ABSTRACT
Failures of complex repairable systems are in many cases observed to follow a nonconstant intensity function. Recent work frequently uses the model Nonhomogeneous Poisson Process with Power Type Intensity Function (NPP-PT), known as Weibull process. This paper approaches parameter estimation using Best Linear Unbiased Estimators (BLUE) based on the observed failure times and on the properties of the logarithmic gamma distribution. The given method profits from the property of the logarithmic gamma distribution that follows it to be approximated by a normal distribution. On this basis approximative simultaneous confidence limits and acceptance tests are available which are very useful in field data processing.
