Method and device for recognizing state of noise-generating machine to be investigated

Patent metadata

Field	Value
Publication number	US-9714884-B2
Application number	US-73667509-A
Country	US
Kind code	B2
Filing date	Apr 29, 2009
Priority date	Apr 29, 2008
Publication date	Jul 25, 2017
Grant date	Jul 25, 2017

Abstract

A statistical basic classification model of acoustic features generated for at least one reference object is automatically adapted by a data processing unit based on acoustic features of a noise generated by an object to be investigated to obtain an individually adapted statistical classification model. The data processing unit then classifies the state of the noise-generating object based on the individually adapted statistical classification model.

First claim

The invention claimed is:

1. A method for recognizing an operational state of a noise-generating object to be investigated that generates noise during operation, where the noise is not purposely generated by the noise-generating object, but results from the operation thereof, comprising:
performing a two-stage process for generating an individually adapted statistical classification model of acoustic features, comprising:
(a) in a first stage, generating a statistical basic classification model of acoustic features based on a plurality of acoustic sound recordings made during operation of at least one reference object in different conditions or at different times, the statistical basic classification model defining one or more acoustic features corresponding to different operational states of the at least one reference object and a statistical parameter for at least one acoustic feature; and
(a) in a second stage, subsequent to the generation of the statistical basic classification model, generating an individually adapted statistical classification model specific to the noise-generating object by:
making sound recordings of an operation of the noise-generating object, the noise-generating object being separate from the at least one reference object;
determining acoustic feature measurements from the sound recordings of the noise-generating object;
automatically adapting the statistical basic classification model of acoustic features generated for the at least one reference object, based on the acoustic feature measurements determined from the sound recording signals of the operation of the noise-generating object to obtain the individually adapted statistical classification model specific to the noise-generating object, wherein adapting the statistical basic classification model includes adjusting the statistical parameter for the at least one acoustic feature;

applying the individually adapted statistical classification model, including the adjusted statistical parameter for the at least one acoustic feature, to further acoustic feature measurements determined from further sound recordings of the operation of the noise-generating object; and
classifying, by the processor, the operational state of the noise-generating object based on the application of the individually adapted statistical classification model to the further acoustic feature measurements.

2. The method as claimed in claim 1 , wherein the at least one reference object is formed by a prototype of the noise-generating object to be investigated.

3. The method as claimed in claim 1 , wherein the noise-generating object has at least one noise-generating module.

4. The method as claimed in claim 1 , wherein said adapting of the statistical basic classification model occurs during commissioning of the noise-generating object.

5. The method as claimed in claim 1 , further comprising storing the statistical basic classification model in a memory.

6. The method as claimed in claim 5 , further comprising buffering the individually adapted statistical classification model.

7. The method as claimed in claim 6 , further comprising making the acoustic sound recordings using acoustic sound pickups which capture at least one of air-borne sound and structure-borne sound.

8. The method as claimed in claim 7 , further comprising at least one of attaching the acoustic sound pickups to the noise-generating object and relative movement of the acoustic sound pickups and the noise-generating object.

9. The method as claimed in claim 8 , further comprising calculating an adapted statistical classification model for different possible positions of the acoustic sound pickup passing the noise-generating object.

10. The method as claimed in claim 9 , wherein the different possible positions of the acoustic sound pickup correspond to a spatial arrangement of different modules of the noise-generating object.

11. The method as claimed in claim 1 , further comprising making the acoustic sound recordings using acoustic sound pickups which capture at least one of air-borne sound and structure-borne sound.

12. The method as claimed in claim 11 , further comprising at least one of attaching the acoustic sound pickups to the noise-generating object and relative movement of the acoustic sound pickups and the noise-generating object.

13. The method as claimed in claim 12 , further comprising calculating an adapted statistical classification model for different possible positions of the acoustic sound pickup passing the noise-generating object.

14. The method as claimed in claim 13 , wherein the different possible positions of the acoustic sound pickup correspond to a spatial arrangement of different modules of the noise-generating object.

15. The method as claimed in claim 1 , wherein said adapting adapts the statistical classification model for the noise-generating object as a whole or for modules of the noise-generating object.

16. The method as claimed in claim 1 , comprising automatically adapting the statistical basic classification model at regular maintenance intervals.

17. The method as claimed in claim 1 , comprising automatically adapting the statistical basic classification model when the noise-generating object is moved.

18. The method as claimed in claim 1 , wherein the noise-generating object and the at least one reference object comprise distinct instances of the same type object.

19. The method as claimed in claim 1 , wherein classifying, by the processor, the operational state of the noise-generating object based on the individually adapted statistical classification model comprises classifying the noise-generating object as defective.

20. A device for recognizing an operational state of a noise-generating object to be investigated that generates noise during operation, where the noise is not purposely generated by the noise-generating object, but results from the operation thereof, the device comprising:
a non-transitory memory unit storing a statistical basic classification model of acoustic features based on a plurality of acoustic sound recordings made during operation of at least one reference object in different conditions or at different times, the statistical basic classification model defining one or more acoustic features corresponding to different operational states of the at least one reference object and a statistical parameter for at least one acoustic feature;
a data processing unit communicatively coupled to the memory and configured to:
receive sound recording signals from sound recordings of an operation of the noise-generating object, the noise-generating object being separate from the at least one reference object;
determine acoustic feature measurements from the sound recording signals of the operation of the noise-generating object;
adapt the statistical basic classification model of acoustic features generated for the at least one reference object, based on the acoustic feature measurements determined from the sound recording signals of the operation of the noise-generating object to obtain an individually adapted statistical classification model specific to the noise-generating object, wherein adapting the statistical basic classification model includes adjusting the statistical parameter for the at least one acoustic feature;
receive further sound recordings of the operation of the noise-generating object, and determine further acoustic feature measurements from the further sound recordings of the noise-generating object;
apply the individually ad

Assignees

• Hofer Joachim
• Leutelt Lutz
• Siemens Ag

Inventors

• Hofer Joachim
• Leutelt Lutz

Classifications

• G06N99/00

  Subject matter not provided for in other groups of this subclass · CPC title

• G01M15/12 Primary

  by monitoring vibrations · CPC title