osman-applicationofneural-2006.pdf (1.76 MB)
Application of neural networks in modelling serviceability deterioration of concrete stormwater pipes
conference contribution
posted on 2006-01-01, 00:00 authored by A Ng, D Tran, N Osman, K McManusStormwater pipe systems in Australia are designed to convey water from rainfall and surface runoff only and do not transport sewage. Any blockage can cause flooding events with the probability of subsequent property damage. Proactive maintenance plans that can enhance their serviceability need to be developed based on a sound deterioration model. This paper uses a neural network (NN) approach to model deterioration in serviceability of concrete stormwater pipes, which make up the bulk of the stormwater network in Australia. System condition data was collected using CCTV images. The outcomes of model are the identification of the significant factors influencing the serviceability deterioration and the forecasting of the change of serviceability condition over time for individual pipes based on the pipe attributes. The proposed method is validated and compared with multiple discriminant analysis, a traditionally statistical method. The results show that the NN model can be applied to forecasting serviceability deterioration. However, further improvements in data collection and condition grading schemes should be carried out to increase the prediction accuracy of the NN model.
History
Event
Proceedings of the 7th WSEAS International Conference on Neural Networks (2006 : Cavtat, Croatia)Pagination
53 - 60Publisher
WSEASLocation
Cavtat, CroatiaPlace of publication
[Cavtat, Croatia]Start date
2006-06-12End date
2006-06-15Language
engNotes
(Won Best Paper Award based on best originality and scientific impact, good presentation (presented by Osman, N.Y) and paper presented by student) Reproduced with the kind permission of the copyright ownerPublication classification
E1.1 Full written paper - refereedCopyright notice
2006, WSEASTitle of proceedings
NN'06 : Proceedings of the 7th WSEAS International Conference on Neural NetworksUsage metrics
Categories
No categories selectedLicence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC