Reliability Analysis of Data with No Failure from Fleet and Proving Ground Endurance Tests 920773
The reliability of an electronic sensor in the automotive applications is assessed using data from Fleet Test and proving ground Vehicle Endurance test. These nonfailure data are multiply censored at different mileage. Reliability analysis of data with no failure is rarely discussed in most reliability literature. This paper applies the Weibull maximum likelihood analysis based on known values of the Weibull shape parameter to extract useful reliability information. The well-known Weibayes and Weibest methods are subsets of the discussed approach. The sensitivity of the change of reliability levels over a range of Weibull shape parameter values is also examined in our case. The Huang-Porter (1991) approach of obtaining a reliability lower bound regardless of the Weibull shape parameter values is also applied and its potential of practical application is discussed. Practical limitations of all methods are discussed.