NASA Power MOSFET Prognostics Data with Thermal Cycling Experiments

NASA Ames research demonstrates prognostics for 100V power MOSFETs using thermal cycling to age devices. The methodology employs Gaussian process regression and validates the approach with experimental data. The dataset includes experimental results and finite element analysis simulations based on a two-transistor model.

Use Cases

• Train Gaussian process regression models on thermal cycling data to predict MOSFET failure.
• Analyze the relationship between ON-state resistance change and junction temperature as a failure precursor.
• Use finite element analysis simulation data to interpret degradation phenomena and safe operation area (SOA) changes.
• Validate prognostics methodologies for die-attachment degradation in 100V power MOSFETs.

Strengths

• Research conducted by the National Aeronautics and Space Administration (NASA).
• Methodology validated with experiments on specific 100V power MOSFETs.
• Data includes both experimental results and finite element analysis simulations.

Limitations

• Data is presented within a research paper (PDF), not in a structured, machine-readable tabular format.
• The specific volume of data points (rows) and features (columns) is not quantified in the description.
• The dataset's primary focus is on a specific failure mechanism (die-attachment degradation) under defined stress conditions.

Provenance

Source

National Aeronautics and Space Administration (NASA Ames via Data.gov)

Collection Method

Experimental thermal cycling of 100V power MOSFETs, augmented with finite element analysis simulation.

Freshness

The record was last updated on September 19, 2025.