Component machine testing using neural network processed vibration data analysis
First Claim
1. Apparatus for evaluation and diagnostics of a component machine that is part of a larger machine where the diagnostics and evaluation are through neural network analysis of a vibration data signal to identify abnormal signal patterns in said vibration data, comprising in combination:
- first sensing means positioned for sensing vibration data signals produced by said component machine;
second sensing means positioned separate from said first sensing means for sensing background noise vibration data in the vicinity of said component machine;
first neural network preprocessing means for rendering as normal any influence of said sensed background noise in said sensed component machine vibration data signals; and
a processing stage for evaluation and diagnostics, said processing stage including;
an interactive interface member having at least the features of simultaneous multiple signal display and operator input and means for receiving said preprocessed component machine vibration data signal, and,a second neural network with associated diagnostic technology operating systems and supporting data bases for evaluation and processing determinations based on signal patterns in said component machine vibration data signal identified in said second neural network.
1 Assignment
0 Petitions
Accused Products
Abstract
A component machine testing technique is provided that performs diagnostic analysis on a vibration signal of the component machine that has been separated from power and load machine background noise in a first neural network. The diagnostic analysis, with operator direction through an interactive interface, uses a second neural network in performing a series of diagnostic operations followed by archival of any experience acquired in the testing operation being performed.
In the diagnostic analysis, both time based and frequency based vibration signal information from the component machine under test are used together through a simultaneous multiple display interactive interface under operator direction.
73 Citations
15 Claims
-
1. Apparatus for evaluation and diagnostics of a component machine that is part of a larger machine where the diagnostics and evaluation are through neural network analysis of a vibration data signal to identify abnormal signal patterns in said vibration data, comprising in combination:
-
first sensing means positioned for sensing vibration data signals produced by said component machine; second sensing means positioned separate from said first sensing means for sensing background noise vibration data in the vicinity of said component machine; first neural network preprocessing means for rendering as normal any influence of said sensed background noise in said sensed component machine vibration data signals; and a processing stage for evaluation and diagnostics, said processing stage including; an interactive interface member having at least the features of simultaneous multiple signal display and operator input and means for receiving said preprocessed component machine vibration data signal, and, a second neural network with associated diagnostic technology operating systems and supporting data bases for evaluation and processing determinations based on signal patterns in said component machine vibration data signal identified in said second neural network. - View Dependent Claims (2, 3, 4, 5)
-
-
6. The method of evaluation and diagnostics of a component machine that is part of a larger machine where the diagnostics and evaluation are through neural network analysis of vibration data signals to identify abnormal signal patterns, comprising in combination the steps of:
-
sensing vibration data signals at the location of a component machine; sensing background noise vibration data signals at a location in the vicinity of said component machine; providing a first neural network preprocessor for rendering as being normal any influence of said sensed background noise in said sensed component machine vibration data signals by autoregression parameter selection; and
,processing said sensed component machine vibration data signals for evaluation and diagnostics, said processing including viewing a simultaneous multiple signal display and inputting responses based on information in said display. - View Dependent Claims (7, 8, 9)
-
-
10. A vibration data processing element for extracting background vibration signals from specific vibration signals that originate at a specific source and are subsequently processed in a neural network, comprising in combination:
-
means for sensing in a first sensing unit said specific vibration signals at a first location said first location being adjacent to said specific source of said specific vibration signals; means for sensing in a second sensing unit said background vibration signals, said second sensing unit being positioned separate and remote from said source of specific vibration signals; means involving autoregression for producing a minimum and maximum expectation signal representing said background vibration signals in said sensed background vibration signals and deriving therefrom normal expectation range signals; a neural network member; said neural network member having an input layer of processing elements each with an input terminal, each said input terminal being supplied with an expectation range signal from said expectation range signals; said neural network member having an output layer of processing elements each with an output terminal; each said processing element in said input layer being connected to each said processing element in said output layer; and
,each said processing element in said output layer being connected to each said processing element in said input layer; and
,a vigilance stage positioned between each said output terminal and a corresponding input terminal of said subsequent neural network processing, said stage being responsive to a vigilance factor and operable to pass output signals based on a required portion of all possible output signals at said output terminals.
-
-
11. An apparatus for diagnostics and evaluation of a machine through neural network analysis of a vibration data signal of said machine to identify abnormal signal patterns, the improvement comprising;
a vibration data signal preprocessing stage including a neural network with selected autoregression and vigilance factors operable to identify specific vibration data signal patterns in said vibration data signal as being normal. - View Dependent Claims (12)
-
13. In a system in which vibration data from a machine is assembled and then processed through a neural network processor in diagnostic and evaluation determinations of said machine;
an interface for operator interaction with said system positioned in said system between the input of said vibration data and said processor, comprising in combination; input means adapted for manual entry of operator direction into said system; and
,display means including a plurality of at least two separate data signal display capabilities and a separate condition indicator display capability, said display means being positioned within operator view when said operator is within reach of said input means. - View Dependent Claims (14, 15)
Specification