Method for predictive determination of a process variable based on an assignment of a discrete measured value

• US 7,941,393 B2
• Filed: 04/27/2009
• Issued: 05/10/2011
• Est. Priority Date: 10/28/2006
• Status: Expired due to Fees

• Current Assignee: ABB Power Grids Switzerland AG (ABB Limited)
• Original Assignee: ABB Technology AG (ABB Limited)
• Inventors: Glock, Jürgen, Heinemann, Lothar
• Primary Examiner(s): Holmes; Michael B.
• Assistant Examiner(s): COUGHLAN, PETER D
• Application Number:
  US12/430,307
• Publication Number:
  US 20090248606A1
• Time in Patent Office: 743 Days
• Field of Search: 706/62
• US Class Current: 706/62
• CPC Class Codes: G05B 13/026 using a predictor
• CPC Class Codes: G05B 17/02 electric
• CPC Class Codes: G05B 19/408 characterised by data handling or data format, e.g. reading, buffering or conversion of data
• CPC Class Codes: G05B 2219/31282 Data acquisition, BDE MDE
• CPC Class Codes: G05B 2219/37542 Curve fitting measured points, predict, extrapolate dimension in time
• 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 23/0283 Predictive maintenance, e.g. involving the monitoring of a system and, based on the monitoring results, taking decisions on the maintenance schedule of the monitored system; Estimating remaining useful life [RUL] preventive maintenance, i.e. planning maintenance according to the available resources without monitoring the system G06Q10/06

Abstract

A system and method are disclosed for predictive determination of a process variable (P) in a technical installation. Measured values for the process variable (P) are recorded at predeterminable points in time. Temporally successive recorded measured values are combined to form a weighted measured value. A discrete measured value is assigned to each weighted measured value. Each weighted measured value that lies in a predeterminable tolerance band is assigned a same discrete measured value. A future temporal progression of the process variable (P) is estimated using a linear regression of the discrete measured values, and a point in time in which the process variable (P) reaches a predeterminable limit value is calculated.

Claims

1. A method for predictive determination of a process variable (P) in a technical installation using a circuit having a database and a processor, the method comprising:

receiving measured values from a sensor;

recording, in the database, the measured values for the process variable (P) of the technical installation at predeterminable points in time during operation of the technical installation; and

 

in the processor;

combining temporally successive recorded measured values to form a weighted measured value;

assigning a discrete measured value to each weighted measured value;

estimating a future time profile of the process variable (P) using a regression of the discrete measured values; and

calculating a point in time at which the process variable (P) reaches a predeterminable limit value,

wherein each weighted measured value that lies in a predeterminable tolerance band (H1 ... H8) is assigned a same discrete measured value.

2. The method as claimed in claim 1, wherein the regression is a linear regression.

3. The method as claimed in claim 2, comprising:

calculating a coefficient of determination (B) of the linear regression; and

calculating, with aid of the coefficient of determination (B) , a time interval within which the process variable (P) is expected to reach a predetermined value.

4. The method as claimed in claim 2, comprising:

detecting damage to the technical installation using the linear regression.

5. The method as claimed in claim 1, wherein the predeterminable tolerance bands (H1 ... H8) have different widths.

6. The method as claimed in claim 1, comprising:

taking the measured values into account more or less strongly depending on a point in time of their recording.

7. The method as claimed in claim 6, wherein the measured values taken into account are only measured values lying temporally after a predeterminable point in time.

8. The method as claimed in claim 1, comprising:

applying the calculating of a point in time at which the process variable (P) reaches the predeterminable limit value to a gas-insulated switching installation as the technical installation.

9. The method as claimed in claim 8, comprising:

receiving pressure of an insulating gas in a gas space of the gas-insulated switching installation as the process variable (P).

10. The method as claimed in claim 9, wherein the regression is a linear regression.

11. The method as claimed in claim 1, comprising:

calculating a coefficient of determination (B) of the linear regression; and

calculating, with aid of the coefficient of determination (B), a time interval within which the process variable is expected to reach a predetermined value.

12. The method as claimed in claim 11, comprising:

taking the measured values into account more or less strongly depending on a point in time of their recording.

13. The method as claimed in claim 12, comprising:

detecting damage to the technical installation using the linear regression.

14. A system for predictive determination of a process variable (P) in a technical installation, comprising:

a sensor;

a database for recording measured values for the process variable (P) of the technical installation at predeterminable points in time during operation of the technical installation; and

a processor for executing a program for;

combining temporally successive recorded measured values to form a weighted measured value;

assigning a discrete measured value to each weighted measured value;

estimating a future time profile of the process variable (P) using a regression of the discrete measured values; and

calculating a point in time at which the process variable (P) reaches a predeterminable limit value,

wherein each weighted measured value that lies in a predeterminable tolerance band (H1 ... H8) is assigned a same discrete measured value.