An approach to quantifying reliability-growth effectiveness

Abstract

This paper presents an approach for quantifying reliability growth effectiveness through the development of a performance parameter and the application of that parameter to an existing reliability growth model, the tracking, growth and prediction (TGP) model. The basis for this performance parameter is presented from two perspectives; based either on previous experience in other reliability growth tests when data are available, or by subjective assessment of a given corporate culture when data are not available. Quantification of this parametric variable with data is presented by normalizing past performance with the reliability growth program goals. Development of the parametric variable without data is accomplished through the characterization of attributes that are indicative of the corporate culture. Five case studies were used to develop the performance parameter based on test data and by subjective assessment of capability. Although the results of these estimates of the performance parameter varied, they were reasonably close. Based on results of uncertainty propagation analyses, it is reasonable to conclude that a modified TGP model provides a conservative estimate of the risk involved in achieving reliability growth goals. This conclusion is based on the model's durability in sustaining the introduction of uncertainty into the existing failure rate parameters as well as into the performance parameter.