Transfer learning-based deep neural network model for performance prediction of hydrogen-fueled solid oxide fuel cells
Bibliographic record
Abstract
Transfer learning (TL) is an effective method for minimizing modeling efforts and data requirements for diverse energy systems. This paper presents use of TL for different hydrogen-fueled solid oxide fuel cells (SOFCs) types (tubular vs planar) with different manufacturers (UAlberta vs Elcogen) and output power ranges. Three different single tubular cells were fabricated and tested under 18 operating conditions. In addition, a planar single cell was tested under 10 operating conditions. The data gathered from the first tubular cell (A1) was used to train a deep neural network (DNN) model with the cell voltage as the output. Then, the DNN model was transferred from this source domain to three different target domains. Two tubular cells with different cell properties such as electrolyte thickness, electrodes’ thickness and porosity, and one planar cell with different microstructural properties and different physical layout were the target domains. Fine-tuning was used for TL, and the effect of different normalization strategies and different amounts of fine-tuning data were compared. The developed DNN was able to capture the nonlinear part in current density–voltage (J–V) curves from the rich dataset available for training. The DNN model trained on cell A1 achieved a high prediction accuracy (R 2 = 0.99). Training the DNN in the source domain and fine-tuning the trained network using 10% of target data results in, on average, 85% less training time. This results in a DNN model developed for the target domain, which is now as accurate as the model developed for the source domain. The applied technique reduces the computational cost by 85% and the data requirement by 90% for developing predictive models for SOFCs. • Collected datasets from three tubular cells and one planar cell. • 23,820 data samples as open-source for the research community. • Designed a DNN model to predict SOFC output voltage with less than 1% error. • Data requirement is reduced by 90% for SOFC data-driven model via transfer learning. • Reduction of training time by 85% compared to the case without a transferred model.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".