Method for compressing wideband sensor signal

Patent metadata

Field	Value
Publication number	US-11378442-B2
Application number	US-202016743888-A
Country	US
Kind code	B2
Filing date	Jan 15, 2020
Priority date	Jan 15, 2020
Publication date	Jul 5, 2022
Grant date	Jul 5, 2022

Abstract

Example implementations described herein are directed to systems and methods for extracting signal in the presence of noise for industrial IoT systems. Through the example implementations described herein, the high-frequency band signal of the sensor output can be maintained despite data compression, while also retaining information regarding the condition of the machine. Example implementations can also generate compressed data pairs to synchronized downsampled envelopes and spectrum data.

First claim

What is claimed is:

1. A method for managing sensor data from a machine having a plurality of machine states, the method comprising:
decomposing sensor data into a plurality of band-limited data to generate a decomposed sensor data; and
executing downsampling on each of the decomposed sensor data to generate downsampled decomposed sensor data;
wherein the downsampling cycle is shorter than a duration of the each of the machine states, and
wherein the downsampled decomposed sensor data is implemented to identify a signal frequency band containing a signal developed by the machine
where the method further includes:
extracting an envelope from the sensor data;
executing downsampling on the extracted envelope to generate a downsampled envelope;
synchronizing the downsampled envelope with the downsampled decomposed sensor data as data pairs,
collecting the data pairs until a predetermined condition is met;
identifying, from the downsampled decomposed sensor data, the signal frequency band having a signal value satisfying another predetermined condition, and
storing the identified signal frequency band.

2. The method of claim 1 , further comprising:
extracting feature values from the each of the decomposed sensor data,
wherein the downsampling is conducted based on the extracted feature values to generate the downsampled decomposed sensor data.

3. The method of claim 1 , further comprising:
extracting feature values from each of the decomposed sensor data,
wherein the downsampling is conducted based on the extracted feature values to generate the downsampled decomposed sensor data, and
wherein the extracting feature values is conducted from the decomposed sensor data of the identified signal frequency band.

4. The method of claim 3 , wherein generation of subsequent data is conducted through a demodulator and a low-pass filter configured according to the identified signal frequency band.

5. The method of claim 1 , wherein the decomposing the sensor data is conducted through a Fast Fourier Transform (FFT).

6. The method of claim 1 , wherein the method is executed by a processor integrated with a sensor configured to provide the sensor data with a common enclosure.

7. A non-transitory computer readable medium, storing instructions for managing sensor data from a machine having a plurality of machine states, the instructions comprising:
decomposing sensor data into a plurality of band-limited data to generate a decomposed sensor data; and
executing downsampling on each of the decomposed sensor data to generate downsampled decomposed sensor data;
wherein the downsampling cycle is shorter than a duration of the each of the machine states,
wherein the downsampled decomposed sensor data is implemented to identify a signal frequency band containing a signal developed by the machine
wherein the instructions further includes:
extracting an envelope from the sensor data;
executing downsampling on the extracted envelope to generate a downsampled envelope; and
synchronizing the downsampled envelope with the downsampled decomposed sensor data as data pairs
collecting the data pairs until a predetermined condition is met;
identifying, from the downsampled decomposed sensor data, the signal frequency band having a signal value satisfying another predetermined condition; and
storing the identified signal frequency band.

8. The non-transitory computer readable medium of claim 7 , the instructions further comprising:
extracting feature values from the each of the decomposed sensor data,
wherein the downsampling is conducted based on the extracted feature values to generate the downsampled decomposed sensor data.

9. The non-transitory computer readable medium of claim 7 , the instructions further comprising:
extracting feature values from each of the decomposed sensor data;
wherein the downsampling is conducted based on the extracted feature values to generate the downsampled decomposed sensor data, and
wherein the extracting feature values is conducted from the decomposed sensor data of the identified signal frequency band.

10. The non-transitory computer readable medium of claim 9 , wherein generation of subsequent data is conducted through a demodulator and a low-pass filter configured according to the identified signal frequency band.

11. The non-transitory computer readable medium of claim 7 , wherein the decomposing the sensor data is conducted through a Fast Fourier Transform (FFT).

12. The non-transitory computer readable medium of claim 7 , wherein the instructions are executed by a processor integrated with a sensor configured to provide the sensor data with a common enclosure.

Assignees

• Hitachi Ltd

Inventors

• Maeda Daisuke
• Gaur Sudhanshu

Classifications

• G01H1/003

  of rotating machines (G01H1/10 takes precedence) · CPC title

• B23Q17/0976

  Detection or control of chatter (B23Q15/12 takes precedence) · CPC title

• Y02P90/80

  Management or planning · CPC title

• G01H1/14 Primary

  Frequency · CPC title

• G01M13/045

  Acoustic or vibration analysis · CPC title