Power MOSFET Thermal Aging Data for Remaining Useful Life Prediction

NASA Ames research presents run-to-failure data from a controlled thermal overstress experiment on power MOSFETs. The study identifies increasing ON-state resistance as a key failure precursor linked to die-attach degradation. Prognostic algorithms, including Gaussian process regression and Kalman filters, were validated using this data.

Use Cases

• Train Gaussian process regression models using ON-state resistance time series to predict remaining useful life.
• Apply extended Kalman or particle filters for model-based prognostics with ON-state resistance as the primary failure feature.
• Analyze the correlation between die-attach degradation and ON-state resistance under controlled thermal stress conditions.
• Benchmark prognostic performance metrics using the collected run-to-failure experimental data.

Strengths

• Data originates from a controlled thermal overstress experiment conducted by NASA.
• The primary failure mode (die-attach degradation) and key precursor (ON-state resistance) are identified and validated with simulation.

Limitations

• The dataset is described in a PDF report; the underlying raw data files are not directly accessible.
• The sample size and specific number of tested devices are not provided.

Provenance

Source

National Aeronautics and Space Administration (NASA Ames)

Collection Method

Data collected from a controlled thermal overstress accelerated aging experiment on power MOSFETs.

Time Range

null

Freshness

null

Geography

null