Non-deposition self-cleansing models for large sewer pipes

Carlos Montes, Sergio Vanegas, Zoran Kapelan, Luigi Berardi, Juan Saldarriaga

Research output: Contribution to journalArticleScientificpeer-review

11 Citations (Scopus)
14 Downloads (Pure)


Multiple models from the literature and experimental datasets have been developed and collected to predict sediment transport in sewers. However, all these models were developed for smaller sewer pipes, i.e. using experimental data collected on pipes with diameters smaller than 500 mm. To address this issue, new experimental data were collected on a larger, 595 mm pipe located in a laboratory at the University of los Andes. Two new self-cleansing models were developed using this dataset. Both models predict the sewer self-cleansing velocity for the cases of non-deposition with and without deposited bed. The newly developed and existing models were then evaluated and compared on the basis of the most recently collected and previously published datasets. Models were compared in terms of prediction accuracy measured by the root mean squared error and mean absolute percentage error. The results obtained show that in the existing literature, self-cleansing models tend to be overfitted, i.e. have a rather high prediction accuracy when applied to the data collected by the authors, but this accuracy deteriorates quickly when applied to the datasets collected by other authors. The newly developed models can be used for designing both small and large sewer pipes with and without deposited bed condition.

Original languageEnglish
Pages (from-to)606-621
Number of pages16
JournalWater science and technology : a journal of the International Association on Water Pollution Research
Issue number3
Publication statusPublished - 2020

Bibliographical note

Accepted Author Manuscript


  • Bedload
  • Deposited bed
  • Non-deposition
  • Sediment transport
  • Self-cleansing


Dive into the research topics of 'Non-deposition self-cleansing models for large sewer pipes'. Together they form a unique fingerprint.

Cite this