Impact of Operations on a Series Per Reliability Perspective

Abstract

The Mean Time Between Failure (MTBF) figures are the average of test durations between failure observations formed into a series. They can be seen to suit classical statistical distributions. An equally possible condition is that they are not stationary as a contradiction to the previous statement. In such worse conditions accepting availability of a concealed statistical property, this paper tries to identify the impact of Bi-sample Differencing and Bi-sample mean manipulations. In other words, operating on reliability observations series to reveal concealed statistical knowledge. Experimentation based on observation over a stationary series as a controlled experiment. As an outcome of experiments, the differencing seems to be alleviating the trend and seasonality to a degree. The bi-sample averaging is observed to be hiding variant conditions.