System for computer-aided measurement of quality and/or process data in a paper machine

• US 7,437,208 B2
• Filed: 09/16/2004
• Issued: 10/14/2008
• Est. Priority Date: 09/16/2003
• Status: Expired due to Fees

• Current Assignee: Voith Paper Patent GmbH
• Original Assignee: Voith Patent Gmbh
• Inventors: Typpo, Pekka M., Muench, Rudolf, Holik, Herbert, Wegmann, Florian, Hardt, Niels
• Primary Examiner(s): Von Buhr; M. N.
• Application Number:
  US10/942,598
• Publication Number:
  US 20050145357A1
• Time in Patent Office: 1,489 Days
• Field of Search: 700 28- 31, 700 47- 50, 700108-110, 700127-129, 702 81- 84, 162/198, 162/199, 162/201, 162/202, 162/263
• US Class Current: 700/128
• CPC Class Codes: G05B 19/41875 characterised by quality surveillance of production
• CPC Class Codes: G05B 2219/32187 Correlation between controlling parameters for influence on quality parameters
• 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 23/0297 Reconfiguration of monitoring system, e.g. use of virtual sensors; change monitoring method as a response to monitoring results
• CPC Class Codes: Y02P 90/02 Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Abstract

A system for the computer-aided measurement of quality and/or process data during the production and/or conversion of a material web, in particular paper or board web, by way of correlation with raw measured data. The raw measured data is present in the form of other quality and process data during the production or conversion process. Measurements of quality parameters in the laboratory may also be incorporated in the raw measured data. The raw measured data is combined to form data sets which in each case are determined simultaneously. Specific laboratory or quality measurements are selected as target data which, by way of at least one soft-sensor algorithm running in a computer-based operation and linking unit, can be calculated from the other data serving as input data, and each data set containing measured data, which relates approximately to the same monitored volume of web and raw material, in particular paper and paper raw material.

Claims

1. A method for the computer-aided measurement of at least one of quality and process data during one of production and conversion of a fiber material web, comprising the steps of:

correlating raw measured data with target data, said raw measured data being present in the form of other quality and process data gathered during one of the production and the conversion of the material web, said correlation including at least one of;

incorporating measurements of quality parameters in a laboratory with said raw measured data; and

combining portions of said raw measured data to form data sets, which are determined simultaneously;

running at least one soft-sensor algorithm in a computer-based operation and linking unit;

selecting specific laboratory or quality measurements to serve as target data, which by way of said soft-sensor algorithm is calculated from said other quality and process data serving as input data, each said data set containing measured data, which relates approximately to a monitored volume of the material web and raw material;

pre-filtering items and data; and

coordinating said items of data with one another taking into account an accuracy with which a simultaneity of said data sets are known.

2. The method of claim 1, wherein said raw measured data is associated with a timestamp.

3. The method of claim 1, wherein said soft-sensor algorithm is determined by way of a mathematical method including at least one of a Principal Component Analysis (PCA), a Partial Least Squares Recursive (PLSR), a Self Organizing Map (SOM), a fuzzy and neural network.

4. The method of claim 1, wherein said soft-sensor algorithm runs automatically as soon as there is new input data.

5. The method of claim 1, wherein said soft-sensor algorithm runs automatically with one of a fixed and predefinable cycle time.

6. The method of claim 1, wherein said soft-sensor algorithm runs in response to a manual request.

7. The method of claim 1, wherein said soft-sensor algorithm runs when at least one condition for execution is satisfied.

8. The method of claim 1, further comprising the step of outputting a result from said soft-sensor algorithm by way of at least one of a display and a printer unit.

9. The method of claim 2, wherein said timestamps are provided with an offset to compensate for transport dead times which said monitored volume runs through during one of the production and the conversion process.

10. The method of claim 1, further comprising the step of determining said items of data from a sliding average filter and a predefinable window size.

11. The method of claim 1, wherein said input data additionally includes data from other sections of one of the production and the conversion process as said target data.

12. The method of claim 1, further comprising a step of determining at least one soft-sensor by way of a design tool.

13. The method of claim 12, wherein said design tool extracts recorded input data and target data from a database having historic data.

14. The method of claim 13, further comprising a step of selecting from said historic data at least one of two groups of data sets including a first group and a second group, said first group containing a training data set used to calculate said soft-sensor algorithm, said second group containing a test data set having other data used to test a calculation of said soft-sensor algorithm.

15. The method of claim 14, wherein said first group containing training data is further subdivided into a plurality of groups of data sets used during a training process.

16. The method of claim 14, further comprising a step of testing said input data by way of said design tool to determine if some of said input data is not correlated with said target data and eliminating such data from said training data set.

17. The method of claim 14, further comprising a step of testing said input data by way of said design tool to determine which of said input data in said training data set is redundant, with continuing use made of only said input data which improves the correlation with said target data.

18. The method of claim 14, further comprising a step of testing said input data by way of said design tool to determine which of said input data in said training data set has a nonlinear relationship with said target data and adding such data to said training data set.

19. The method of claim 14, further comprising a step of determining a calculation formula of said soft-sensor algorithm by way of said design tool with said training data set.

20. The method of claim 14, further comprising a step of testing said soft-sensor algorithm by way of said design tool using a test data set.

21. The method of claim 14, wherein by way of said design tool a calculation formula of said soft-sensor algorithm is implemented in an associated computer, said algorithm configured such that in the event of a respective change in said input data then said target data is output such that said output behaves like a sensor.

22. The method of claim 12, wherein said soft-sensor algorithm is automatically adapted to slow changes in one of the production and conversion process by way of repeated automatic starting of said design tool allowing data from a specific time window of the past to continue to be used.

23. The method of claim 1, further comprising using at least one soft-sensor based on said at least one soft-sensor algorithm during one of production and conversion of the material web.

24. The method of claim 23, wherein at least one said soft-sensor is used to check at least one real sensor.

25. The method of claim 24, wherein at least one said soft-sensor determines at least one of a grammage and a thickness of the material web in order to check a scanner.

26. The method of claim 25, wherein said at least one soft-sensor is two soft-sensors used in a two-of-three rule to decide if there is a possible fault in one of said two soft-sensors and said scanner.

27. The method of claim 26, wherein said two soft-sensors include a first soft-sensor and a second soft-sensor, said first soft-sensor utilizes at least one of flow and consistency data from a region of a flow box, said second soft-sensor utilizing at least one of flow and consistency data from a thick stock feed.

28. The method of claim 23, further comprising the steps of:

predicting a grammage of the material web apart from a flow and consistency measurement; and

measuring one of ash and ash metering quantity.

29. The method of claim 23, further comprising the step of predicting a thickness of the material web using at least one process variable including stock composition, press settings and calender settings.

30. The method of claim 24, further comprising the step of activating an alarm system if a fault is found in checking at least one said real sensor.

31. The method of claim 24, further comprising the step of substituting at least one said soft-sensor for at least one said real sensor in the event of a fault.

32. The method of claim 31, wherein said substituting step is carried out automatically when said at least one said real sensor is identified as faulty by using a two-of-three rule.

33. The method of claim 24, further comprising the step of checking a first of said at least one soft-sensors and substituting a second of said at least one said soft-sensors and one of said real sensors in the event of a fault of said first of said at least one soft-sensors.

34. The method of claim 24, further comprising the step of grade-independent calibrating of at least one said real sensor using at least one said soft-sensor.

35. The method of claim 34, wherein said at least one real sensor has at least one internal measuring channel that is extended by incorporating further process and quality data in order to calculate an actual measured variable.

36. The method of claim 35, wherein said at least one real sensor includes a real moisture sensor having grammage and type of fiber data incorporated.

37. The method of claim 23, further comprising the step of providing a dewatering process wherein at least one said soft-sensor is used to measure a dryness of the material web, said at least one soft-sensor having raw measurement channels additionally incorporates further process parameters, said ray measurement channels including at least one of optical, electromagnetic and radioactive absorption, said further process parameters including at least one of stock consistency in a flow box, type of stock and settings dewatering parameters.

38. The method of claim 23, wherein said at least one soft-sensor is used instead of a respective real sensor as long as said real sensor is not ready for use.

39. The method of claim 23, wherein said at least one soft-sensor is configured to be used to measure the quality data which can usually only be measured by one of a laboratory and online using real sensors.

40. The method of claim 23, wherein said soft-sensor is used to measure a porosity, printability and strength parameters of the material web.

41. The method of claim 39, wherein the quality data that is expected for current measured variables is calculated, as a calculated value, and monitored aided by a calculation formula of said soft-sensor algorithm.

42. The method of claim 41, wherein if said calculated value lies outside of tolerance limits then one of a warning signal is generated and predefined measures are automatically taken.

43. The method of claim 42, wherein said predefined measures includes a change in process parameters to directly eliminate a cause of deviation.

44. The method of claim 43, wherein said predefined measures include a change in parameters from process steps that follow a subprocess responsible for said deviation.

45. The method of claim 39, further comprising the step of measuring quality data for each spool of the material web, said quality data including strengths, porosity and surface properties.

46. The method of claim 39, further comprising the step of measuring quality data which was not previously measured for each spool of the material web, said quality data including printability values being at least one of mottling and missing dots on a portion of the material web using one of a laboratory and industrial printing process.

47. The method of claim 23, further comprising the step of automatically regulating quality parameters using said at least one soft-sensor, said quality parameters including porosity, printability and strength parameters.

48. The method of claim 23, wherein said at least one soft-sensor is used for determining process characteristic curves.

49. The method of claim 23, further comprising the step of predicting an anticipated effect on changes of one of the production and conversion process before said changes are made to process variables, wherein said at least one soft-sensor is used in the context of an if-then analysis to effect said predicting step.

50. The method of claim 49, wherein said predicting step is used for at least one process and quality control system.

51. The method of claim 23, further comprising the step of determining one of optimal machine settings and optimal process parameters using said at least one soft-sensor together with at least one mathematical optimization method.

52. The method of claim 51, wherein at least one said soft-sensor is determined by way of a design tool, said input data and said target data recorded by way of said design tool are extracted from a database having historic data, at least two groups of data sets being selected from said historic data including a first group and a second group, said first group including a training data set used to calculate said soft-sensor algorithm, said second group including a test data set having other data used to test the calculated soft-sensor algorithm.

53. The method of claim 51, wherein one of said machine settings and said process parameters have specific quality properties which are optimized by way of at least one mathematical optimization method.

54. The method of claim 51, further comprising the step of carrying out at least one confirmation trial for a verification purpose.

55. The method of claim 51, wherein one of the production and the conversion process is changed to one of said optimal machine settings and said optimal process parameters.