The smaller assets of US Navy require health and utilization monitoring systems (HUMS) to prolong their effectiveness. For example, vehicles contain pumps, motors, gearboxes, alternators, etc. that experience degradation. Similarly, Naval ship experiences vibration due to hull impacts from wave impact and operation of machinery like marine propulsion shafts, gearboxes, propellers, pumps and main engines. This vibration can lead to structural fatigue as well as wear. The cumulative degradation impacts the cost and mission. The first goal is to present a report describing the hardware and micro-electronics designed to achieve the system goals of: 1) low-cost, 2) small size, low weight, and continuous power, 3) ruggedness in harsh settings, and 4) wired and wireless features. The second goal is to present a report that can detect real-time anomaly outputs and threshold particular anomalous behavior. The display will make use of data fusion to differentiate among nuisance anomalies and impending component failures from the multiple component-specific devices. The spatially aware devices will catalogue all device information within the software. The third goal is to develop a system architecture in the form of machine learning of the structure and weights of a neural network that can be applied to analyze vibration metadata.
Benefit: The anticipated benefits include enhancing the real-time situational awareness for the navy personnel when monitoring vibration anomalies and novelty anomaly detection for vehicle components or structures. The limited time and available resources will be used more effectively by the data, metadata, and anomaly sharing inherent in the Zansors approach. The Zansors solution can be utilized to establish a standardized approach for low-cost and small size HUMS for lower valued assets. The proposed work has broader benefits. For example, the proposed solution can be expanded into commercial applications like commercial trucking, heavy construction equipment, manufacturing, aircraft and related equipment, commercial maritime, and infrastructure monitoring (e.g., bridges, locks, damns).
Keywords: anomaly detection, anomaly detection, Machine Learning, Condition Based Monitoring, mechanical, Sensors, Neural networks, HUMS, Algorithms
