Concentric Tube Heat Exchanger Benchmark Dataset: Experimental Measurements, CFD Configuration, and PINN Surrogate
Fuente:
Zenodo
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Recurso digital |
| Language: | English |
| Published: |
Zenodo
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866901555678019584 |
|---|---|
| author | DA COSTA, MAYSE COSTA, ÂNGELO LIRA, JOSÉ CLAUDINO |
| author_facet | DA COSTA, MAYSE COSTA, ÂNGELO LIRA, JOSÉ CLAUDINO |
| contents | <p>This dataset accompanies the manuscript "PINN surrogate for concentric tube heat exchangers: experimental validation and operational envelope mapping" submitted to the Journal of the Brazilian Society of Mechanical Sciences and Engineering (2026).</p> <p>It provides synchronized experimental measurements, CFD-derived results, and Physics-Informed Neural Network (PINN) predictions for a laboratory-scale concentric tube heat exchanger (TecQuipment TD360a) operating with water in both circuits. Eight experimental cases were measured, combining four cold-side mass flow rates (0.050, 0.033, 0.017, 0.008 kg/s) with parallel and counterflow configurations, while keeping the hot-side flow rate constant at 0.050 kg/s.</p> <p>The dataset includes:<br>- Experimental temperature and flow-rate measurements with calibrated instrumentation<br>- Moffat-propagated uncertainties for heat-transfer rates and effectiveness<br>- Energy balance analysis for the eight cases<br>- CFD validation results from ANSYS CFX with SST k-omega turbulence model and mesh-independence study<br>- Trained PINN surrogate (Jupyter notebook) with full reproducibility through fixed random seeds<br>- PINN predictions for the eight experimental cases<br>- Comparative analysis tables (Exp, CFD, PINN)</p> <p>The PINN surrogate, with architecture [5, 32, 32, 2] and 1,314 parameters, was trained against epsilon-NTU theoretical targets with energy-conservation and effectiveness-consistency constraints. It reproduces experimental outlet temperatures with mean absolute percentage errors of 0.98% (hot stream) and 3.00% (cold stream), and reduces computational cost by approximately one order of magnitude relative to CFD.</p> <p>The dataset is intended to support reproducibility of the published results and to serve as a reference benchmark for subsequent methodological studies on surrogate modeling of concentric tube heat exchangers.</p> <p> </p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_20112557 |
| institution | Zenodo |
| language | eng |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Concentric Tube Heat Exchanger Benchmark Dataset: Experimental Measurements, CFD Configuration, and PINN Surrogate DA COSTA, MAYSE COSTA, ÂNGELO LIRA, JOSÉ CLAUDINO Physics-informed neural networks Concentric tube heat exchanger Experimental validation Computational fluid dynamics CFD benchmarking Thermal effectiveness Surrogate model Heat transfer Uncertainty propagation Moffat method Open dataset <p>This dataset accompanies the manuscript "PINN surrogate for concentric tube heat exchangers: experimental validation and operational envelope mapping" submitted to the Journal of the Brazilian Society of Mechanical Sciences and Engineering (2026).</p> <p>It provides synchronized experimental measurements, CFD-derived results, and Physics-Informed Neural Network (PINN) predictions for a laboratory-scale concentric tube heat exchanger (TecQuipment TD360a) operating with water in both circuits. Eight experimental cases were measured, combining four cold-side mass flow rates (0.050, 0.033, 0.017, 0.008 kg/s) with parallel and counterflow configurations, while keeping the hot-side flow rate constant at 0.050 kg/s.</p> <p>The dataset includes:<br>- Experimental temperature and flow-rate measurements with calibrated instrumentation<br>- Moffat-propagated uncertainties for heat-transfer rates and effectiveness<br>- Energy balance analysis for the eight cases<br>- CFD validation results from ANSYS CFX with SST k-omega turbulence model and mesh-independence study<br>- Trained PINN surrogate (Jupyter notebook) with full reproducibility through fixed random seeds<br>- PINN predictions for the eight experimental cases<br>- Comparative analysis tables (Exp, CFD, PINN)</p> <p>The PINN surrogate, with architecture [5, 32, 32, 2] and 1,314 parameters, was trained against epsilon-NTU theoretical targets with energy-conservation and effectiveness-consistency constraints. It reproduces experimental outlet temperatures with mean absolute percentage errors of 0.98% (hot stream) and 3.00% (cold stream), and reduces computational cost by approximately one order of magnitude relative to CFD.</p> <p>The dataset is intended to support reproducibility of the published results and to serve as a reference benchmark for subsequent methodological studies on surrogate modeling of concentric tube heat exchangers.</p> <p> </p> |
| title | Concentric Tube Heat Exchanger Benchmark Dataset: Experimental Measurements, CFD Configuration, and PINN Surrogate |
| topic | Physics-informed neural networks Concentric tube heat exchanger Experimental validation Computational fluid dynamics CFD benchmarking Thermal effectiveness Surrogate model Heat transfer Uncertainty propagation Moffat method Open dataset |
| url | https://doi.org/10.5281/zenodo.20112557 |