Physics-informed machine learning for advanced thermal storage-integrated energy networks, and sector-coupled flexibility services through collaborative business models

Title: Physics-informed machine learning for advanced thermal storage-integrated energy networks, and sector-coupled flexibility services through collaborative business models
Acronym: ThermFlex
Project leader: Qian Wang
Participating researcher:  Yangzhe Chen
Counterpart: RISE
Financed by: Swedish Energy Agency
Period: 2024-11-01 – 2028-12-31

Thermal storage technologies can bridge thermal networks, power grids, distributed renewables, and low-exergy resources. With the right system integration and operational logic, they can provide cross-sector energy flexibility services - increasing energy efficiency and sustainability of the overall energy system. Emerging advanced thermal storage solutions offer a variety of benefits, such as higher energy densities at lower temperatures and longer storage periods. However, their complex heat transfer phenomena necessitate new methods for estimating state-of-charge and power output - critical to reach market maturity. ThermFlex introduces novel data-driven modeling methods, including physics-informed machine learning, to address these challenges. ThermFlex will also explore new business and pricing models for flexibility services for district heating/cooling networks, and how to automate delivery of such flexibility services at scale leveraging ontologies.