Confidence intervals for the Crow (AMSAA) reliability growth model based on grouped data

In many reliability growth applications the actual failures times are not known but the number of failures over intervals of time are known. This is called grouped data. For grouped data the maximum likelihood estimates of the parameters of the Crow (AMSAA) reliability growth model yield a point estimate of the current system mean time between failures (MTBF). In practical applications an interval estimate or confidence interval is also often of interest in order to make appropriate management decisions regarding the reliability of the system. To address this need, easy to use confidence interval procedures for the grouped data case of the Crow (AMSAA) model were developed and will be discussed in this paper. These confidence interval procedures have been accepted for use by the International Electrotechnical Commission (IEC) and have been included in the IEC International Standard on reliability growth. This paper gives an overview of the procedures for calculating the confidence intervals and illustrates these methods through numerical example.