Industrial control system with machine learning for compressors

Patent metadata

Field	Value
Publication number	US-11340592-B2
Application number	US-201916518523-A
Country	US
Kind code	B2
Filing date	Jul 22, 2019
Priority date	Jul 22, 2019
Publication date	May 24, 2022
Grant date	May 24, 2022

Abstract

A compressor controller for operating a compressor within an industrial automation environment is provided. The compressor controller includes a control module, configured to control the compressor via control settings, and a machine learning module, coupled with the control module. The machine learning module is configured to receive a set of supervised data related to the compressor, and to train with the supervised data to produce a Newtonian physics model representing the inputs and outputs of the compressor within the industrial automation environment. The machine learning module is also configured to receive performance data related to the compressor, receive environment data related to the compressor, and to process the performance data and environment data to produce predicted future performance data for the compressor, and to produce control settings for the compressor.

First claim

What is claimed is:

1. A compressor controller for operating a compressor within an industrial automation environment, the compressor controller comprising:
a control module, configured to control the compressor via control settings; and
a machine learning module, coupled with the control module, and configured to:
receive a set of supervised data related to the compressor;
train with the supervised data to produce a Newtonian physics model representing the inputs and outputs of the compressor within the industrial automation environment;
receive performance data related to the compressor;
receive environment data related to the compressor; and
process the performance data and environment data to produce predicted future performance data for the compressor, and to produce control settings for the compressor.

2. The compressor controller of claim 1 , wherein the performance data comprises compressor status, guide vane position, blow off valve position, discharge pressure, flow rates, power supply data, and power consumption.

3. The compressor controller of claim 1 , wherein the control settings for the compressor are optimized to maximize through-put, minimize energy consumption, and to minimize starts and stops of the compressor.

4. The compressor controller of claim 1 , wherein the future performance data comprises predictions of bearing wear, turbine blade wear, required maintenance, and failure of the compressor.

5. The compressor controller of claim 1 , wherein the machine learning module is further configured to:
produce a schedule for compressor maintenance based at least in part on the future performance data for the compressor.

6. The compressor controller of claim 1 , wherein the machine learning module is further configured to:
display a warning when the future performance data predicts required maintenance or a failure in the compressor.

7. The compressor controller of claim 1 , wherein the environment data comprises vibration data.

8. A method for operating a compressor within an industrial automation environment, the method comprising:
receiving a set of supervised data related to the compressor;
training a machine learning module with the supervised data to produce a Newtonian physics model representing the inputs and outputs of the compressor within the industrial automation environment;
receiving performance data related to the compressor;
receiving environment data related to the compressor; and
processing the performance data and environment data in the trained machine learning module to produce predicted future performance data for the compressor, and to produce control settings for the compressor.

9. The method of claim 8 , wherein the performance data comprises compressor status, guide vane position, blow off valve position, discharge pressure, flow rates, power supply data, and power consumption.

10. The method of claim 8 , wherein the control settings for the compressor are optimized to maximize through-put, minimize energy consumption, and to minimize starts and stops of the compressor.

11. The method of claim 8 , wherein the future performance data comprises predictions of bearing wear, turbine blade wear, required maintenance, and failure of the compressor.

12. The method of claim 8 , further comprising:
producing a schedule for compressor maintenance based at least in part on the future performance data for the compressor.

13. The method of claim 8 , further comprising:
displaying a warning when the future performance data predicts required maintenance or a failure in the compressor.

14. The method of claim 8 , wherein the environment data comprises vibration data.

15. One or more non-transitory computer-readable media having stored thereon program instructions to operate a compressor within an industrial automation environment, wherein the program instructions, when executed by a computing system, direct the computing system to at least:
receive a set of supervised data related to the compressor;
train a machine learning module with the supervised data to produce a Newtonian physics model representing the inputs and outputs of the compressor within the industrial automation environment;
receive performance data related to the compressor;
receive environment data related to the compressor; and
process the performance data and environment data in the machine learning module to produce predicted future performance data for the compressor, and to produce control settings for the compressor.

16. The one or more non-transitory computer-readable media of claim 15 , wherein the performance data comprises compressor status, guide vane position, blow off valve position, discharge pressure, flow rates, power supply data, and power consumption.

17. The one or more non-transitory computer-readable media of claim 15 , wherein the control settings for the compressor are optimized to maximize through-put, minimize energy consumption, and to minimize starts and stops of the compressor.

18. The one or more non-transitory computer-readable media of claim 15 , wherein the future performance data comprises predictions of bearing wear, turbine blade wear, required maintenance, and failure of the compressor.

19. The one or more non-transitory computer-readable media of claim 15 , further comprising program instructions, which when executed by the computing system, direct the computing system to at least:
produce a schedule for compressor maintenance based at least in part on the future performance data for the compressor.

20. The one or more non-transitory computer-readable media of claim 15 , further comprising program instructions, which when executed by the computing system, direct the computing system to at least:
display a warning when the future performance data predicts required maintenance or a failure in the compressor.

21. The one or more non-transitory computer-readable media of claim 15 , further comprising program instructions, which when executed by the computing system, direct the computing system to at least:
produce the predicted future performance data for the compressor.

22. The one or more non-transitory computer-readable media of claim 15 , further comprising program instructions, which when executed by the computing system, direct the computing system to at least:
produce the control settings for the compressor.

23. The one or more non-transitory computer-readable media of claim 15 , wherein the environment data comprises vibration data.

Assignees

• Rockwell Automation Tech Inc

Inventors

• Mazur David C
• Marshall Steven
• Bromfield Scotty
• Entzminger Rob Alan

Classifications

• Y02P90/80

  Management or planning · CPC title

• G05B13/0265

  the criterion being a learning criterion · CPC title

• F05D2260/82

  Forecasts · CPC title

• F05D2260/80

  Diagnostics · CPC title

• F05D2260/81

  Modelling or simulation · CPC title