A Class of Mappings for Assessment of Aircraft Pneumatic Actuator Degradation Parameters

The aim of this paper is to propose a method to create features that can be used as prognostic indicators in the prognostics and health management context. The features are described by combinations of n-dimensional mappings, which are related to the information embedded into the different axis. Each axis are derived from a reduced set of signals strongly related with the performed task, which, in turn, are associated with specific degradation types. In order to illustrate the method, a case study of an aircraft pneumatic actuator is presented, based on tasks that are closed loop control operations. In this case, because the tasks are segments of normal operation, no special experimentation needs to be implemented. Moreover, it was observed that using only the already available sensors a set of features could be created to characterize the valve's degradation. The adopted approach is hybrid in the sense that one part is derived from simulations awhile another part is based on the actual aircraft flight data.