Technology has been developed that provides a way to compute the remaining useful life (RUL) of a component or system. The estimation of the RUL of a degraded or faulty component is at the center of condition-based maintenance, and prognostics and health management. It gives operators a potent tool in decision-making by quantifying how much time is left until functionality is lost. This is especially important for aerospace systems, where unanticipated subsystem or component failure may lead to failure of the system as a whole, which in turn may adversely affect the safety of operation.

The technique described here has a training (off-line) mode and a subsequent run-time (on-line) mode for estimating an RUL of an object that is in active use for at least part of the time (an “active object”). In the training mode, the system collects training data, including operating conditions of the object, measurements from sensors monitoring the system, and the ground truth indicating the true extent of damage. The system extracts or identifies precursor features of failure from the sensor data by analyzing their correlation to the ground truth. The feature domain size is optionally reduced by eliminating one or more features that are highly correlated to other features, such that the exclusion does not diminish information about damage progression in the system.

The prognostic technique decomposes the problem of estimating the RUL of a component or subsystem into two separate regression problems: the feature-to-damage mapping and the operational conditions-to-damage-rate mapping. These maps are initially generated in offline mode. One or more regression algorithms are used to generate each of these maps from measurements (and features derived from these), operational conditions, and ground truth information. The regression can be carried out using either physics-based models or — where the cost/benefit analysis of designing physics-based damage propagation algorithms is not favorable, and when sufficient test data are available that map out the damage space — one can employ data-driven approaches or a combination of data-driven and model-based (hybrid) techniques. The decomposition technique allows for the explicit quantification and management of different sources of uncertainty present in the process. Next, the maps are used in an on-line mode where run-time data (sensor measurements and operational conditions) are used in conjunction with the maps generated in off-line mode to estimate both current damage state as well as future damage accumulation. Remaining life is computed by subtracting the instance when the extrapolated damage reaches the failure threshold from the instance when the prediction is made.

This work was done by Kai F. Goebel of NASA, Bhaskar Saha of MCT, Inc., Abhinav Saxena, and Jose R. Celaya of SGT Inc. for Ames Research Center. NASA invites companies to inquire about partnering opportunities. Contact the Ames Technology Partnerships Office at 1-855-627-2249 or This email address is being protected from spambots. You need JavaScript enabled to view it.. Refer to ARC-16273-1