A Novel GRU-Augmented Time-Frequency Estimator for IGBT Remaining Useful Life Prediction
Bibliographic record
Abstract
Aging-induced failure of Insulated Gate Bipolar Transistors (IGBTs) significantly restricts the reliability of power electronic systems. Accurate and efficient prediction of IGBT Remaining Useful Life (RUL) is critical for proactive risk mitigation and ensuring system stability. Despite numerous existing aging models and data-driven methodologies, maintaining prediction robustness and accuracy under diverse and complex operational scenarios remains challenging. To overcome these limitations, we introduce a novel GRU-Augmented Time-Frequency Estimator (GATE) tailored for IGBT lifetime prediction. GATE utilizes an autoregressive time-series prediction framework trained via the Teacher Forcing strategy to recursively decode the electrical parameters indicative of the IGBT’s aging state from rich time-frequency features. Experimental validations are performed using the square-wave power cycling dataset from the NASA Prognostics Data Repository. The results demonstrate that GATE significantly enhances prediction accuracy, reducing Mean Squared Error (MSE) to 0.0026 and Mean Absolute Error (MAE) to 0.045, representing improvements of 38.1% and 19.6%, respectively, compared to the leading baseline method. Moreover, recursive forecasting experiments show that GATE precisely predicts the remaining power cycles until the aging threshold (defined as a 15% increase in Vce(on)) at various aging stages (10–60%). Ablation analyses further underline the critical contribution of the frequency-domain component. Collectively, these findings underscore GATE’s capability to reliably decode IGBT RUL directly from historical operational data, bypassing intricate electrical or mechanical modeling, thereby offering a practically deployable and broadly generalizable solution for lifetime management in power electronic devices.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 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 | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".