Assessing trends in Duane plots using robust fits

The Duane plot is a simple and widely used graphical technique in the analysis of repairable systems. The fitting of a straight line to points in that graph serves to determine the behavior of the system assuming a power law process. However, the classical least squares fitting suffer from lack of robustness and it is specially heavily affected by the more remote and perhaps least interesting points. In order to solve this drawback we propose a robust fit based on Least Median of Squares (LMS) or Least Trimmed Squares (LTS) regression estimators of [1]. ethodology has been applied in a project aimed to study the reliability, availability and maintainability of the aerial contact line of the Spanish railways.