VAMPIRE 3: A Decentralized Platform for Acoustic Diagnostics
Abstract
With previous support from the Office of Naval Research, we have developed a new approach for vibration assessment that can be used to characterize the health of mechanical mounts and the extent of imbalance for shipboard rotating electric machinery. This new approach uses a custom self-powered instrumentation platform that can be used to make measurements of vibration and other machine parameters called VAMPIRE (vibration assessment monitoring point with integrated recovery of energy). This instrumentation platform has been demonstrated as a tool for new approaches to in-situ, vibration monitoring that could support real-time acoustic signature assessment.To generate real time assessments VAMPIRE must process high bandwidth acoustic data locally without relying on remote servers for computation. The need for decentralized realtime processing is not unique to VAMPIRE. Any shipboard sensor that provides real time feedback faces a similar problem. Currently most sensor platforms, including VAMPIRE, use vertically integrated software that performs both the data capture and analysis in a single executable program. The leads to a large, tightly coupled code base that is time consuming and therefore expensive to develop. Vertically integrated software is also complex to maintain because changes often produce difficult to detect side effects elsewhere in the code base. A new modular software framework is underdevelopment at the US Naval Academy for decentralized signal processing. We propose to develop and provide a high-speed time-series database, NILMdb, to interface with this signal processing framework and provide the necessary core data storage, archiving, and access functions to make near real-time monitoring possible. In addition to developing this software framework we also propose new research into non-intrusive condition based maintenance. Our field tests during the first VAMPIRE programs were focused primarily on motors onboard ship, e.g., fire pumps. Experiments conducted on board a USN MCM and USN YP vessel during our previous effort has revealed a new technical opportunity for improving operational readiness and supporting a tactical decision aid for USN warships. Generators on board ships reveal complex vibration patterns duringnormal operation, not only because of vibration in the electric machine (generator) itself, but also because of the prime mover (e.g., a diesel engine) when it shares the same mechanical mount with the electric machine.We propose to extend and demonstrate the role of real-time vibration monitoring for "go/no-go" acoustic signature assessment and for condition-based maintenance, with an eye on developing focused observations for multi-machine systems like diesel engine/generator pairs.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 26, 2018
- Source ID
- N000141812496
Entities
People
- Steven Bruce Leeb
Organizations
- Massachusetts Institute of Technology
- Office of Naval Research
- United States Navy