A DATA-DRIVEN METHODOLOGY FOR DETERMINING THE CAUSES OF DISCOLOURATION IN DISTRIBUTION NETWORKS
A DATA-DRIVEN METHODOLOGY FOR DETERMINING THE CAUSES OF DISCOLOURATION IN DISTRIBUTION NETWORKS
复制标题
用于确定配电网络变色原因的数据驱动方法
DOI：
--
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
J. Boxall
中科院分区：
文献类型：
--
作者：
W. Furnass;S. Mounce;J. Boxall
变色水的供应是消费者和自来水公司都非常关心的问题。提出了一种数据驱动的、主要可自动化的方法，用于识别大量记录的变色事件的潜在原因。该案例不仅要考虑配水系统中可能原因（例如阀门操作、泵跳闸或维修）和变色事件的时间和空间平面分离，还要考虑因果关系的水力联系。说明了使用关系 GIS 数据库来整理和处理事件和水力数据，以及用于确定给定事件上游影响区域的水力回溯算法。对于特定的 WDS，该方法确定了一项或多项工作，1285 个记录的变色问题中有 18.0% 的问题，其中 12.1% 的问题不明确，而平面因果联系方法发现了 54.6% 的问题的可能原因，其中 30% 的问题不明确。这表明面向网络的方法识别出的误报关联较少。
The supplying of discoloured water is of significant concern to both consumers and water companies. A datadriven, primarily automatable method is presented for identifying the potential causes of large numbers of recorded discolouration events. The case is made for considering not just the temporal and spatio-planar separation of possible causes in a water distribution system (such as valve manipulation, pump trips or repairs) and discolouration events but also the hydraulic linking of cause and effect. The use of relational GIS databases for collating and processing event and hydraulic data is illustrated, as is a hydraulic back-tracing algorithm for determining the region of influence upstream of a given event. For a particular WDS the methodology identified one or more work jobs with 18.0% of 1285 recorded discoloration issues, 12.1% of which were ambiguous, whereas a planar causal linkage method found possible causes for 54.6% of the issues, 30% of which were ambiguous. This suggests that fewer false positive linkages are recognised by the network-orientated method.