|
Integrated System Health Management (ISHM) is a global application developed by NASA that uses data mining techniques to produce real time health monitoring capability for complex systems
This system automatically learns typical system behavior by discovering parameter relationships in historical or real time data and constructs general classes or clusters representing normal system operation.
Operational data is then compared to these classes to detect off-nominal system behavior and to alert operators or other automated systems to possible subsystem failures or impending failures.
IMS has been used on all major NASA platforms as well as many manned projects in both fixed wing and rotorcraft environments. Further IMS operates on-line or off-line. .. |
|
 |
|
This is an inference engine designed for monitoring, analysis and diagnosis of real-time and non-real-time system performance.
It is a reusable software tool using advanced state-of-the art technology and enabling solutions to a broad class of problems, previously considered impossible because of real-time system requirements, small size constraints, portability and hardware incompatibility
Shine uses a highly advanced prognostic estimator for hard real-time diagnostics and interpretation of the system state.
This technology provides insight into the system operation using channel-based diagnostic techniques making near zero false alarms attainable.
|
|
|
| |
This system - Advanced and Optimized Sensor System – is designed by NASA for the optimization of any diagnostic or state awareness sensor network. The quality and efficiency of a diagnostic system depends on the availability and relevance of the information it can retrieve. The different issues regarding sensor placement can be summarized as follows: |
Determining the degree of a system information capacity, i.e., characterizing the set of the faults that can be discriminated.
Finding a minimal additional sensor set that guarantees a specific degree of diagnostic capacity
Finding the minimal cost additional sensors which achieve a specific degree of diagnostic capacity |
|
|
| |
BEAM is a complete data analysis system for real-time or off-line fault detection and characterization, which takes data as input and reports fault status as output.
Implementation of this software depends on the application, but a typical application would have a system with a number of individual components, each of which reports health or performance data to a local computer.
To accommodate such a wide range of possibilities,
the engine of BEAM itself is highly adaptable with respect to subsystem size and complexity.
BEAM constructs a strategy to characterize a system from all available observations, and train that characterization with respect to normal phases of operation.
BEAM functions much as a human operator. Through experience and other available resources an allowed set of behavior is “learned” and deviations from this are noted and examined. |
|
