Method and apparatus for system state classification

• US 7,065,422 B1
• Filed: 06/28/2002
• Issued: 06/20/2006
• Est. Priority Date: 06/28/2002
• Status: Active Grant

• Current Assignee: Advanced Micro Devices, Inc.
• Original Assignee: Advanced Micro Devices, Inc.
• Inventors: Green, Eric O.
• Primary Examiner(s): Gandhi, Jayprakash N.
• Application Number:
  US10/185,471
• Time in Patent Office: 1,453 Days
• Field of Search: 700/12, 700 32- 34, 700/49, 700/52, 700/73, 700/95, 700/96, 700/108, 700/111, 700/121, 702/181, 706/45
• US Class Current: 700/108
• CPC Class Codes: G05B 23/0221 Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
• CPC Class Codes: G05B 23/024 Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks

Abstract

A method for identifying a state of a manufacturing system includes defining at least one virtual sensor having a value. At least one state descriptor including a plurality of condition terms and a trigger probability is defined. Each condition term includes a function of the value of the virtual sensor. The condition terms are evaluated based on the value to determine a classification probability. The state is identified responsive to the classification probability being greater to or equal to the trigger probability. A manufacturing system including a plurality of tools, at least one virtual sensor, and a state classification unit is also provided.

Claims

1. A method for using a computer processor to identify a state of a manufacturing system, comprising;

defining at least one virtual sensor having a value;

defining at least one state descriptor including a plurality of Boolean condition terms and a trigger probability, each condition term including a logical function of the value of the virtual sensor;

evaluating the condition terms based on the value to determine a classification probability; and

identifying the state responsive to the classification probability being greater to or equal to the trigger probability.

2. The method of claim 1, further comprising generating a state classification message responsive to identifying the state.

3. The method of claim 2, wherein generating the state classification message further comprises generating at least one of an email message, a visual display, and a report.

4. The method of claim 1, wherein evaluating the condition terms further comprises determining a ratio between condition terms that are satisfied to a total number of condition terms to generate the classification probability.

5. The method of claim 1, wherein the virtual sensor is associated with a tool in the manufacturing system and defining the virtual sensor further comprises determining the value based on a process variable measured during a processing run of the tool.

6. The method of claim 1, wherein the virtual sensor is associated with a tool in the manufacturing system and defining the virtual sensor further comprises determining the value based on an output characteristic measured on a workpiece processed in the tool.

7. The method of claim 1, herein the virtual sensor is associated with a tool in the manufacturing system and defining the virtual sensor further comprises determining the value based on an environmental condition experienced by the tool.

8. The method of claim 1, wherein defining the virtual sensor further comprises associating the virtual sensor with a physical sensor on a tool in the manufacturing system.

9. The method of claim 1, wherein defining the viral sensor further comprises receiving a plurality of sensor values associated with a tool in the manufacturing system and determining the value based on a combination of the plurality of sensor values.

10. The method of claim 1, further comprising evaluating the condition terms and identifying the state at predetermined time intervals.

11. The method of claim 1, further comprising evaluating the condition terms and identifying the state responsive to a change in the value of the virtual sensor.

12. A method for using a computer processor to classify states of a manufacturing system, comprising:

providing a classification tree having a plurality of nodes for at least a portion of the manufacturing system;

defining a plurality of virtual sensors for the manufacturing system, each virtual sensor having a value;

defining a state descriptor including a plurality of Boolean condition terms and a trigger probability for each of a subset of the nodes, each condition term including a logical function of the value of at least one of the virtual sensors; and

for each of the state descriptors;

evaluating the condition terms based on the value to determine a classification probability; and

identifying a state associated with the state descriptor responsive to the classification probability being greater to or equal to the trigger probability.

13. The method of claim 12, further comprising generating a state classification message responsive to identifying the state.

14. The method of claim 13, wherein generating the state classification message further comprises generating at least one of an email message, a visual display, and a report.

15. The method of claim 12, wherein evaluating, the condition terms further comprises determining a ratio between condition terms that are satisfied to a total number of condition terms to generate the classification probability.

16. The method of claim 12, wherein at least one of the virtual sensors is associated with a tool in the manufacturing system and defining the at least one virtual sensor further comprises determining the value based on a process variable measured during a processing run of the tool.

17. The method of claim 12, wherein at least one of the virtual sensors is associated with a tool in the manufacturing system and defining the at least one virtual sensor further comprises determining the value based on an output characteristic measured on a workpiece processed in the tool.

18. The method of claim 12, wherein at least one of the virtual sensors is associated with a tool in the manufacturing system and defining the at least one virtual sensor further comprises determining the value based on an environmental condition experienced by the tool.

19. The method of claim 12, further comprising associating at least one of the virtual sensors with a physical sensor on a tool in the manufacturing system.

20. The method of claim 12, wherein deter the value for at least one of the virtual sensors further comprises receiving a plurality of sensor values associated with a tool in the manufacturing system and determining the value based on a combination of the plurality of sensor values.

21. The method of claim 12, further comprising evaluating the condition terms and identifying the state for each of the state descriptors at predetermined time intervals.

22. The method of claim 12, further comprising evaluating the condition terms and identifying the state for at least one of the state descriptors responsive to a change in the value of the associated virtual sensor.

23. A manufacturing system, comprising:

a plurality of tools;

at least one virtual sensor having a value associated with at least one of the tools; and

a state classification unit configured to define at least one state descriptor including a plurality of Boolean condition terms and a trigger probability, each condition term including a logical function of the value of the virtual sensor, evaluate the condition terms based on the value to determine a classification probability, and identify a state responsive to the classification probability being greater to or equal to the trigger probability.

24. The system of claim 23, wherein the state classification unit is further configured to generate a state classification message responsive to identifying the state.

25. The system of claim 24, wherein the state classification message further comprises at least one of an email message, a visual display, and a report.

26. The system of claim 23, wherein the state classification unit is further configured to determine a ratio between condition terms that are satisfied to a total number of condition terms to generate the classification probability.

27. The system of claim 23, wherein the viral sensor is associated with a tool in the manufacturing system and the value comprises a process variable measured during a processing run of the tool.

28. The system of claim 23, wherein the virtual sensor is associated with a tool in the manufacturing system and the value comprises an output characteristic measured on a workpiece processed in the tool.

29. The system of claim 23, wherein the virtual sensor is associated with a tool in the manufacturing system and the value comprises an environmental condition experienced by the tool.

30. The system of claim 23, further comprising a physical sensor associated with one of the plurality of tools, wherein the virtual sensor is associated with the virtual sensor.

31. The system of claim 23, wherein the state classification unit is further configured to receive a plurality of sensor values associated with one of the tools and determine the value based on a combination of the plurality of sensor values.

32. The system of claim 23, wherein the state classification unit is further configured to evaluate the condition ems and identify the state at predetermined time intervals.

33. The system of claim 23, wherein the state classification unit is further configured to evaluate the condition terms and identify the state responsive to a change in the value of the virtual sensor.

34. A manufacturing system, comprising:

a plurality of tools;

a plurality of virtual sensors, each having a value associated with at least one of the tools; and

a state classification unit configured to manufacturing a classification tree having a plurality of nodes for at least a portion of the manufacturing system, define a state descriptor including a plurality of Boolean condition terms and a trigger probability for each of a subset of the nodes, each condition term including a logical function of the value of at least one of the virtual sensors, and for each of the state descriptors, evaluate the condition terms based on the value to determine a classification probability and identify a state associated with the state descriptor responsive to the classification probability being greater to or equal to the trigger probability.

35. A system for identifying a state of a manufacturing system comprising:

means for defining at least one virtual sensor having a value;

means for defining at least one state descriptor including a plurality of Boolean condition terms and a trigger probability, each condition term including a logical function of the value of the virtual sensor;

means for evaluating the condition terms based on the value to determine a classification probability; and

means for ide identifying the state responsive to the classification probability being greater to or equal to the trigger probability.

36. A system for classifying states of a manufacturing system, comprising:

means for providing a classification tree having a plurality of nodes for at least a portion of the manufacturing system;

means for defining a plurality of virtual sensors for the manufacturing system, each virtual sensor having a value;

means for defining a state descriptor including a plurality of Boolean condition terms and a trigger probability for each of a subset of the nodes, each condition term including a logical function of the value of at least one of the virtual sensors; and

for each of the state descriptors;

means for evaluating the condition terms based on the value to determine a classification probability; and

means for identifying a state associated with the state descriptor responsive to the classification probability being greater to or equal to the trigger probability.