Dynamic Physical/Data-Driven Models for System-Level Prognostics and Health Management

The proposed effort leverages the capabilities of data-driven and physics of failure (PoF) based prognostic techniques for electronic systems by combining them within a hybrid approach. Data-driven and PoF-based techniques both have shortcomings; combining them into a hybrid framework allows using their capabilities in a complementary fashion, and thereby providing a reliable way of prognostics and health management in electronic systems (e-PHM). The approach will not only estimate the remaining useful life (RUL) and forecast the performance and health condition of electronic systems, but will also identify the potential source(s) of degradation or failure that might impact their future health (and consequently, performance and RUL). The approach will result an e-PHM solution that complies with the standards and software development environment of DoD’s Automatic Test Systems (ATS) and Navy’s Consolidated Automated Support System (CASS). Therefore, it can be easily incorporated into a Test Program Set (TPS). Such incorporation capacity provides very wide scope of application of the solution across electronic systems used by Navy and other DoD agencies. For obtaining CBM+ decision support, faster maturation, and easier incorporation into a TPS, the envisioned prognostics solution will be implemented on QSI’s TEAMS diagnostic design and analytic platform.