Self tuning system for industrial surveillance

• US 6,131,076 A
• Filed: 07/25/1997
• Issued: 10/10/2000
• Est. Priority Date: 07/25/1997
• Status: Expired due to Term

• Current Assignee: ARCH Development Corporation (University of Chicago)
• Original Assignee: ARCH Development Corporation (University of Chicago)
• Inventors: Jarman, Kristin K., Stephan, Wegerich W, Gross, Kenneth C.
• Primary Examiner(s): SHAH, KAMINI S
• Application Number:
  US08/900,545
• Time in Patent Office: 1,173 Days
• Field of Search: 702/89, 702/189, 702/190, 702/191, 364/149, 364/148, 340/164, 340/506, 340/651
• US Class Current: 702/189
• 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 23/0254 based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks

Abstract

A method and system for automatically establishing operational parameters of a statistical surveillance system. The method and system performs a frequency domain transition on time dependent data, a first Fourier composite is formed, serial correlation is removed, a series of Gaussian whiteness tests are performed along with an autocorrelation test, Fourier coefficients are stored and a second Fourier composite is formed. Pseudorandom noise is added, a Monte Carlo simulation is performed to establish SPRT missed alarm probabilities and tested with a synthesized signal. A false alarm test is then emperically evaluated and if less than a desired target value, then SPRT probabilities are used for performing surveillance.

Claims

1. A method of automatically establishing operational parameters of a surveillance system and monitoring a process, comprising the steps of:

(a) using an automated system for acquiring time dependent data characteristic of a state of interest;

(b) using an automated system for performing domain transformation on said time dependent data;

(c) using an automated system for constructing a Fourier composite using identified Fourier modes, by iteratively finding the optimal number of modes that are furthest away from the median PSD;

(d) using an automated system for filtering serial correlation from time dependent data;

(e) using an automated system for performing at least one Gaussian test, at least one whiteness noise test and an autocorrelation test and repeating steps (c) and (d) until these tests are satisfied;

(f) using an automated system for storing Fourier coefficients;

(g) using an automated system for generating a Fourier composite from said Fourier coefficients;

(h) using an automated system for adding pseudorandom noise to said Fourier composite;

(i) using an automated system for generating observations of a synthesized signal for Monte Carlo simulation;

(j) using an automated system for establishing SPRT missed alarm probabilities;

(k) using an automated system for testing said SPRT probabilities with a synthesized signal;

(l) using an automated system for evaluating empirically a false alarm rate and if less than desired target values, outputting said SPRT probabilities for use in performing surveillance by said system;

(m) if said false alarm rate is greater than said target values, reducing said SPRT probabilities and repeating steps (j) and (k) until obtaining said false alarm rate less than said desired target value; and

(n) upon establishing operational parameters in said steps (a)-(m) for the surveillance system, carrying out an automated sequential probability ratio test to monitor a process for alarm conditions.

2. The method as defined in claim 1 further including the step of using an automated system for forming a Fourier composite without pseudorandom noise to produce a real time residual function.

3. The method as defined in claim 2 wherein a SPRT methodology operates on said filtered data without serial correlation.

4. The method as defined in claim 3 wherein said SPRT methodology outputs one of normal SPRT output, degraded SPRT output and a no decision SPRT output.

5. The method as defined in claim 4 wherein the no decision SPRT output is transmitted for additional sampling.

6. The method as defined in claim 4 further including operating on one of said outputs using a real time surveillance expert system.

7. The method as defined in claim 6 wherein said real time surveillance expert system compares output from parallel deployed SPRT modules.

8. A system for automatically establishing operational parameters of a surveillance system and monitoring a process using the operational parameters, comprising:

(a) an automated system for acquiring time dependent data characteristic of a state of interest for the process;

(b) an automated system for performing domain transformation on said time dependent data;

(c) an automated system for constructing a Fourier composite using identified Fourier modes, by iteratively finding the optimal number of modes that are furthest away from the median PSD;

(d) an automated system for filtering serial correlation from time dependent data;

(e) an automated system for performing at least one Gaussian test, at least one whiteness noise test and an autocorrelation test and repeating steps (c) and (d) until these tests are satisfied;

(f) an automated system for storing Fourier coefficients;

(g) an automated system for generating a Fourier composite from said Fourier coefficients;

(h) an automated system for adding pseudorandom noise to said Fourier composite;

(i) an automated system for generating observations of synthesized signal for Monte Carlo simulation;

(j) an automated system for establishing SPRT missed alarm probabilities;

(k) an automated system for testing said SPRT probabilities with a synthesized signal;

(l) an automated system for evaluating empirically a false alarm rate and if less than desired target values, outputting said SPRT probabilities for use in performing surveillance by said system and if said false alarm rate is greater than said target values, reducing said SPRT probabilities and using the automated system of (j) and (k) until obtaining said false alarm rate less than said desired target value; and

(m) an automated system for carrying out a SPRT procedure using the operational parameters for the surveillance system to monitor a process for alarm conditions.