Development of models for prediction of water distribution pipeline failures for isolated settlement using artificial neural networks
by Paul Amaechi Ozor; Solomon Onyekachukwu Onyedeke; Charles Mbohwa
International Journal of Technology, Policy and Management (IJTPM), Vol. 22, No. 4, 2022

Abstract: The development of artificial neural network (ANN) models for the prediction of failure of water pipelines in rural communities is undertaken. The research use analytical procedures. This approach was illustrated using a 10-year primary pipelines failure data taken from a typical isolated settlement. Four ANN models were established; three models based on specific material type and a combinatorial model. Validation of the models on unseen data gave impressive results, on the basis of the performance indices. The sensitivity analysis shows that the pipe thickness impacts the model output most, whereas the installation year had the least relative importance in predicting pipeline failure. The models predicted failure parameters appreciably, and determined the benefit index, which allows the elaboration of a strategy for achieving realistic maintenance policies. The study shows that the ANN supported with EPANET software can provide a good means of analysing pipeline failures and optimising the maintenance function.

Online publication date: Thu, 13-Oct-2022

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Technology, Policy and Management (IJTPM):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com