Researchers create a digital twin to run DAWN, Synhelion’s solar fuels plant
A solar research team at the Solar-Institute Jülich (SIJ) of FH Aachen, in collaboration with Synhelion engineers, presented two papers at SolarPACES 2025, assessing the accuracy and reliability of dynamic process models for a solar fuel production system. They utilized operating data from Synhelion’s DAWN pilot plant.
Synhelion’s solar fuel plant, DAWN, employs concentrated solar power to drive endothermic chemical reactions, converting biogenic methane, carbon dioxide, and steam into synthesis gas (syngas). This syngas can be processed into liquid hydrocarbon fuels, such as kerosene for aviation or gasoline.
By utilizing bio-waste-derived methane and carbon, and generating heat from concentrated solar heat, Synhelion is pioneering a sustainable drop-in replacement for conventional liquid fuels.
The DAWN pilot plant in Jülich, which will be open for tours at the next SolarPACES conference in the fall of 2026, utilizes RED-certified sustainable biogenic waste as its methane and carbon source. It has demonstrated the complete production chain under real solar operation conditions.
Synhelion’s solar-absorbing gas receiver is a unique design capable of converting solar irradiance from heliostats to temperatures as high as 1500°C.
To achieve real-time simulation, the team took a new approach, recognizing that traditional physics-based models can be fast but may sacrifice accuracy. They needed a system to offer simulation results in real-time to aid plant operators in decision-making.
The team created a machine-learning model of the thermal energy storage using synthetic data from the physics-based model, successfully validating it. In the overall system study, all three key components—the solar receiver, the reforming reactor, and the thermal energy storage—were validated against real-time operation.
The thermal energy storage model successfully reproduced charge-discharge transients over 8 cycles across 10 days, achieving errors of less than 3% of the temperature swing in the upper storage layers. However, deviations increased toward the bottom due to potential material property variations and mass flow measurement uncertainties.
The surrogate model yielded results in milliseconds on synthetic validation data, compared to an average of over five seconds for a physics-based model, making it 100 to 1000 times faster. On operational validation data from DAWN, it was up to 50,000 times faster.
Since solar flux onto the receiver wasn’t directly measured during operation, the researchers developed a simplified data-driven surrogate solar field model using a 4th-degree polynomial with ridge regression.
The model validation for the reforming reactor revealed major discrepancies, especially under partial-load conditions, where the model overestimated the reaction rate and heat consumption, with residual methane present, contrary to equilibrium predictions, and lower CO₂ conversion than expected.
The team continues to work on the solar receiver and the reforming reactor, addressing issues with the heliostat field model and the reactor kinetics. They are also exploring the use of machine-learning models to accurately reproduce the chemistry.
The long-term goal is to develop a functioning digital twin that enables real-time plant monitoring, predictive control, and virtual testing of control strategies before implementation on an actual solar fuels plant’s system.
