Contributions to planning and analysis of accelerated testing

With the prevailing interest in saving time and expense,Accelerated Life Testing (ALT) and/or Accelerated Degradation Testing (ADT) are becoming an integral part of many programs for improving productsa??quality and reliability.However,appropriate analysis of ALT/ADT data and planning of ALT/ADT experiments are still a challenging task in the face of cost and estimation precision considerations.This dissertation first reviews the basic concepts and existing methods related to ALTs/ADTs,and then proposes a new approach to designing Constant Stress ALT (CSALT) by using some measure to quantify the departure from optimal solution values. Employing stochastic processes to model degradation paths,an efficient method is proposed for Step Stress ADT (SSADT) data analysis.A general formulation for planning both non-destructive Constant Stress ADT (CSADT) and non-destructive SSADT is also conceived by considering the trade-off between total cost of testing and the attainable estimate precision.This planning strategy is further extended to destructive-CSADT. Finally,concluding remarks are composed with highlights.

Author dc:creator

• YANG GUIYU

• Accelerated Testing, Test Planning, Data Analysis, Stochastic Models, Cost Functions, Precision Constraint