On-line monitoring and diagnostics of a process using multivariate statistical analysis

• US 7,853,431 B2
• Filed: 03/20/2007
• Issued: 12/14/2010
• Est. Priority Date: 09/29/2006
• Status: Active Grant

• Current Assignee: Fisher-Rosemount Systems Incorporated (Emerson Electric Company)
• Original Assignee: Fisher-Rosemount Systems Incorporated (Emerson Electric Company)
• Inventors: Samardzija, Nikola, Miller, John Philip
• Primary Examiner(s): Wachsman; Hal D
• Assistant Examiner(s): Huynh; Phuong
• Application Number:
  US11/688,737
• Publication Number:
  US 20080082299A1
• Time in Patent Office: 1,365 Days
• Field of Search: 702179-189, 700/9, 700/10, 700/96, 700/108, 700/116, 700/117, 700/118, 700/146
• US Class Current: 702/179
• CPC Class Codes: G05B 23/021 adopting a different treatment of each operating region or a different mode of the monitored system, e.g. transient modes; different operating configurations of monitored system
• 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
• CPC Class Codes: G05B 23/0245 based on a qualitative model, e.g. rule based; if-then decisions
• CPC Class Codes: Y10S 715/965 for process control and configuration

Abstract

A system and method of monitoring and diagnosing on-line multivariate process variable data in a process plant, where the multivariate process data comprises a plurality of process variables each having a plurality of observations, includes collecting on-line process data from a process control system within the process plant when the process is on-line, where the collected on-line process data comprises a plurality of observations of a plurality of process variables and where the plurality of observations of the set of collected process data comprises a first data space having a plurality of dimensions, performing a multivariate statistical analysis to represent the operation of the process based on a set of collected on-line process data comprising a measure of the operation of the process when the process is on-line within a second data space having fewer dimensions than the first data space, performing a univariate analysis to represent the operation of the process as a multivariate projection of the on-line process data by a univariate variable for each of the process variables, where the univariate variable unifies the process variables, and generating a visualization comprising a first plot of a result generated by the multivariate statistical representation of the operation of the process and a second plot of a result generated by the univariate representation of the operation of the process.

Claims

1. A system for monitoring and diagnosing on-line multivariate process variable data in a process plant, wherein the multivariate process data comprises a plurality of process variables each having a plurality of observations, the system comprising:

a data collection tool adapted to collect on-line process data from a process control system within the process plant when the process is on-line, wherein the collected on-line process data comprises a plurality of observations of a plurality of process variables and wherein the plurality of observations of a set of collected on-line process data comprises a first data space having a plurality of dimensions;

a first analysis tool comprising a multivariate statistical analysis engine adapted to represent the operation of the process based on the set of collected on-line process data comprising a measure of the operation of the process when the process is on-line within a second data space having fewer dimensions than the first data space;

a second analysis tool comprising a univariate analysis engine adapted to represent the operation of the process as a multivariate projection of the set of collected on-line process data by a univariate variable for each of the process variables, wherein the univariate variable unifies the process variables, and the univariate analysis engine is adapted to monitor the set of collected on-line process data with the univariate variable; and

a display tool adapted to generate a visualization comprising a first plot of a result generated by the multivariate statistical representation of the operation of the process and a second plot of a result generated by the univariate representation of the operation of the process.

2. The system of claim 1, wherein the first plot of a result generated by the multivariate statistical representation of the operation of the process comprises a plot of a score generated by the multivariate statistical representation of the operation of the process in relation to at least one component used to define the second data space.

3. The system of claim 2, wherein the display tool is adapted to enable a user to select the at least one component used to define the second data space.

4. The system of claim 2, wherein the score comprises a representation of one or more of the plurality of observations in the second data space.

5. The system of claim 1, wherein the second analysis tool is adapted to calculate a multivariable transformation as a function of a plurality of process variable vector transformations each corresponding to one of the process variables, wherein each process variable vector transformation is a function of the univariate variable unifying the process variables, and wherein the second plot comprises a plot of the multivariate transformation as a function of the univariate variable.

6. The system of claim 1, wherein the second plot comprises a plot of each process variable with a unique position relative to the univariate variable as determined by a position parameter unique to a process variable vector.

7. The system of claim 1, wherein the second plot comprises a plot of a multivariate transformation as a function of the univariate variable.

8. The system of claim 1, wherein the data collection tool is adapted to collect on-line process data representative of a real-time on-line operation of the process, and wherein the display tool is adapted to generate a real-time visualization of the operation of the process comprising the first and second plots.

9. The system of claim 8, wherein the display tool is adapted to update the real-time visualization of the operation of the process in real-time as the data collection tool collects the on-line process data representative of the real-time on-line operation of the process.

10. The system of claim 1, wherein the first analysis tool is adapted to perform a statistical analysis on the collected on-line process data to determine a first threshold value, and wherein the second analysis tool is adapted to normalize the plurality of process variables to determine a second threshold value, wherein the second threshold value comprises a common process variable control limit for the process variables.

11. The system of claim 10, wherein the first plot comprises a plot of the result generated by the multivariate statistical representation of the operation of the process in relation to the first threshold value.

12. The system of claim 10, wherein each process variable comprises a process variable control limit and corresponds to a process variable vector transformation as a function of a univariate variable unifying the process variables, wherein the second analysis tool is adapted to normalize each process variable vector transformation to the same process variable control limit using a scaling parameter unique to each process variable to generate the common process variable control limit for the process variables, wherein the second plot comprises a plot of a result generated by the univariate representation of the operation of the process in relation to the common process variable limit.

13. The system of claim 10, wherein the first analysis tool is adapted to determine the presence of an abnormal process condition if a result generated by the multivariate statistical representation of the operation of the process exceeds the first threshold value.

14. The system of claim 10, wherein the second analysis tool is adapted to determine the presence of an abnormal process condition if a result generated by the univariate representation of the operation of the process exceeds the second threshold value.

15. The system of claim 10, wherein the display tool is adapted to generate an alarm if a result generated by the multivariate statistical representation exceeds the first threshold value and if a result generated by the univariate representation exceeds the second threshold value.

16. The system of claim 15, wherein the display tool is adapted to identify the process variable corresponding to an abnormal process condition if a result generated by the multivariate statistical representation exceeds the first threshold value and if a result generated by the univariate representation exceeds the second threshold value.

17. The system of claim 15, wherein the first analysis tool is adapted to generate a plurality of components that define the second data space, wherein each component corresponds to a degree of variance in the process data for a process variable indicating a degree of significance of the process variable on the process, wherein the display tool is adapted to classify a severity of the alarm based on a process variable associated with the alert and based on the degree of significance of the process variable on the process.

18. The system of claim 10, wherein the display tool is adapted to diagnose a false alarm if a result generated by the multivariate statistical representation exceeds the first threshold value and if a result generated by the univariate representation does not exceed the second threshold value.

19. The system of claim 10, wherein the display tool is adapted to diagnose a missed alarm if a result generated by the multivariate statistical representation does not exceed the first threshold value and if a result generated by the univariate representation exceeds the second threshold value.

20. The system of claim 10, wherein the first threshold value comprises a statistical limit of a process variable.

21. The system of claim 10, wherein the second threshold value comprises a physical limit of a process variable.

22. The system of claim 1, wherein the display tool is adapted to enable a user to select a set of previously collected on-line process data, wherein the multivariate statistical analysis engine is adapted to represent a previous operation of the process based on the set of the previously collected on-line process data comprising a measure of the operation of the process when the process was on-line, wherein the univariate analysis engine is adapted to represent the previous operation of the process as a multivariate projection of the previously collected on-line process data by a univariate variable for each of the process variables, and wherein the display tool is adapted to generate a visualization comprising a first plot of a result generated by the multivariate statistical representation of the previous operation of the process and a second plot of a result generated by the univariate representation of the previous operation of the process.

23. A method of monitoring and diagnosing on-line multivariate process variable data in a process plant, wherein the multivariate process data comprises a plurality of process variables each having a plurality of observations, the method comprising:

collecting on-line process data from a process control system within the process plant when the process is on-line, wherein the collected on-line process data comprises a plurality of observations of a plurality of process variables and wherein the plurality of observations of a set of collected on-line process data comprises a first data space having a plurality of dimensions;

performing a multivariate statistical analysis to represent the operation of the process based on the set of collected on-line process data comprising a measure of the operation of the process when the process is on-line within a second data space having fewer dimensions than the first data space;

performing a univariate analysis to represent the operation of the process as a multivariate projection of the set of collected on-line process data by a univariate variable for each of the process variables, wherein the univariate variable unifies the process variables, and the univariate analysis monitors the set of collected on-line process data with the univariate variable; and

generating a visualization comprising a first plot of a result generated by the multivariate statistical representation of the operation of the process and a second plot of a result generated by the univariate representation of the operation of the process.

24. The method of claim 23, wherein the first plot of a result generated by the multivariate statistical representation of the operation of the process comprises a plot of a score generated by the multivariate statistical representation of the operation of the process in relation to at least one component used to define the second data space.

25. The method of claim 24, further comprising enabling a user to select the at least one component used to define the second data space.

26. The method of claim 24, wherein the score comprises a representation of one or more of the plurality of observations in the second data space.

27. The method of claim 23, wherein performing a univariate analysis comprises calculating a multivariable transformation as a function of a plurality of process variable vector transformations each corresponding to one of the process variables, wherein each process variable vector transformation is a function of the univariate variable unifying the process variables, and wherein the second plot comprises a plot of the multivariate transformation as a function of the univariate variable.

28. The method of claim 23, wherein the second plot comprises a plot of each process variable with a unique position relative to the univariate variable as determined by a position parameter unique to the process variable vector.

29. The method of claim 23, wherein the second plot comprises a plot of a multivariate transformation as a function of the univariate variable.

30. The method of claim 23, wherein collecting on-line data comprises collecting on-line process data representative of a real-time on-line operation of the process, and wherein generating a visualization comprises generating a real-time visualization of the operation of the process comprising the first and second plots.

31. The method of claim 30, wherein generating a visualization comprises updating the real-time visualization of the operation of the process in real-time concurrently with collecting the on-line process data representative of the real-time on-line operation of the process.

32. The method of claim 23, wherein performing a multivariate statistical analysis comprises performing a statistical analysis on the collected on-line process data to determine a first threshold value, and wherein performing a univariate analysis comprises normalizing the plurality of process variables to determine a second threshold value, wherein the second threshold value comprises a common process variable control limit for the process variables.

33. The method of claim 32, wherein the first plot comprises a plot of the result generated by the multivariate statistical representation of the operation of the process in relation to the first threshold value.

34. The method of claim 32, wherein each process variable comprises a process variable control limit and corresponds to a process variable vector transformation as a function of a univariate variable unifying the process variables, wherein performing a univariate analysis comprises normalizing each process variable vector transformation to the same process variable control limit using a scaling parameter unique to each process variable to generate the common process variable control limit for the process variables, wherein the second plot comprises a plot of a result generated by the univariate representation of the operation of the process in relation to the common process variable limit.

35. The method of claim 32, further comprising determining the presence of an abnormal process condition if a result generated by the multivariate statistical representation of the operation of the process exceeds the first threshold value.

36. The method of claim 32, further comprising determining the presence of an abnormal process condition if a result generated by the univariate representation of the operation of the process exceeds the second threshold value.

37. The method of claim 32, further comprising generating an alarm if a result generated by the multivariate statistical representation exceeds the first threshold value and if a result generated by the univariate representation exceeds the second threshold value.

38. The method of claim 37, further comprising identifying the process variable corresponding to an abnormal process condition if a result generated by the multivariate statistical representation exceeds the first threshold value and if a result generated by the univariate representation exceeds the second threshold value.

39. The method of claim 37, wherein performing a multivariate statistical analysis comprises generating a plurality of components that define the second data space, wherein each component corresponds to a degree of variance in the process data for a process variable indicating a degree of significance of the process variable on the process, the method further comprising classifying a severity of the alarm based on a process variable associated with the alert and based on the degree of significance of the process variable on the process.

40. The method of claim 32, further comprising diagnosing a false alarm if a result generated by the multivariate statistical representation exceeds the first threshold value and if a result generated by the univariate representation does not exceed the second threshold value.

41. The method of claim 32, further comprising diagnosing a missed alarm if a result generated by the multivariate statistical representation does not exceed the first threshold value and if a result generated by the univariate representation exceeds the second threshold value.

42. The method of claim 32, wherein the first threshold value comprises a statistical limit of a process variable.

43. The method of claim 32, wherein the second threshold value comprises a physical limit of a process variable.

44. The method of claim 23, wherein performing a multivariate statistical analysis comprises performing a multivariate statistical analysis to represent a previous operation of the process based on a set of previously collected on-line process data comprising a measure of the operation of the process when the process was on-line, wherein performing a univariate analysis comprises performing a univariate analysis to represent the previous operation of the process as a multivariate projection of the previously collected on-line process data by a univariate variable for each of the process variables, the method further comprising:

enabling a user to select a previous set of previously collected on-line process data; and

generating a visualization comprising a first plot of a result generated by the multivariate statistical representation of the previous operation of the process and a second plot of a result generated by the univariate representation of the previous operation of the process.