CUED Publications database

Optimising the design of energy piles using machine learning

Makasis, N and Narsilio, GA and Bidarmaghz, A (2017) Optimising the design of energy piles using machine learning. In: 19th International Conference on Soil Mechanics and Geotechnical Engineering, 2017-9-17 to 2017-9-22 pp. 3455-3458..

Full text not available from this repository.


© 2017 19th ICSMGE Secretariat. All rights reserved. Closed loop ground-source heat pump (GSHP) systems can efficiently provide clean and renewable energy for heating and cooling purposes, by using the ground as a heat source/sink. One interesting application is to implement ground heat exchangers (GHEs) directly into building foundations (e.g., energy piles), thus eliminating the highest cost associated with GSHP technology (i.e., drilling). However, this application introduces certain design constrains, including limiting the ability to optimise typical geothermal design parameters, such as GHE separation and depth, and therefore the provision of all required thermal demand may not be guaranteed. Thus, a different design approach is required, one that maximises the amount of energy that can be provided by GSHP systems. This article introduces a new methodology for optimising this amount of energy, using machine learning techniques. The approach is based on 'learning' the relationship between the thermal load and the fluid temperatures in the GHEs, and creating a model that quickly estimates the performance of the GSHP system, for any thermal load. The new approach can be used alongside a more complex numerical simulation (e.g. FEM) and it can significantly speed up the design optimisation process.

Item Type: Conference or Workshop Item (UNSPECIFIED)
Divisions: Div D > Structures
Depositing User: Cron Job
Date Deposited: 03 May 2018 01:32
Last Modified: 13 Apr 2021 07:08