|
|
| Prognostics Framework Prognostics, the
identification of incipient faults, is usually considered and treated as a
component/part problem rather than a system problem.
When cast as a component/part problem, prognostic techniques are limited
to those using failure statistics (e.g. wear and tear distributions) and
measurements that are specifically correlated with a failure process
(e.g. measurements wear and tear).
However, due to the reduced DOD budgets, the need for the
capability to monitor the health of a weapon system is now more critical for the
system readiness. Moreover, systems
such as spacecraft, satellite systems, aircraft, submersible vehicles, and
armored combat vehicles often have to operate in hostile environments with
little or no human interaction. These
require an autonomous real-time health monitoring system which provides
automatic fault detection, isolation, reconfiguration, and reporting to increase
the probability of overall mission success. This, therefore, requires dealing with prognostics as a
system-level problem. The
Prognostics Framework interrogates and integrates many measurements to
determine the overall health condition of the system and hence to determine
mission readiness of the system and required maintenance tasks.
Various component/parts prognostic techniques have been developed in various laboratories, universities and industry. Techniques such as neural networks, time/stress measurement devices, vibration monitoring, oil monitoring, sensors, trend analysis, statistical analysis, etc. are, by their very nature, equipment-specific and very few applications have been implemented. Moreover, they are primarily point solutions. There is no overall framework to manage the information provided by the individual techniques. The Prognostics Framework was developed to integrate the various sources of diagnostic/prognostic data into useful information. The Prognostics Framework is a generic, tailorable software tool that uses model-based reasoning to integrate embedded test and sensor data into diagnostic and prognostic information. Use of the Prognostic Framework will save time, money, and program-specific funds and allow systems to converge upon a prognostic capability over time. Furthermore, the Prognostics Framework will provide information to operational and maintenance crew with normalized/standardized data across a fleet of systems. This Prognostics Framework will also integrate prognostic mechanisms at various stages of maturity. The Prognostics Framework will be integrated with diagnostics to provide a total "Health Management” Capability, and will also tie-in to logistics infrastructure (e.g., IETM, logistics planning, mission planning, spare parts provisioning). The Prognostics Framework can be used to develop a health information manager for any system.
The predictive techniques to be incorporated into the Diagnostician include: 1) advanced, item-specific prognostic mechanisms (such as artificial neural networks, etc.); 2) linear degradation of signals / measurements over time; 3) historical conclusions /statistics; and 4) engineering correlations.
The underlying design
foundations for the Prognostics Framework are
generic, hierarchical, and open architecture design strategies, with a
single knowledge base for both embedded and off-line applications. It is an object-oriented software structure. The
Prognostic Framework consists of a suite of development tools that support developers in the application of the framework,
and run-time tools that support embedding the framework into an
operational system. A synopsis of the tools is provided below. Prognostics Framework development tool suite Prognostics Framework Run-Time Software: Health Management System Preprocessing Techniques for Prognostic Framework Inputs Download Prognostics Framework Health Monitoring System Demonstration Return
to Products Page
|
|
Send mail to CompanyWebmaster with
questions or comments about this web site.
|
