My client had just begun the first phase of a Navy program to
develop a Condition-Based Maintenance (CBM) system to predict
failure probabilities for large motors, gearboxes, and other
equipment on Navy ships. The replacement and/or repair of this
equipment when a ship is docked is expensive, thus delaying this
maintenance saves money. A failure at sea, however, is much more
expensive and can be, in some cases, catastrophic. The Navy wanted a
system that predicted when a piece of equipment would fail. In this
way, the maintenance people could assess the risk of deploying a
ship without replacing equipment or other key parts, knowing that it
may not return to its home base for up to two years.
My client, the primary contractor on the program, hired a firm
with many years of experience in testing vibration and other
environmental parameters on Navy ships. I was contracted to develop
the functional requirements for the system based on the original
technical proposal to the Navy, and to interface with the
subcontractor to ensure their technical plans met the proposed
requirements.
The proposed system would continuously monitor critical equipment
and maintain a record of operational signatures such as temperature
profile, vibration levels, and frequency content of vibration. When
a change takes place in a moving component, the parametric signature
changes, and the system would recognize it. Based on known failure
analysis and artificial intelligence technology, the system would
produce a prognosis, an estimate of the time to failure for a
specific component.
The system concept had several major elements:
Clusters of sensors, attached to a piece of equipment, capable
of communications over a network
A ship-wide computer network for communicating the sensor's
data
A central computer that stored the data, analyzed the current
data against a database, and produced diagnostic and prognostic
estimates
I analyzed the original proposal to the Navy, assessed the ideas
that the team was producing, and wrote a document that derived a
logical system architecture based on a methodical analysis of the
proposed requirements. I also worked with the team to develop a set
of requirements to deliver to the Navy at the end of the first phase
of the project.
