METHOD FOR EARLY WARNING CHATTER DETECTION AND ASSET PROTECTION MANAGEMENT

• US 20160340830A1
• Filed: 08/01/2016
• Published: 11/24/2016
• Est. Priority Date: 10/20/2011
• Status: Active Grant

• Current Assignee: Ecolab USA Incorporated (Ecolab, Inc.)
• Original Assignee: Nalco Company (Ecolab, Inc.)
• Inventors: Archer, Sammy L., Von Drasek, William A., Furman, Gary S. Jr.
• Granted Patent: US 10,604,896 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: D21G 3/00 Doctors
• CPC Class Codes: D21G 3/04 for drying cylinders
• CPC Class Codes: D21G 9/0036 controlling the press or drying section
• CPC Class Codes: D21G 9/0045 controlling the calendering or finishing
• CPC Class Codes: G01N 29/14 using acoustic emission techniques echo of particles G01N29/046; measuring mechanical vibrations or acoustic waves in solids in general G01H1/00
• CPC Class Codes: G01N 29/4418 with a model, e.g. best-fit, regression analysis
• CPC Class Codes: G01N 29/4481 Neural networks
• CPC Class Codes: G01N 29/46 by spectral analysis, e.g. Fourier analysis or wavelet analysis spectral signal processing per se G06F17/14
• CPC Class Codes: G06N 3/02 Neural networks
• CPC Class Codes: G06N 7/08 using chaos models or non-linear system models

Abstract

The invention embodies the application of different combinations of the monitoring and data processing aspects as a means to develop an early warning chatter alarming system. Configuring an early warning chatter alarming system can be as simple as using nσ alarm settings to develop an alarming strategy from different trend conditions such as overall RMS, selected vibration frequencies, slope analysis, and wavelet analysis. A higher level of alarming is provided by using a time integrated approach to account for both intensity of the alarm variable and duration. Combining these different aspects with a predictive model incorporates process-operating conditions to enhance the alarming sensitivity for earlier detection and reduce false positives. Finally, combining the different alarming aspects with a rule-based decision making approach such as fuzzy logic allows alarming based on qualitative analysis of different data streams.

Claims

1. A method of detecting and addressing chatter from Yankee dryer doctor blades used in the creping process, cleaning, or cut-off operations, the method comprising the steps of over a period of time, with an sensor constructed and arranged to measure the frequencies and amplitudes of vibrations in a doctor blade as it crepes a paper product, measuring the frequencies and amplitudes of the vibrations indexed by time, collecting the measurements into a time waveform, converting the waveform into a fast-Fourier transform having a frequency spectrum which includes distinct vibration bands, correlating characteristics of the vibration bands with acceptable performance properties of the doctor blade and to define a baseline of acceptable vibration bands, predicting from the correlated characteristics the degree of deviation from the baseline of acceptable vibration bands associated with doctor blade chatter, and outputting when a data point on a vibration band exceeds the degree of deviation excessive chatter has occurred.

2. The method of claim 1 where the sensor is an accelerometer.

3. The method of claim 1 where the sensor is a piezoelectric accelerometer.

4. The method of claim 1 where the measurements are analyzed and modeled by a data processing device constructed and arranged to utilize one process selected from the group consisting of:

RMS data trending, neural network techniques, multiple regression analysis, AR, ARMAX, partial least squares, and any combination thereof.

5. The method of claim 1 where at least one of the correlations is determined by comparing characteristics of the vibration bands with blade age.

6. The method of claim 5 where the measurements are analyzed and modeled by a data processing device constructed and arranged to utilize RMS data trending and where the determination is made at least in part by noting that the slope in a saw tooth shaped vibration band continuously increases over time with the same blade and becomes discontinuous when the blade is changed.

7. The method of claim 6 further comprising the step of defining a deviation from the baseline due to chatter to only occur when a deviation exceeds the mean and standard deviation of the baseline due to both an increase in magnitude and a duration of that increase greater than the mean duration of all data spikes in the waveform.

8. The method of claim 6 further comprising the steps of pre-determining the slope at which the blade is too old to be desired for use and replacing the blade when such a slope manifests on the waveform.

9. The method of claim 1 where at least one of the correlations is determined by comparing characteristics of the vibration bands with one factor selected from:

track bearing, balance, dryer lubricity, dust levels, moisture levels, temperature, felt age, grade, furnish composition, coating chemistry, cleaning blade status (on or off), machine speed, external source vibrations, external pressure sources, and any combination thereof.

10. The method of claim 1 where the range of characteristics of the vibration bands caused by the factor is so broad that the sensor must be capable of detecting frequency bandwidth spanning four orders of magnitude.

11. The method of claim 1 where the sensor only indirectly measures vibrations of the doctor blade because it is engaged not to the blade itself but to a blade holder which is engaged to and provides more rigid support to the blade but which does not dampen the vibration to such an extent that an accurate measurement cannot be taken.

12. The method of claim 1 where the measurements are taken synchronously.

13. The method of claim 1 where the measurements are taken asynchronously.

14. The method of claim 1 where the output is an alarm.