From Models To Experiments: Shallow Recurrent Decoder Networks on the DYNASTY Experimental Facility
Authors: Stefano Riva, Andrea Missaglia, Carolina Introini, J. Nathan Kutz, Antonio Cammi
Organizations: Politecnico di Milano, Deptartment of Energy, CeSNEF - Nuclear Engineering Division, 20156 Milan, Italy · Autodesk Research, 6 Agar Street, London UK · Emirates Nuclear Technology Center (ENTC), Department of Mechanical and Nuclear Engineering, Khalifa University, Abu Dhabi, 127788, United Arab Emirates
The Shallow Recurrent Decoder networks are a novel paradigm recently introduced for state estimation, combining sparse observations with high-dimensional model data. This architecture features important advantages compared to standard data-driven methods including: the ability to use only three sensors (even randomly selected) for reconstructing the entire dynamics of a physical system; the ability to train on compressed data spanned by a reduced basis; the ability to measure a single field variable (easy to measure) and reconstruct coupled spatio-temporal fields that are not observable and minimal hyper-parameter tuning. This approach has been verified on different test cases within different fields including nuclear reactors, even though an application to a real experimental facility, adopting the employment of in-situ observed quantities, is missing. This work aims to fill this gap by applying the Shallow Recurrent Decoder architecture to the DYNASTY facility, built at Politecnico di Milano, which studies the natural circulation established by internally heated fluids for Generation IV applications, especially in the case of Circulating Fuel reactors. The RELAP5 code is used to generate the high-fidelity data, and temperature measurements extracted by the facility are used as input for the state estimation. The results of this work will provide a validation of the Shallow Recurrent Decoder architecture to engineering systems, showing the capabilities of this approach to provide and accurate state estimation.
Figures & tables
Figure 1: SHRED architecture applied to the DYNASTY facility. Three out of the four available thermocouples are used to measure the temperature in the fluid T , at fixed locations. The sensors time series are used to construct a latent temporal sequence model which is mapped to the compressive representations of all spatio-temporal field variables. The compressive representations can then be mapped to the original state space by the Singular Value Decomposition (SVD).
Figure 2: DYNASTY natural circulation loop. The photograph on the left shows the facility installed at the Energy Laboratories of Politecnico di Milano, while the schematic on the right highlights the loop layout and the main components, including the heated sections and the upper air-cooled heat sink.
Figure 3: R5 nodalization of the DYNASTY facility used as a high-fidelity model for state estimation. The heated section corresponding to the VHHC-GV1 configuration is highlighted in red, while the upper air-cooled heat sink is shown in blue.
Figure 4: Division of the parametric space for training, validation and testing: the experimental case is always in the test set and in the middle of the parametric range.
Figure 5: Definition of the 4 sensor configurations investigated in this work: in each one, three thermocouples are used as input to reconstruct the full state and one is left out for validation, in addition to the mass flow rate meter (not displayed).
Figure 6: Decay of the singular values for the temperature field (left) and retained energy based on their cumulative sum (right).
Method
Configuration
Time (min)
Epochs
Time/Epoch (min)
Full
Unseen TC1
8.4324
76
0.1110
Full
Unseen TC2
7.5419
70
0.1077
Full
Unseen TC3
8.4453
78
0.1083
Full
Unseen TC4
10.2979
95
0.1084
Average
8.6794
79.75
0.1088
Std
1.1589
10.72
0.0014
Table 1: Training performance for full and compressive strategies, for a single SHRED model.
Figure 7: Relative errors of the SHRED reconstruction for the test set on each transient across the parametric space (bars), for both the full (red) and compressive (blue) training approaches. The dotted lines represents the average value of either the full or compressive approach over the different transients.
Figure 8: Reconstruction of the temperature field at final time for the HHHC scenario at the experimental test point, for the compressive (a) and full (b) training approaches. The first row shows the contour plots of the true state (first column) and the SHRED reconstructions (from the second column onwards), the second row shows the corresponding error fields and the first column of the third row shows the temperature evolution over time for the true state and the SHRED reconstructions at the TC locations.
Figure 9: Comparison of the mass flow rate time series for all the scenarios at the experimental test point, between the R5 simulation and the SHRED reconstructions obtained with the compressive (first row) and full (second row) training approaches. This includes also the ensemble SHRED reconstruction.
Configuration
HHHC
DH
VHHC-GV1
VHHC-GV2
Avg
Std
FOM
0.127
0.143
0.069
0.053
0.098
0.044
Unseen TC1 – Full
0.056
0.057
0.120
0.099
0.083
0.032
Unseen TC1 – Comp.
0.057
0.058
0.115
0.098
0.082
0.029
Unseen TC2 – Full
0.038
0.070
0.054
0.125
0.072
0.038
Unseen TC2 – Comp.
0.045
0.087
0.120
0.128
0.095
0.038
Unseen TC3 – Full
0.061
0.057
0.094
0.062
0.069
0.017
Table 2: Relative RMSE of the FOM and SHRED reconstructions for the experimental test point, for both the full and comp. training approaches. The ensemble SHRED reconstruction is also included. The last two columns show the average and standard deviation across sensors.
Figure 10: Comparison of the TC measurements for the HHHC scenario at the experimental test point, between the experimental data, the FOM and the SHRED reconstructions obtained with the compressive (a) and full (b) training approaches. This includes also the ensemble SHRED reconstruction.
Figure 11: Comparison of the TC measurements for the VHHC-GV1 scenario at the experimental test point, between the experimental data, the FOM and the SHRED reconstructions obtained with the compressive (a) and full (b) training approaches. This includes also the ensemble SHRED reconstruction.
Figure 12: Comparison of the mass flow rate time series for all the scenarios at the experimental test point, between the experimental data, the FOM simulation and the SHRED reconstructions obtained with the full (first column) and compressive (second column) training approaches.
State estimation in magnetohydrodynamic flows is critical for real-time monitoring of liquid metal blankets in tokamak fusion reactors. Due to the multiphysics nature of these phenomena, high-fidelity simulations are computationally prohibitive for real-time applications. This work investigates a data- driven Reduced Order Model framework: the Shallow Recurrent Decoder (SHRED) coupled with Principal Component Analysis, to map sparse temperature measurements to the full thermo-hydraulic system's state. The major contribution of this work lies in the two-parameter analysis of a fully three-dimensional domain representative of the DEMO breeding blanket configuration. Here, the flow is subjected to an external magnetic field varying in direction and intensity and is hindered by two cylinders acting as a water-cooling system, which impose a temperature boundary condition on their surfaces. This double-parametric magnetic variation induces nonlinear transitions in the flow dynamics, ranging from chaotic behavior at low magnetic field intensities to laminarized regimes at high intensities, characterized by the formation of asymmetric side layers at an inclination angle of 30 degrees. SHRED reconstruction maintains a mean relative error of approximately 5% for the temperature, pressure, and velocity fields. This accuracy is maintained across both weak and strong magnetic fields, ranging from 0.075 T to 0.300 T, and for inclination angles from 5 to 30 degrees, reflecting its dominant toroidal component. These errors are only slightly larger than the lower error bound dictated by low-rank truncation. The results establish SHRED as a reliable state estimator for complex and realistic engineering applications involving completely unseen parametric scenarios and validate it as an accurate real-time state estimation technique suitable for online monitoring and control of real facilities.
Energy Department, CeSNEF - Nuclear Engineering Division, Politecnico di Milano, Milano · Autodesk Research, 6 Agar Street, London, WC2R 0QE, UK · Argonne National Laboratory, Lemont, IL, 60439, United States of America +1
Reconstructing spatially resolved plasma dynamics from few sensors is essential for diagnostics, reduced-order modelling and control, yet remains difficult because the sparse measurements incompletely constrain multiscale, regime-dependent degrees of freedom. The Shallow Recurrent Decoder (SHRED) partially addresses spatial sparsity by using measurement histories; however, its fully connected decoder provides no explicit mechanism for resolving spatial structure across scales or explicit parametric dependency. We introduce the Recurrent Multiscale Affine-modulated Inference Network (ReMAIN), which preserves SHRED's recurrent temporal encoding but replaces its decoder with a U-Net whose feature hierarchy is conditioned by the recurrent state through feature-wise linear modulation. The temporal representation supplies both a dense prior and scale-specific modulation throughout the U-Net. A parametric extension jointly embeds the operating condition and sensor history, enabling reconstruction to adapt as the governing dynamics change with operating regime. ReMAIN is first benchmarked against SHRED on six one-dimensional nonlinear PDEs representing diverse dynamics. Across all benchmarks, it reduces reconstruction errors on unseen trajectories and more faithfully resolves sharp transitions, localized extrema and fine-scale variations. The parameter-conditioned model is then demonstrated on a collisionless E×B plasma subject to perpendicular axial electric and radial magnetic fields, with the electric-field strength serving as the operating parameter. ReMAIN reconstructs the high-dimensional, multiscale plasma state and recovers its regime-dependent spatiotemporal dynamics at electric-field strengths withheld from training. Together, ReMAIN improves sparse-sensor full-state reconstruction and, through parameter conditioning, generalizes across plasma operating regimes.
Maryam Reza, Farbod Faraji
Independent Researcher · Department of Computing, Huxley Building, Imperial College London, London SW7 2RH, United Kingdom
In reactor physics, neutronics can be treated with different fidelity levels, according to the needs of the user. On one hand, the precise modeling of neutrons' behaviour in reactor physics is often expensive and time-consuming due to the high computational costs to numerically solve the Boltzmann transport equation. Conversely, by adopting suitable assumptions, such as the SPN, diffusion theory, and point kinetics, it is possible to generate efficiently low-fidelity data. From the perspective of surrogate models, this computational limitation translates into a scarcity of high-fidelity data and a significant amount of low-fidelity data. Given this difference in fidelity levels, it would be interesting to develop a suitable procedure to map low-fidelity models towards higher fidelity models; for instance, one could obtain the solution to a multi-group diffusion equation starting from time-series data obtained from a point kinetics model. Indeed, this work investigates this possibility by leveraging multi-fidelity information with Shallow Recurrent Decoders, a novel machine learning architecture able to map time-series observations to the full state of the reactor. This technique has been designed to use local or global measurements as input and map their temporal trajectories to the high-dimensional state; by the same logic, in principle, this architecture can also be used when the input is formed by the solution of a lumped model. This work applies this idea to a benchmark reactor geometry, mapping the point kinetics model to the diffusion solution under various input conditions, with much less computational costs.
Stefano Riva, Carolina Introini, J. Nathan Kutz +1
Autodesk Research, London, UK · Department of Energy, Nuclear Engineering Division, Politecnico di Milano, Milan, Italy · Department of Mechanical and Nuclear Engineering and Emirates Nuclear Technology Center, Khalifa University, Abu Dhabi (127788), United Arab Emirates