Interface design for prognostic asset maintenance

1. Scope We report here part of a project funded by INNOVATE-TSB (UK) concerning the implementation of a prognostic (that is, predictive) maintenance system for escalator assets on the underground railway. The present approach to asset maintenance is primarily based on scheduled maintenance and repair at regular intervals which is wasteful in terms of both human and physical resources. Although humble in machine terms, escalators are essential for moving passengers around underground stations at an appropriate rate of flow and their good maintenance is essential for public safety and the wider life of mass transit systems as a whole (Campbell, 2002). Maintaining them is a demanding task owing to their typical placement, accessibility and limited scheduling windows for work and their sheer ubiquity escalators in various settings around the world means that interventions concerning them have considerable economic implications. The project aimed instead to demonstrate the use of condition monitoring data processed via prognostic computational intelligence to move to a system where maintenance occurs on the basis of identified Condition Indicators and a time-risk based assessment of Remaining Useful Life. Arguably the success of this project hinged upon solving Human Factors challenges in designing interfaces to convey the output of computational intelligence in such a way that it can be used to support a new engineering decision making paradigm. In this context, user interfaces serve as the critical ‘last mile’ in terms of delivering on the potential of investments into sensor technology and computational intelligence in decision making. We anticipate that as digital sensing and the “Internet of Things” become ever more widespread, there will be significant future demand for similar interventions combing data representation with support for changing decision making processes across a range of industries. 2. Project organisation