IAHR Document Library

« Back to Library Homepage « Hydrolink 2021-2 Artificial Intelligence

[Ahead of Print] Machine learning for early stage piping design

Author(s): Ilker Telci

Linked Author(s): Ilker Telci

Keywords: Hydrolink; pipeline systems; Front-End Engineering Design (Pre-FEED); hydraulic transients; water hammer; Machine Learning (ML); water quality modeling; Artificial Neural Networks (ANN)


The design and construction of pipeline systems consist of several steps beginning from before Front-End Engineering Design (Pre-FEED) to Engineering-Procurement-Construction (EPC). Pre-FEED and FEED stages involve the initial conception and feasibility assessment of the systems. Their detailed design is finalized in the EPC stage of a project. In the early stages (e.g. Pre-FEED) of a project the main concern is the estimation of the project cost. These design stages have high impact on the sustainability, performance, and the cost of the final product [1]. Although the design is at a conceptual level at this stage, an accurate cost estimate is crucial for the success of the overall project. The cost of piping systems mainly depends on the pipe class (pressure capacity) to be used and the piping supports. The decision on the pipe class requires information on the expected maximum pressure in the pipe segments; the support capacity can be determined based on the expected maximum dynamic load the pipes may encounter.


Year: 2021

Copyright © 2024 International Association for Hydro-Environment Engineering and Research. All rights reserved. | Terms and Conditions