By Dr. Wayne Nelson
Many materials and products degrade slowly and fail after a long time under normal use conditions. Usually tests under such conditions cannot run long enough to yield useful reliability information. Accelerated tests quickly yield such information. Run at high stress (mechanical load, temperature, etc.), specimens degrade and fail sooner. Then product life at use stress levels is estimated with a model fitted to the high stress data as described by Nelson (2004) and Meeker, Escobar, and Pascual (2021).This talk describes recent advances in statistical models and data analyses and plots for accelerated tests. Using these advances, scientists and engineers can run faster tests that are more informative and accurate.