Periodic Reporting for period 1 - SHINE (Multi-level approach for the up-scaling of ultra high temperature energy storage and conversion)
Berichtszeitraum: 2024-06-01 bis 2026-05-31
Europe’s transition to a climate-neutral energy system requires new ways to store and deliver renewable energy on demand. Many industrial sectors and power applications operate at very high temperatures (>1000oC) and cannot rely solely on electricity from intermittent sources, such as solar and wind. Conventional batteries are unsuitable for these conditions, while existing thermal storage systems often lack efficiency and flexibility.
SHINE addressed this challenge by exploring latent-heat thermophotovoltaic (LHTPV) batteries, an emerging ultra-high-temperature energy storage technology. These systems store thermal heat in metallic-based phase-change materials and convert it directly into electricity using solid-state converters, i.e. thermophotovoltaic (TPV) converters in this case. They have the potential to support dispatchable renewable power, improve energy efficiency and contribute to the decarbonisation of energy-intensive industries, aligning strongly with the goals of the European Green Deal. Their applications can be various from waste heat recovery, to power-to-heat-to-power and solar-to-heat-to-power concepts, with the last two being investigated within SHINE.
Before such systems can be deployed, however, the scientific community needs accurate yet computationally efficient modelling tools capable of predicting their behaviour under real operating conditions, as well as accessible datasets that support transparency and further innovation. SHINE was therefore designed to fill a critical knowledge gap by developing advanced multi-scale numerical models for the battery two key components (storage system and converter), public datasets and a computationally efficient reduced-order model capable of simulating the full LHTPV battery. The produced-order model has retrieved information from full order models leading to next generation simulation tools.
SHINE overall objectives were:
1. To build and validate through in-house experiments full order models (FOMs) of the system’s two key components: the latent-heat thermal storage unit (thermal model) and the TPV converter (electric model);
2. To generate and publish numerical datasets using the FOMs;
3. To generate Artificial Neural Networks (ANNs)-based multivariable polynomial functions predicting the sub-components performance, using data from the produced datasets;
4. To integrate the models into a fast, simplified in-house simulation platform suitable for optimisation, system análisis, measured data post-processing and clean-up and future integration with experimental measurements;
5. To study the dynamic behaviour of the LHTPV battery under different charging and discharging conditions.
The expected impacts include:
• Improved scientific understanding of ultra-high-temperature energy storage and conversion;
• A new modelling paradigm of using AI to feed ROMs without losing information from rigorous models;
• The importance of deploying multi-scale, multi-disciplinary research for next-generation energy storage and covnersion technologies;
• Generation of new tools for supporting system design, optimisation and monitoring;
• Data sets and in-house codes that enable future digital twins and improved control strategies (posible integration into open and closed-loop systems);
• Long-term benefits for renewable energy integration in industrial and power sectors.
While SHINE is a technical project, it also contributes to broader societal needs by supporting the transition toward sustainable and secure energy systems. Through open data, communication actions and training activities, the project has encouraged knowledge sharing across both industrial and academic stakeholders.
SHINE addressed this challenge by exploring latent-heat thermophotovoltaic (LHTPV) batteries, an emerging ultra-high-temperature energy storage technology. These systems store thermal heat in metallic-based phase-change materials and convert it directly into electricity using solid-state converters, i.e. thermophotovoltaic (TPV) converters in this case. They have the potential to support dispatchable renewable power, improve energy efficiency and contribute to the decarbonisation of energy-intensive industries, aligning strongly with the goals of the European Green Deal. Their applications can be various from waste heat recovery, to power-to-heat-to-power and solar-to-heat-to-power concepts, with the last two being investigated within SHINE.
Before such systems can be deployed, however, the scientific community needs accurate yet computationally efficient modelling tools capable of predicting their behaviour under real operating conditions, as well as accessible datasets that support transparency and further innovation. SHINE was therefore designed to fill a critical knowledge gap by developing advanced multi-scale numerical models for the battery two key components (storage system and converter), public datasets and a computationally efficient reduced-order model capable of simulating the full LHTPV battery. The produced-order model has retrieved information from full order models leading to next generation simulation tools.
SHINE overall objectives were:
1. To build and validate through in-house experiments full order models (FOMs) of the system’s two key components: the latent-heat thermal storage unit (thermal model) and the TPV converter (electric model);
2. To generate and publish numerical datasets using the FOMs;
3. To generate Artificial Neural Networks (ANNs)-based multivariable polynomial functions predicting the sub-components performance, using data from the produced datasets;
4. To integrate the models into a fast, simplified in-house simulation platform suitable for optimisation, system análisis, measured data post-processing and clean-up and future integration with experimental measurements;
5. To study the dynamic behaviour of the LHTPV battery under different charging and discharging conditions.
The expected impacts include:
• Improved scientific understanding of ultra-high-temperature energy storage and conversion;
• A new modelling paradigm of using AI to feed ROMs without losing information from rigorous models;
• The importance of deploying multi-scale, multi-disciplinary research for next-generation energy storage and covnersion technologies;
• Generation of new tools for supporting system design, optimisation and monitoring;
• Data sets and in-house codes that enable future digital twins and improved control strategies (posible integration into open and closed-loop systems);
• Long-term benefits for renewable energy integration in industrial and power sectors.
While SHINE is a technical project, it also contributes to broader societal needs by supporting the transition toward sustainable and secure energy systems. Through open data, communication actions and training activities, the project has encouraged knowledge sharing across both industrial and academic stakeholders.
SHINE developed a multi-level modelling framework for ultra-high-temperature latent-heat thermophotovoltaic batteries. The work included thermal, optical and electrical modelling, multi-physics simulations and dynamic system-level analisis using in-house codes in MatlabTM.
-Development of component models:
A detailed model of the latent-heat storage unit was built using computational fluid dynamics, representing phase change and natural convection in high-temperature conditions. In parallel, a thermophotovoltaic model was developed in two stages: a validated 0-dimensional version for initial analyses and a more advanced 1-dimensional drift–diffusion solver, which was significantly improved during the project.
-Parametric simulations and database creation:
Extensive simulations were carried out for two system concepts (S2H2P and P2H2P). The results were integrated into databases (MySQL and EXCEL) and later published in open repositories with versioning. The datasets include temperature fields, melt fractions, system efficiencies and other variables required for future AI-based surrogate models.
-Reduced-order model (ROM):
A fast 0D/1D reduced-order model combining the thermal and electrical sub-models was developed and verified against the FOMs. This ROM enabled rapid calculations, dynamic simulations, sensitivity analyses and optimisation. It was also used as the basis for a coupled thermal–technoeconomic tool developed in TRNSYS for assessing the LHTPV battery at system level.
-Scientific advances and dissemination:
The project generated an open dataset and in-house codes some of which were made open to the public. A methodology was also demonstrated to estimate the state of charge of latent-heat storage systems using only temperature and heat-flux signals, representing an important scientific innovation for future coupling of real-time experiments with the model and closed-loop control systems.
Despite the early ending of the fellowship, all major technical objectives were advanced, and the ROM-based framework provides a strong foundation for future work, including the planned incorporation of artificial neural networks.
-Development of component models:
A detailed model of the latent-heat storage unit was built using computational fluid dynamics, representing phase change and natural convection in high-temperature conditions. In parallel, a thermophotovoltaic model was developed in two stages: a validated 0-dimensional version for initial analyses and a more advanced 1-dimensional drift–diffusion solver, which was significantly improved during the project.
-Parametric simulations and database creation:
Extensive simulations were carried out for two system concepts (S2H2P and P2H2P). The results were integrated into databases (MySQL and EXCEL) and later published in open repositories with versioning. The datasets include temperature fields, melt fractions, system efficiencies and other variables required for future AI-based surrogate models.
-Reduced-order model (ROM):
A fast 0D/1D reduced-order model combining the thermal and electrical sub-models was developed and verified against the FOMs. This ROM enabled rapid calculations, dynamic simulations, sensitivity analyses and optimisation. It was also used as the basis for a coupled thermal–technoeconomic tool developed in TRNSYS for assessing the LHTPV battery at system level.
-Scientific advances and dissemination:
The project generated an open dataset and in-house codes some of which were made open to the public. A methodology was also demonstrated to estimate the state of charge of latent-heat storage systems using only temperature and heat-flux signals, representing an important scientific innovation for future coupling of real-time experiments with the model and closed-loop control systems.
Despite the early ending of the fellowship, all major technical objectives were advanced, and the ROM-based framework provides a strong foundation for future work, including the planned incorporation of artificial neural networks.
SHINE delivered several advances that go beyond current knowledge in ultra-high-temperature thermal storage and TPV energy conversion.
- New modelling capabilities:
The project developed one of the first integrated modelling frameworks for LHTPV batteries, covering detailed CFD, TPV modelling and reduced-order simulation. The ability to capture phase-change behaviour, radiative effects and electrical conversion within a unified approach represents a clear step forward.
- Novel method for system characterisation:
A key scientific result is the demonstration that the state of charge of a latent-heat storage system can be determined solely from measured temperature and heat-flux signals. This finding opens new possibilities for real-time monitoring, digital twins and model-based control of high-temperature storage systems.
- Public, FAIR-compliant datasets:
The project provides openly accessible and versioned datasets that can be reused by the community, supporting transparency and reducing the need for repeated numerical campaigns.
- Foundations for future AI-based surrogates:
Although full ANN-based models were not completed during the project, SHINE prepared the data structures and in-house algorithms codes required for future surrogate modelling. These will help reduce computational costs and improve prediction accuracy.
-Pathways to uptake:
To advance further towards industrial deployment, several steps are identified, including:
1. completing experimental validation with in-house lab tests (TRL3-4) and upcoming demonstrators (TRL5-6);
2. incorporating ANN-based models into the ROM;
3. full techno-economic assessments using realistc data;
4. collaboration with industry for pilot-scale testing;
5. development of an accessible front-end interface for wider use of the tools;
6. extensión of the model to other types of thermal storage systems (sensible, themochemical) and solid-state converters (e.g. hybrid thermionic-thermophotovoltaic converters)
To advance further towards research and innovation the following steps are identified:
1. inclusión of detailed phenomena for the phase-change process under various scales;
2. simulation of the phase change interface with higher fidelity models
3. coupling of the model with real-time data in AI-assisted control algorithms.
These actions will support commercialisation pathways, regulatory discussions and industrial-scale feasibility analyses.
- New modelling capabilities:
The project developed one of the first integrated modelling frameworks for LHTPV batteries, covering detailed CFD, TPV modelling and reduced-order simulation. The ability to capture phase-change behaviour, radiative effects and electrical conversion within a unified approach represents a clear step forward.
- Novel method for system characterisation:
A key scientific result is the demonstration that the state of charge of a latent-heat storage system can be determined solely from measured temperature and heat-flux signals. This finding opens new possibilities for real-time monitoring, digital twins and model-based control of high-temperature storage systems.
- Public, FAIR-compliant datasets:
The project provides openly accessible and versioned datasets that can be reused by the community, supporting transparency and reducing the need for repeated numerical campaigns.
- Foundations for future AI-based surrogates:
Although full ANN-based models were not completed during the project, SHINE prepared the data structures and in-house algorithms codes required for future surrogate modelling. These will help reduce computational costs and improve prediction accuracy.
-Pathways to uptake:
To advance further towards industrial deployment, several steps are identified, including:
1. completing experimental validation with in-house lab tests (TRL3-4) and upcoming demonstrators (TRL5-6);
2. incorporating ANN-based models into the ROM;
3. full techno-economic assessments using realistc data;
4. collaboration with industry for pilot-scale testing;
5. development of an accessible front-end interface for wider use of the tools;
6. extensión of the model to other types of thermal storage systems (sensible, themochemical) and solid-state converters (e.g. hybrid thermionic-thermophotovoltaic converters)
To advance further towards research and innovation the following steps are identified:
1. inclusión of detailed phenomena for the phase-change process under various scales;
2. simulation of the phase change interface with higher fidelity models
3. coupling of the model with real-time data in AI-assisted control algorithms.
These actions will support commercialisation pathways, regulatory discussions and industrial-scale feasibility analyses.