Water researchSep 5, 2024

A detailed model to speed up water flow simulations in large pipe networks

Updated Mar 16, 2026

Abstract

A new framework can reduce hydraulic calculation time in large-scale water distribution networks by up to 51.93% compared to traditional methods.

The framework utilizes a joint topology-calculation decomposition method to streamline the hydraulic calculation process.
A high-performance decomposed gradient algorithm is integrated with parallel computation to enhance efficiency.
Tests conducted on four water distribution networks of varying sizes, ranging from 8 to 85,118 nodes, showed high calculation accuracy consistent with EPANET.
Factors influencing calculation acceleration include the level of decomposition and the consistency of sub-model sizes and structures.
The framework is designed to support the development of rapid-responding models for improved management of large-scale water distribution networks.

Simplified

As the size of water distribution network (WDN) models continues to grow, developing and applying real-time models or digital twins to simulate hydraulic behaviors in large-scale WDNs is becoming increasingly challenging. The long response time incurred when performing multiple hydraulic simulations in large-scale WDNs can no longer meet the current requirements for the efficient and real-time application of WDN models. To address this issue, there is a rising interest in accelerating hydraulic calculations in WDN models by integrating new model structures with abundant computational resources and mature parallel computing frameworks. This paper presents a novel and efficient framework for steady-state hydraulic calculations, comprising a joint topology-calculation decomposition method that decomposes the hydraulic calculation process and a high-performance decomposed gradient algorithm that integrates with parallel computation. Tests in four WDNs of different sizes with 8 to 85,118 nodes demonstrate that the framework maintains high calculation accuracy consistent with EPANET and can reduce calculation time by up to 51.93 % compared to EPANET in the largest WDN model. Further investigation found that factors affecting the acceleration include the decomposition level, consistency of sub-model sizes and sub-model structures. The framework aims to help develop rapid-responding models for large-scale WDNs and improve their efficiency in integrating multiple application algorithms, thereby supporting the water supply industry in achieving more adaptive and intelligent management of large-scale WDNs.

Full Text

Full text is available at the source.

View Full Text ↗

Funding

Competing interests

Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

PubMed

Abstract

Full Text

Funding

Competing interests

Related papers

What Lands in Your Inbox Each Week: