Calibration of friction and roughness in an urban water distribution network using an LSTM neural network-based framework

Abstract

© 2025 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

This paper proposes a method to estimate both the friction factor and the roughness coefficient of the pipes in a water distribution network (WDN). Pressure head data is used for the training of a Long-Short Term Memory (LSTM) neural network to predict the pressure head at all of the other nodes in the network, as well as the flow rate through the pipes. When all the data is predicted, the estimated pressure head drops, and the flow rate for the entire water distribution network is used to compute both the friction factor and the flow rate. The proposed method is tested using the well-known Hanoi DMA (District Metered Area) as a case study. Satisfactory results are obtained when comparing the predicted friction factor and roughness coefficient against the expected values.

The authors gratefully acknowledge the financial support provided by the Tecnologico Nacional de M ´ exico and SECIHTI (Mexico). The authors also ácknowledge the support provided by SECTEI-CDMX through project eSAST, number 1564c23, DGAPA-UNAM through project IT1000724, and by the Spanish project SEAMLESS: Sustainable learning-based Management of Multi-resource Large-scale Systems (ref. PID2023-148840OB-I00), funded by MCIN/AEI/10.13039/501100011033/FEDER, UE.

Peer Reviewed

Postprint (author's final draft)