Health Assessment, Prognostics and Remaining Useful Life – Part A

The implementation of a data-based prognostics approach supposes that the system can be equipped with sensors in order to gather raw data in real time, which makes it possible to observe the evolution of the behavior. Regarding the specifications of a prognostics tool, a central issue can be raised: the utility of such a system is deeply connected to its ability to approximate and predict the degradation of equipment. Tools of the kind of neural networks or neurofuzzy networks have been used a lot for prognostics applications, due to their ability to approximate nonlinear signals. This chapter addresses the problem of the health state assessment of a monitored system based on extracted and/or predicted features. The deployment of the data-based prognostics is usually performed in two phases: a learning phase and a test one. Monitoring and prognostics of an asset are intimately connected processes.