network is used in this paper to show that pressure and flow data monitored at three different locations each can inform the roughness values of all the pipelines in the system. Firstly, the roughness coefficients of all the pipelines in the chosen water network are randomly reduced within a certain reasonable range in order to characterize the real-world system behavior. Subsequently, synthetic monitoring data for pipeline pressure and flow is generated using hydraulic simulations for 200 scenarios with varied nodal demands. Finally, an optimization algorithm is developed based on a reverse engineering approach to predict pipeline roughness coefficients using the synthetic monitoring data. Least of the minimum squared error between the modeled and synthetic flow (and pressure) data for all the 200 scenarios are considered the objectives in the optimization process. MATLAB programming interface is used in this study in conjunction with EPANET hydraulic modeling tools. The proposed scheme and the included results of this paper will offer a promising and powerful approach towards asset management where system-wide monitoring data could inform the condition of the assets and subsequently support prioritization for rehabilitation.
影响因子：
3.5
作者：
M. Moglia;S. Burn;S. Meddings
通讯作者：
S. Meddings
DOI：
--
发表时间：
2005
期刊：
MIT International Conference on Information Quality
影响因子：
--
作者：
A. Koronios;Shien Lin;Jing Gao
通讯作者：
Jing Gao
工程资产管理的定义：回顾
DOI：
--
发表时间：
2015
期刊：
影响因子：
--
作者：
M.S. Somia Alfatih;M. Salman Leong;L. M. Hee
通讯作者：
L. M. Hee