A novel approach to the integration of posterior knowledge into condition monitoring systems: theory and practice

Considers the problem of the integration of posterior knowledge into condition monitoring systems from both the theoretical and practical points of view. The problem is framed in the context of elementary probability theory where the task of posterior knowledge representation is examined. A methodology for updating posterior probabilities is proposed for cases where fault conditions are rejected or retained on the basis of external knowledge supplied by an end-user. A possible condition-class ranking is generated following the estimation of condition-class probability functions. It is shown that a simple renormalisation of existing probabilities does not apply in the dependent condition-class case and can lead to erroneous results; the condition-class ranking may change following the exclusion of condition-classes known not to have occurred. An artificial example is used to illustrate the theoretical principles. Simulations are then used to show the effect of posterior knowledge as part of a maintenance methodology. Preliminary results indicate that posterior knowledge reduces the average search depth for faults. This may represent a maintenance saving in terms of reduced number of sub-unit inspections