Data processing: measuring – calibrating – or testing – Measurement system – Performance or efficiency evaluation
Reexamination Certificate
2002-09-06
2004-02-03
Hoff, Marc S. (Department: 2857)
Data processing: measuring, calibrating, or testing
Measurement system
Performance or efficiency evaluation
C714S025000
Reexamination Certificate
active
06687654
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to monitoring machinery; and, in particular, to automated distributed detection of faults that constitute conditions for performing corrective maintenance before failure of the machinery.
2. Description of the Related Art
High-valued complex machinery often constitutes a major investment to its owner and is not easily replaced. Examples include consumer items such as automobiles and farm equipment, heavy equipment such as trains, cranes, drills and earthmovers, as well as special purpose factory installations such as power generators, assembly line equipment, and power train equipment, such as transmissions, for delivering power to assembly line equipment. Owners of such machinery desire to detect and correct small problems with individual components of such machinery before the small problem leads to catastrophic failure of the machine. However, it is often impractical to inspect each small component subject to failure on a frequent basis. The component may be buried deep in the machinery and require many person-hours to remove, inspect and re-install or replace. In addition to the costs of the person-hours, there is the cost of having the high-valued equipment non-operational for the duration of the inspection procedure. Such costs are only warranted when the part is sufficiently defective that failure to replace may lead to failure of the high-valued complex machine of which it is part.
There is a clear need for systems that can monitor the high-valued complex machinery for failure of individual components while the machinery is operating for its intended purpose.
One approach is to build-in special purpose sensors that detect the correct operation of each individual component, and have those sensors report when the associated component fails. This approach is impractical for many reasons and is not taken in practice. In many machines, there are so many moving components, some very small, that special purpose sensors attached to each one may interfere with required motions, violate required spatial tolerances, increase the cost of the machinery, and otherwise render the machinery unsuitable for its purpose. Another problem with this approach is that some failure modes are not determined until after the machine is built and operated, and it is impossible to guarantee a sensor that will detect such failure modes before they are discovered.
Another approach is to attach vibration sensors to the machinery and analyze vibration data from such sensors. Changes in operation of one or more components of the machinery associated with failure of that component may change one or more characteristics of the vibration data. This approach has been taken by many conventional systems. However, the changes that can be detected depend on the characteristics of the vibration data and the processing of the vibration data.
Some conventional systems process vibration data by measuring the shape and size of vibration amplitude with time. Such systems have been used to determine gross transient properties of machinery, such as firing of a circuit breaker, or approach of a train on train rails. However, such systems have not been shown to detect small changes in minor components of the machinery. Such small changes are often dwarfed by the vibrations caused by larger, more energetic components, such as drive shafts.
Some systems process vibration data by determining statistics of the vibration in the frequency domain. However, such systems have not been shown to detect small changes in minor components of complex machinery. Such small changes are often dwarfed by the vibrations caused by larger, more energetic components, in the same frequency band. Some small components may vary their frequency signature with time during normal operations, so that it is difficult, using fixed frequency bands, to distinguish normal variations from variations associated with an approaching failure of a minor component.
Based on the foregoing, there is a clear need for a machinery monitoring system that can detect problems in minor components of complex machines, which warrant maintenance actions directed to those minor components.
The past approaches described in this section could be pursued, but are not necessarily approaches that have been previously conceived or pursued. Therefore, unless otherwise indicated herein, the approaches described in this section are not to be considered prior art to the claims in this application merely due to the presence of these approaches in this background section.
SUMMARY OF THE INVENTION
Techniques are provided for monitoring a machine for significant deviations from normal operations, which indicate conditional maintenance should be performed. In one aspect of the invention, techniques include collecting, at a field processing element, sensor data about the machine. The field processing element performs narrowband frequency domain processing to determine a segment of sensor data that indicates a deviation from normal operations that exceeds a threshold deviation. A message including the segment of sensor data is sent to a base processing element. In response to receiving the message, the base processing element performs different narrowband frequency domain processing to determine whether the deviation from normal operations is significant for maintaining the machine. If the deviation from normal operations is determined to be significant for maintaining the machine, then the deviation is reported to cause the machine to be maintained.
According to another aspect of the invention, techniques include collecting, at a field processing element, sensor data about the machine. The field processing element performs first narrowband frequency domain processing to determine a segment of sensor data that indicates a deviation from normal operations that exceeds a threshold deviation. The field processing element sends a message including the segment of sensor data to a base processing element for performing different narrowband frequency domain processing to determine whether the deviation from normal operations is significant for maintaining the machine. Narrowband frequency domain processing includes normalizing a frequency spectrum for a temporal portion of sensor data. Normalizing the frequency spectrum includes determining a broad background spectrum and removing the broad background spectrum from the frequency spectrum to produce a residual spectrum. Narrowband frequency domain processing also includes determining in the residual spectrum a peak set of one or more spectral peaks that exceed a threshold amplitude.
According to another aspect of the invention, techniques include receiving at a base processing element a message including a segment of sensor data about the machine from a field processing element. The field processing element determined that the segment indicated a deviation from normal operations that exceeds a threshold deviation. In response to receiving the message, the base processing element performs first narrowband frequency domain processing on the segment of sensor data to determine whether the deviation from normal operations is significant for maintaining the machine. The first narrowband frequency domain processing is different from second processing performed in the field processing element. If the deviation from normal operations is determined to be significant for maintaining the machine, then the deviation is reported to cause the machine to be maintained. Narrowband frequency domain processing includes normalizing and determining in a residual spectrum a peak set of one or more spectral peaks that exceed a threshold amplitude.
These techniques allow the detection of subtle changes in sensor data that indicate problems in minor components of a complex machine. The techniques also allow false alarm rates to be reduced to desirable levels.
This system scales well with an increasing number of machines and machine components to be monitored because the techniques can be u
Peacock George S.
Smith, Jr. George A.
Vojtech, Jr. George L.
Cooch Francis A.
Hoff Marc S.
Raymond Edward
The Johns Hopkins University
LandOfFree
Techniques for distributed machinery monitoring does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Techniques for distributed machinery monitoring, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Techniques for distributed machinery monitoring will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3316318