Condition monitoring in technical processes

• US 20050171705A1
• Filed: 11/15/2004
• Published: 08/04/2005
• Est. Priority Date: 11/18/2003
• Status: Active Grant

• Current Assignee: Peter Renner
• Original Assignee: Peter Renner
• Inventors: Renner, Peter
• Granted Patent: US 7,113,871 B2
• Time in Patent Office: Days
• US Class Current: 702/33
• CPC Class Codes: G05B 2219/33027 Artificial neural network controller
• CPC Class Codes: G05B 2219/42271 Monitor parameters, conditions servo for maintenance, lubrication, repair purposes
• 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

The invention makes a contribution in order that machines and installations are better utilized, their service life is extended, reliability is improved, the loss of valuable materials is avoided, test results are optimized, and energy consumption is diminished. It is an expert system that acquires and assesses the condition of installations, machines and apparatuses. The mechanism proceeds from a normal condition and permanently observes short-term, medium-term and long-term changes. If variations in the normal condition are detected promptly, major instances of damage can be avoided by promptly performing corrective actions before major instances of damage associated with prolonged downtimes occur.

Claims

1. A system for monitoring technical processes, machines, installations, apparatuses or the like, in which measured data are acquired via measurement channels, which contain sensors, sensor lines and evaluation electronics, and with governing software, wherein the measurement channels of mutually dependent measured variables are combined into reference groups (module 26).

2. The monitoring system of claim 1 wherein the measurement channels are split into the category of identification channels for defining a quasi-static phase of operation and the category of condition channels (module 27), whose measured data, after computational and comparison procedures, supply information about the condition of the underlying process.

3. The monitoring system of one of claims 1 or 2 wherein a quasi-static phase of operation is defined through identification conditions (module 28), the quasi-static phase of operation being in effect if the identification conditions are fulfilled.

4. The monitoring system of one of the foregoing claims wherein gaps arising as a result of downtimes and times of the non-quasi-static phase of operation are filled in with dummy values (module 29).

5. The monitoring system of one of the foregoing Claims wherein, during the running time of a process, tests are continually performed to ascertain whether the identification conditions are fulfilled (module 28).

6. The monitoring system of one of the foregoing claims wherein analog measured data, binary data, as well as machine running times are linked to form true/false conditions in order to identify the quasi-static phase of operation.

7. The monitoring system of one of the foregoing claims wherein the machine condition is monitored only during the quasi-static phase of operation.

8. The monitoring system of one of the foregoing claims wherein the measured values of the condition channels of the reference groups are compressed into statistical values, preferably means (module 30).

9. The monitoring system of one of the foregoing claims wherein the statistical condition values are time-stamped.

10. The monitoring system of one of the foregoing claims wherein the statistical condition values of the individual time series are aggregated into test blocks (module 31) that contain a fixed number of values.

11. The monitoring system of one of the foregoing claims wherein the test blocks are set up in moving fashion (module 31) so that the oldest value is dropped and the newest is incorporated.

12. The monitoring system of one of the foregoing claims wherein the swing is compared with a limiting value (module 32) and, in the case of exceedance, the swing is compared with the mean of the swings of the other test blocks of the same time series of the reference group and, in the case of exceedance of a further limiting value, the shape of the curve is tested.

13. The monitoring system of one of the foregoing claims wherein, if a reference group comprises just one measurement channel, the swing is compared with a limiting value (module 32).

14. The monitoring system of one of the foregoing claims wherein a further comparison with the swing of the test block of the next higher or still higher time series is performed (module 32) and, in the case of exceedance of a limiting value, the shape of the curve is tested if the limiting value is exceeded.

15. The monitoring system of one of the foregoing claims wherein the first and second derivatives are employed for defining the shape of the curve (module 33), limiting values of the derivatives determining whether a linear or a progressive behavior is present.

16. The monitoring system of one of the foregoing claims wherein the shape of the curve is investigated to determine whether it is sloping linearly upward or downward and progressively upward or downward.

17. The monitoring system of one of the foregoing claims wherein the exceedance of limiting values, depending on the severity of the malfunctions established, entries are initiated in observation lists (modules 34, 36), warning lists (modules 35, 37), or alarms.

18. The monitoring system of one of the foregoing claims wherein linear changes are rated as less than progressive changes.

19. The monitoring system of one of the foregoing claims wherein, in normal operation free of malfunctions, there is a learning phase, permanent data being acquired during this learning phase and the quasi-static phase of operation, by which the limits of normal operation are revealed and the limiting values can be automatically positioned in each case at the maximal and minimal values with allowance for a margin of safety.