Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Predicting the Timing of Water Main Failure Using Artificial Neural Networks
Effective management of aging water distribution infrastructure is essential for preserving the economic vitality of North American municipalities. Historical failures within Scarborough, Ontario, Canada, reveal a seasonal pattern to water main failures, with the majority of failures occurring during the very cold winter months. Extensive installation of cement mortar lining and cathodic protection have extended the life span of aging water mains and reduced escalating failure rates. Artificial neural networks are found to be capable of predicting the time to failure for individual pipes using a range of pipe-specific attributes, including diameter, length, soil type, construction year, and the number of previous failures. The developed models have correlation coefficients ranging from 0.70–0.82 on instances reserved for evaluating predictive performance and have utility on an asset-by-asset basis when planning water main inspection, maintenance, and rehabilitation. Simulated failure scenarios indicate a return to high failure rates if cement mortar lining and cathodic protection are not extended to all candidate pipes in the distribution system.
Predicting the Timing of Water Main Failure Using Artificial Neural Networks
Effective management of aging water distribution infrastructure is essential for preserving the economic vitality of North American municipalities. Historical failures within Scarborough, Ontario, Canada, reveal a seasonal pattern to water main failures, with the majority of failures occurring during the very cold winter months. Extensive installation of cement mortar lining and cathodic protection have extended the life span of aging water mains and reduced escalating failure rates. Artificial neural networks are found to be capable of predicting the time to failure for individual pipes using a range of pipe-specific attributes, including diameter, length, soil type, construction year, and the number of previous failures. The developed models have correlation coefficients ranging from 0.70–0.82 on instances reserved for evaluating predictive performance and have utility on an asset-by-asset basis when planning water main inspection, maintenance, and rehabilitation. Simulated failure scenarios indicate a return to high failure rates if cement mortar lining and cathodic protection are not extended to all candidate pipes in the distribution system.
Predicting the Timing of Water Main Failure Using Artificial Neural Networks
Harvey, Richard (Autor:in) / McBean, Edward A. (Autor:in) / Gharabaghi, Bahram (Autor:in)
Journal of Water Resources Planning and Management ; 140 ; 425-434
07.02.2013
102013-01-01 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Unbekannt
Predicting the Timing of Water Main Failure Using Artificial Neural Networks
| British Library Online Contents | 2014
Predicting the Timing of Water Main Failure Using Artificial Neural Networks
| Online Contents | 2014
Predicting Water Levels Using Artificial Neural Networks
| British Library Conference Proceedings | 2003
Forecasting Water Main Failure Using Artificial Neural Network and Generalized Linear Models
| British Library Conference Proceedings | 2013
Predicting the Caspian Sea surface water level using artificial neural networks
| British Library Conference Proceedings | 1995