(3) Condition-Based Predictive Maintenance for Mission Critical Systems with Probabilistic Knowledge Graph and Deep Learning

Introducing the I-SEER, Intelligent Fusion Technology’s - SErvice Enhanced Recommender, a data analysis web application for condition-based predictive maintenance (CBPM). CBPM provides insight into naval assets, such as fixed wing, rotary, auxiliary power units, and ground vehicles, by predicting the most likely point and time of failure. I-SEER uses CBPM to provide a window into the future of system performance, drive efficiencies, increase combat readiness, and reduce mishaps. I-SEER is based on a state-of-the-art knowledge graph (KG) and deep learning (DL) framework for automated diagnostics and maintenance scheduling, based on data and equipment conditions. The core technology can be applied to different Navy’s maintenance needs. The I-SEER application is designed for complex mechanical systems, and uses historical and streaming physical sensor data and unstructured text from technical documents and past records to predict mechanical failures, minimize the time for determining the root cause of system failure, reduce time waiting for parts, make the most inferences from the available sensors, and recommend maintenance activities based on a user-defined acceptable risk. The proposed cognitive-based decision support system is not meant to replace human interaction and decision-making, rather, support the operator with fused data, rapidly identify potential failures, and provide timely recommended actions. I-SEER is designed for incorporation and integration into the algorithmic base of the standardized Navy condition-based maintenance systems, using pre-established guidelines, processes, and procedures. Due to the flexible web-based or application service architecture, I-SEER can be deployed anywhere the asset maintenance is performed such as the Naval Depot-Level, Field-Level, or vendor customer providers.