Generalized lensing angular similarity operator

• US 20040260515A1
• Filed: 07/16/2004
• Published: 12/23/2004
• Est. Priority Date: 03/09/2000
• Status: Abandoned Application

• Current Assignee: SmartSignal Corporation (Emerson Electric Company)
• Original Assignee: SmartSignal Corporation (Emerson Electric Company)
• Inventors: Wegerich, Stephan W., Wolosewicz, Andre, Pipke, R. Matthew
• Application Number:
  US10/892,808
• Publication Number:
  US 20040260515A1
• Time in Patent Office: Days
• US Class Current: 702/182
• CPC Class Codes: G05B 17/02 electric
• 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

In a machine for monitoring an instrumented process or for analyzing one or more signals, an empirical modeling module for modeling non-linearly and linearly correlated signal inputs using a non-linear angular similarity function with variable sensitivity across the range of a signal input. A different angle-based similarity function can be chosen for different inputs for improved sensitivity particular to the behavior of that input. Sections of interest within a range of a signal input can be lensed for particular sensitivity.

Claims

1. A tool for monitoring system operation, said tool comprising:

a data acquisition unit, said data acquisition unit receiving signals from sensors, said sensors being disposed upon a system being monitored and providing parametric snapshots of system operation;

a memory storing a training set, said training set containing a plurality of system vectors, each of said system vectors being representative of an expected operating state of said system being monitored;

a processor receiving snapshots from said data acquisition unit and comparing received snapshots with system vectors from said memory, said processor comprising a similarity engine, said similarity engine receiving said snapshots from said acquisition unit and selectively applying a lensing similarity function to said comparison; and

an output unit using said similarity engine to apply the lensing similarity function to the received vectors to generate a similarity vector, said processor providing results of said comparison to said output unit.

2. A tool as in claim 1, said processor further comprising an estimated state generator receiving said similarity vector from said similarity engine and training vectors from said memory and generating an estimated state therefrom, said estimated state being selectively provided to said output device.

3. A tool as in claim 2, said processor further comprising a deviation detection engine, said deviation detection engine receiving snapshots from said data acquisition unit and estimated states from said estimated state generator and determining deviation therefrom, said deviation detection engine selectively providing said determined deviation to said output device.

4. A tool as in claim 3 wherein said lensing similarity function defines a similarity domain, vectors belonging to said training set falling on said similarity domain, snapshots being expected to fall within said similarity domain, each said snapshot'"'"'s location within said similarity domain being a basis of said comparison by said processor.

5. A tool as in claim 4 wherein the lensing similarity function is representable as a line segment selected from the group consisting of a polynomial segment, an elliptical arc, a trigonometric segment and a circular arc, said line segment defining said similarity domain.

6. A tool as in claim 4 wherein said lensing similarity function comprises selecting a line segment from a non-planar surface, said line segment defining said similarity domain.

7. A tool as in claim 3 wherein the lensing similarity function comprises moving the comparison angle apex with respect to a similarity domain, vectors belonging to said training set falling on said similarity domain, snapshots being expected to fall within said similarity domain, each said snapshot'"'"'s location within said similarity domain being a basis of said comparison by said processor.

8. A tool as in claim 3 wherein the lensing similarity function comprises extending a comparison angle range beyond 90°

, rays from said comparison angle contacting outer limits of a similarity domain, vectors belonging to said training set falling on said similarity domain, snapshots being expected to fall within said similarity domain, each said snapshot'"'"'s location within said similarity domain being a basis of said comparison.