JPWO2020095303A5

JPWO2020095303A5 -

Info

Publication number: JPWO2020095303A5
Application number: JP2021525314A
Authority: JP
Publication date: 2022-08-29
Anticipated expiration: 2039-11-07

Claims

A method for monitoring at least one machine, the method comprising:
causing at least a first sensor to obtain at least a first non-stationary signal from at least one machine operating in a non-stationary manner during at least one operating time frame; the first sensor provides at least a first non-stationary output;
causing at least a second sensor to obtain at least a second non-stationary signal from the at least one machine during the operating time frame, wherein the at least second sensor receives at least a second providing a non-stationary output;
fusing said at least first non-stationary output with said at least second non-stationary output to produce a fused output;
extracting at least one feature of at least one of said first and second non-stationary signals based on said fused output;
analyzing said at least one characteristic to ascertain the health of said at least one machine; and
performing at least one of a repair action, a maintenance action and a modification of an operating parameter of said at least one machine based on said health found by said analyzing;
the aforementioned method.

2. A method for monitoring at least one machine according to claim 1, wherein said at least one characteristic is insensitive to the level of constancy of said non-stationary operation of said machine.

3. The at least one machine of claim 1 or claim 2, wherein said fusing comprises modifying said at least first non-stationarity output based on said at least second non-stationarity output. Methods for monitoring.

The at least first non-stationary signal represents a mechanical state of the machine and the at least a second non - stationary signal represents an operational state of the machine. A method for monitoring at least one machine according to clause .

The fusing comprises applying a wavelet transform to the first and second nonstationary outputs, and the modifying comprises one of the first and second nonstationary outputs. 5. The monitoring of at least one machine according to claim 3 or claim 4, comprising multiplying the wavelet transform of the first and second nonstationary outputs by a binary mask of the wavelet transform of the other of the first and second nonstationary outputs. How to.

3. The method for monitoring at least one machine of claim 2, wherein said fusing employs deep learning.

said deep learning combining said at least first and second non-stationary outputs into a vector having one or more dimensions; and training a neural network to automatically classify said vector. 7. A method for monitoring at least one machine according to claim 6, comprising:

training said neural network to convert said non-stationary output to a corresponding stationary output from at least one machine sharing at least one common characteristic with said at least one machine being monitored; 8. A method for monitoring at least one machine according to claim 7, comprising training the neural network to classify based on.

said fusing comprises combining said at least first and second non-stationary outputs in an interwoven arrangement having alternating repetitions of said first and second non-stationary outputs; and said deep learning comprises employing an RNN network for time series prediction.

A method according to any preceding claim , wherein said extracting comprises extracting said at least one feature directly from said fused output.

A method according to any preceding claim, wherein said at least one machine comprises a group of machines performing a collaborative process.

A system for monitoring at least one machine, the system comprising:
a first sensor operative to obtain at least a first non-stationary signal from at least one machine operating in a non-stationary manner during at least one operating time frame; , said at least first sensor providing at least a first non-stationary output;
a second sensor, the second sensor operable to obtain at least a second non-stationary signal from the at least one machine during the operating time frame, the at least second sensor comprising: , providing at least a second non-stationarity output;
a signal processor operable to fuse the at least first non-stationary output with the at least second non-stationary output to produce a fused output;
a feature extractor operable to extract at least one feature of at least one of said first and second non-stationary signals; operable to analyze said at least one characteristic to ascertain health; and
a machine control module, the machine control module operable to control execution of at least one of a repair operation, a maintenance operation and a modification of an operating parameter of the at least one machine based on the condition; ,
said system.

13. The system for monitoring at least one machine according to claim 12 , wherein said signal processor is operable to modify said at least first non-stationary output based on said at least second non-stationary output. .

14. A method according to claim 12 or claim 13 , wherein said at least first non-stationary signal represents a mechanical state of said machine and said at least a second non-stationary signal represents an operational state of said machine. A system for monitoring at least one machine.

A system according to any one of claims 12 to 14 , wherein said at least one feature is extracted directly from said fused output.