System, method and apparatus for determining properties of product or process streams

• US 10,175,661 B2
• Filed: 02/05/2014
• Issued: 01/08/2019
• Est. Priority Date: 02/05/2013
• Status: Active Grant

• Current Assignee: YokoGawa Corporation of America (Yokogawa Electric Corporation)
• Original Assignee: YokoGawa Corporation of America (Yokogawa Electric Corporation)
• Inventors: Trygstad, W. Marcus
• Primary Examiner(s): Dao, Tuan C
• Application Number:
  US14/173,301
• Publication Number:
  US 20140309756A1
• Time in Patent Office: 1,798 Days
• Field of Search: None
• CPC Class Codes: G05B 13/04 involving the use of models or simulators
• CPC Class Codes: G05B 13/048 using a predictor
• CPC Class Codes: G06F 30/20 Design optimisation, verification or simulation optimisation, verification or simulation of circuit designs G06F30/30

Abstract

Systems, methods, and apparatuses are provided for determining properties of process streams, in particular, hydrocarbon processing streams. The systems, methods, and apparatuses frequently, for example, substantially in real-time, determine measurements for the properties of the process stream. The systems, methods, and apparatuses provide features that allow such properties of process streams to be accurately measured even as process conditions and other parameters that affect process operations change. More specifically, an analyzer having a measurement device configured to detect one or more independent variables of a process stream, a model configured to determine one or more analyzer measurements from the one or more independent variables, and a procedure to adjust the model using a corresponding primary measurement is disclosed.

Claims

1. A system for determining a property of a fluid process stream, the system comprising:

a hard analyzer including hardware, configured to generate one or more dependent variables associated with a sample of the fluid process stream, wherein said dependent variables are properties that correspond to composition of the sample, and wherein said dependent variables exclude independent process variables of pressure, temperature, and flow rate;

a distributed control system (DCS) having a plurality of sensors including at least one temperature sensor, at least one pressure sensor, and at least one flow rate sensor, communicably coupled to the fluid process stream;

the DCS configured to use said sensors to capture a plurality of independent process variables associated with the fluid process stream, and to use said plurality of independently process variables as process control inputs to control the fluid process stream;

an inferential analyzer including a model of aspects of the fluid process stream, the inferential analyzer being configured to capture from said DCS said plurality of independent process variables associated with the process stream and to use said plurality of independent process variables to generate one or more inferred dependent variables, wherein said plurality of independent process variables are not said dependent variables;

a comparison module configured to generate a comparison of the dependent variables and the inferred dependent variables;

the system being configured to use the comparison to adjust the model; and

the DCS configured to use the comparison to adjust one or more parameters of the fluid process stream.

2. The system of claim 1, wherein the system is configured to control the process stream by adjusting values for the inferred dependent variables.

3. The system of claim 1, further comprising a cognitive quality manager (CQM) configured to use the comparison to adjust the model.

4. The system of claim 1, further comprising a process optimizer configured to use the comparison to control the process stream.

5. The system of claim 4, further comprising a process control system operatively engaged with the process optimizer, the process control system configured to adjust one or more parameters of the process stream in response to receipt of instructions from the process optimizer.

6. The system of claim 1, wherein the inferential analyzer is configured to capture the plurality of independent process variables over a discrete interval.

7. The system of claim 6, wherein the discrete interval is at most about 240 minutes.

8. The system of claim 7, wherein the discrete interval is at most about 180 minutes.

9. The system of claim 8, wherein the discrete interval is at most about 120 minutes.

10. The system of claim 9, wherein the discrete interval is at most about 90 minutes.

11. The system of claim 10, wherein the discrete interval is at most about 60 minutes.

12. The system of claim 1, wherein the model comprises at least one of a statistical model and a statistical multivariate model.

13. The system of claim 1, wherein at least one of the independent process variables, the dependent variables, and the inferred dependent variables is discretized.

14. The system of claim 1, wherein the independent process variables include one or more of temperature, pressure, rate of addition or consumption of a catalyst or reagent, flow rate, the input or output of thermal energy to the process, refractive index, conductivity, viscosity, opacity, pH, density, intensity responses measured by photometric or spectrometric devices at particular wavelengths or frequencies, and combinations thereof.

15. The system of claim 1, being further configured to aggregate the independent process variables into a property spectrum.

16. The system of claim 15, wherein the system is configured to use the property spectrum to update the model.

17. The system of claim 1, wherein the comparison includes a validation of at least one of the independent process variables, the dependent variables, and the inferred dependent variables.

18. The system of claim 17, being configured to adjust the model and/or control the process stream once the dependent variables are validated.

19. The system of claim 18, being configured to determine whether to use the validated dependent variables to adjust the model based upon at least one of an age of the validated dependent variables and a current state of the process as compared to a state of the process at the time the validated dependent variable was measured.

20. The system of claim 17, wherein the comparison module is configured to generate the comparison using a statistical multivariate procedure.

21. The system of claim 17, wherein the comparison module is configured to generate the comparison using a procedure selected from the group consisting of:

(a) a multivariate method to determine whether the sample spectrum belongs to a population of sample spectra which includes a calibration sample set;

(b) multivariate method to classify any spectral outlier according to cause;

(c) replacing any false independent variable within the outlier sample spectrum;

(d) determining whether any outlier condition is due to a particular sensor by comparing sensor output to a baseline;

(e) correcting the outlier condition by replacement of a false independent variable with a model value;

(f) using averaging, regression, a damping function, or a vector-based algorithm, to provide a mathematical relationship between the inferred dependent variables and the dependent variables, which mathematical relationship is then applied to subsequent inferential predictions to produce a corrected inferential value;

(g) estimating a lag time of the dependent variables relative to the inferred dependent variables;

(h) applying a statistical analysis to the dependent variables and the inferred dependent variables;

(i) updating the model with the addition of recent data to and the elimination of older data from the data set used in creation of the model;

(j) updating the model by adding sample data for spectral outliers and/or prediction outliers to the data set used in creation of the model;

(k) capturing a sample that has been identified as a spectral outlier or a property outlier, and using it to adjust outputs of the inferential analyzer;

(l) performing model updates on the basis of schedule;

(m) performing model updates on the basis of need as determined by factors including frequency, severity of spectral outliers and/or prediction outliers, and/or changing of process operation or configuration;

(n) applying a goodness of fit test procedure; and

combinations thereof.

22. The system of claim 1, wherein the hard analyzer comprises a gas chromatograph configured to generate the dependent variables by capturing and evaluating a chromatogram of a sample of the process stream.

23. The system of claim 1, wherein the hard analyzer is at least one of a gas chromatograph (GC), a fast GC, an ultrafast GC, a Fourier Transform Infrared (FTIR) analyzer, a Fourier Transform Near-Infrared (FTNIR) analyzer, a nuclear magnetic resonance (NMR) analyzer, a Raman analyzer, a flash point analyzer, a distillation property analyzer, a cloud point analyzer, a freezing point analyzer, a cold filter plugging point analyzer, a pour point analyzer, and a combination thereof.

24. The system of claim 1, wherein the dependent variables include one or more of octane, percent total aromatics, percent benzene, percent olefins, distillation properties, Reid Vapor Pressure (RVP), cetane number, distillation properties, polynuclear aromatics, polycyclic aromatics, cloud point, pour point, cold filter plugging point, PIONA, PONA, distillation properties, and combinations thereof.

25. The system of claim 1, wherein the inferential analyzer is configured to generate the one or more inferred dependent variables in real time, and the hard analyzer is an online analyzer that generates dependent variables with a frequency that is lower than the frequency with which inferred dependent variables are generated, and wherein the comparison is used to adjust the model and/or to control the process stream in real time.

26. A method for determining a property of a process stream, the method comprising:

generating, with a hard analyzer including hardware, one or more dependent variables associated with a sample of the process stream, wherein said dependent variables are properties that correspond to composition of the sample, and wherein said dependent variables exclude independent process variables of pressure, temperature, and flow rate;

communicably coupling a distributed control system (DCS) having a plurality of sensors including at least one temperature sensor, at least one pressure sensor, and at least one flow rate sensor, to the fluid process stream;

using the DCS and the sensors to capture a plurality of independent process variables associated with the fluid process stream, and using said plurality of independent process variables as process control inputs to control the fluid process stream;

capturing from the DCS, with an inferential analyzer including a model of aspects of the process stream, said plurality of independent process variables associated with the process stream and using said plurality of independent process variables to generate one or more inferred dependent variables, wherein said plurality of independent process variables are not said dependent variables;

generating, with a comparison module, a comparison of the dependent variables and the inferred dependent variables;

using the comparison to adjust the model; and

with the DCS, using the comparison to adjust one or more parameters of the fluid process stream.

27. The method of claim 26, wherein said using the comparison further comprises using the comparison to control the process stream by adjusting values for the inferred dependent variables.

28. The method of claim 26, further comprising using a cognitive quality manager (CQM) to adjust the model.

29. The method of claim 26, further comprising using a process optimizer to control the process stream.

30. The method of claim 29, further comprising using a process control system operatively engaged with the process optimizer, to adjust one or more parameters of the process stream in response to receipt of instructions from the process optimizer.

31. The method of claim 26, further comprising using the inferential analyzer to capture the one or more independent process variables over a discrete interval.

32. The method of claim 31, wherein the discrete interval is at most about 240 minutes.

33. The method of claim 32, wherein the discrete interval is at most about 180 minutes.

34. The method of claim 33, wherein the discrete interval is at most about 120 minutes.

35. The method of claim 34, wherein the discrete interval is at most about 90 minutes.

36. The method of claim 35, wherein the discrete interval is at most about 60 minutes.

37. The method of claim 26, further comprising the inferential analyzer using a model in the form of at least one of a statistical model and a statistical multivariate model.

38. The method of claim 26, further comprising discretizing at least one of the independent process variables, the dependent variables, and the inferred dependent variables.

39. The method of claim 26, wherein the independent process variables include one or more of temperature, pressure, rate of addition or consumption of a catalyst or reagent, flow rate, the input or output of thermal energy to the process, refractive index, conductivity, viscosity, opacity, pH, density, intensity responses measured by photometric or spectrometric devices at particular wavelengths or frequencies, and combinations thereof.

40. The method of claim 26, further comprising aggregating the independent process variables into a property spectrum.

41. The method of claim 40, further comprising using the property spectrum to update the model.

42. The method of claim 26, wherein said generating a comparison comprises validating at least one of the independent process variables, the dependent variables, and the inferred dependent variables.

43. The method of claim 42, comprising adjusting the model and/or controlling the process stream once the dependent variables are validated.

44. The method of claim 43, comprising determining whether to use the validated dependent variables to adjust the model based upon at least one of an age of the validated dependent variables and a current state of the process as compared to a state of the process at the time the validated dependent variable was measured.

45. The method of claim 42, comprising generating the comparison using a statistical multivariate procedure.

46. The method of claim 42, comprising generating the comparison using a procedure selected from the group consisting of:

(a) a multivariate method to determine whether the sample spectrum belongs to a population of sample spectra which includes a calibration sample set;

(b) multivariate method to classify any spectral outlier according to cause;

(c) replacing any false independent variable within in the sample spectrum;

(d) determining whether any outlier condition is due to a particular sensor by comparing sensor output to a baseline;

(e) correcting the outlier condition by replacement with a model value;

(f) using averaging, regression, a damping function, or a vector-based algorithm, to provide a mathematical relationship between the inferred dependent variables and the dependent variables, which mathematical relationship is then applied to subsequent inferential predictions to produce a corrected inferential value;

(g) estimating a lag time of the dependent variables relative to the inferred dependent variables;

(h) applying a statistical analysis to the dependent variables and the inferred dependent variables;

(i) updating the model with the addition of recent data to and the elimination of older data from the data set used in creation of the model;

(j) updating the model by adding sample data for spectral outliers and/or prediction outliers to the data set used in creation of the model;

(k) capturing a sample that has been identified as a spectral outlier or a property outlier, and using it to adjust outputs of the inferential analyzer;

(l) performing model updates on the basis of schedule;

(m) performing model updates on the basis of need as determined by factors including frequency, severity of spectral outliers and/or prediction outliers, and/or changing of process operation or configuration;

(n) applying a goodness of fit test procedure; and

combinations thereof.

47. The method of claim 26, wherein the hard analyzer comprises a gas chromatograph configured to generate the dependent variables by capturing and evaluating a chromatogram of a sample of the process stream.

48. The method of claim 26, wherein the hard analyzer is at least one of a:

gas chromatograph (GC), a fast GC, an ultrafast GC, a Fourier Transform Infrared (FTIR) analyzer, a Fourier Transform Near-Infrared (FTNIR) analyzer, a nuclear magnetic resonance (NMR) analyzer, a Raman analyzer, a flash point analyzer, a distillation property analyzer, a cloud point analyzer, a freezing point analyzer, a cold filter plugging point analyzer, a pour point analyzer, and a combination thereof.

49. The method of claim 26, wherein the dependent variables include one or more of octane, percent total aromatics, percent benzene, percent olefins, distillation properties, Reid Vapor Pressure (RVP), cetane number, distillation properties, polynuclear aromatics, polycyclic aromatics, cloud point, pour point, cold filter plugging point, PIONA, PONA, distillation properties, and combinations thereof.

50. The method of claim 26, wherein said capturing and using is effected using an on-line hard analyzer in real time, and the comparison is used to adjust the model and/or to control the process stream in real time.