Methods for Analyzing Reliability Engineering: Probabilistics and Accelerated Testing in Early Phases of Product Emergence Process
摘要
The product emergence process of a product is significantly influenced by its reliability, which in turn determines both customer satisfaction and operating costs. It is therefore evident that reliability engineering plays a pivotal role in the development and optimization of products. This paper presents the application of reliability engineering methods using three representative case studies based on real test data, which have been enhanced by simulation. The data presented here originates from fatigue tests of shape memory actuators. In order to facilitate comprehension, an initial presentation will be made of the fundamentals of the shape memory effect, together with an account of the nature of shape memory actuators and the procedures employed in their testing. The initial case study entails an examination of the failure behavior exhibited by three distinct test channels of a test rig. The application of descriptive statistics (boxplots) and significance tests (Kruskal-Wallis test with Conover post-hoc analysis) enables the identification and interpretation of differences in the failure behavior of the channels. The second case study focuses on accelerated testing. The determination of suitable acceleration levels is achieved through the analysis of different load levels. The investigation is conducted using the two-parameter Weibull distribution, including the calculation of confidence intervals for the model parameters and the determination of acceleration factors. In the third case study, a Sudden Death test is conducted in which candidate data is extrapolated using the Johnson rank correction in order to make reliable statements about the service life distribution. The methods presented illustrate the practical advantages of reliability engineering approaches in the analysis and optimization of system reliability, and offer valuable insights for product development and testing.