Autonomous Monitoring and Assessment of Sensor Data in Support of Calibration and CBM

Impact Technologies, LLC, in collaboration with Engineering Reliability Software Group, LLC (ESRG), propose development of a widely applicable anomaly detection and sensor/data health assessment software product in support of condition based maintenance. The technology is innovative in that it uses a data-driven modeling methodology which has generic application to both high and low bandwidth signals and is capable of detecting subtle changes in signals resulting from calibration or sensor drift issues. This approach is highly scalable (not limited to the number of sensors) so it can be applied across any ship system and readily supports the condition-based maintenance (CBM) and mission readiness philosophies of the US Navy. Specific signal processing and data-driven modeling techniques will be developed for this purpose. The software modules will utilize open-systems protocols to support integration into Navy systems such as ICAS, MELS and the I/E-PARS reporting system. The signal health software module will focus on early Fault Detection and Isolation (FDI) of faulty and corrupted data, with reporting functions. The module will be capable of real-time sensor fault detection (wired or wireless) and recovery from sensor faults using the signal’s predicted value until the sensor is repaired or replaced. This will allow the ship systems being monitored to continue to operate as normal without the need for abrupt maintenance actions that are costly to mission readiness and effective operations.