System for surveillance of spectral signals

• US 6,240,372 B1
• Filed: 11/14/1997
• Issued: 05/29/2001
• Est. Priority Date: 11/14/1997
• Status: Expired due to Term

• Current Assignee: ARCH Development Corporation (University of Chicago)
• Original Assignee: ARCH Development Corporation (University of Chicago)
• Inventors: Wilks, Alan D., Criss-Puszkiewicz, Cynthia, Wegerich, Stephan W., Gross, Kenneth C.
• Primary Examiner(s): WACHSMAN, HAL D
• Application Number:
  US08/970,873
• Time in Patent Office: 1,292 Days
• Field of Search: 702/71, 702/183-185, 702/113, 702/116, 702/75-77, 702/124, 702/126, 702/189, 702/190, 702/198, 702/60-62, 702/155, 702/179-182, 702/FOR 103 , 702/FOR 104, 702/FOR 105, 702/FOR 107, 702/FOR 108, 702/FOR 133, 702/FOR 111-FOR 119, 702/FOR 134, 702/FOR 135 , 702/FOR 139, 702/FOR 145, 702/FOR 168, 702/FOR 170, 702/FOR 171, 324/76.19, 324/76.21, 324/76.22, 376/215-217, 340/506, 340/514, 340/516, 340/517, 340/524, 340/525, 340/540, 340/286.01, 340/825.36, 700/51, 700/79, 700/80, 700/83, 700/286-290
• US Class Current: 702/71
• CPC Class Codes: G01R 23/16 Spectrum analysis; Fourier analysis

Abstract

A method and system for monitoring at least one of a system, a process and a data source. A method and system have been developed for carrying out surveillance, testing and modification of an ongoing process or other source of data, such as a spectroscopic examination. A signal from the system under surveillance is collected and compared with a reference signal, a frequency domain transformation carried out for the system signal and reference signal, a frequency domain difference function established. The process is then repeated until a full range of data is accumulated over the time domain and a SPRT sequential probability ratio test methodology applied to determine a three-dimensional surface plot characteristic of the operating state of the system under surveillance.

Claims

1. A method of testing a process and/or a data set for determining a particular condition thereof, comprising the steps of:

collecting data from at least a first source of data and a second source of data to redundantly detect at least one variable of the process and/or the data set to provide a first signal from said at least a first source of data and a second signal from said at least a second source of data, each of said signals being characteristic of the at least one variable;

providing frequency domain transformation data for each of said first and second signals at a given time;

obtaining a frequency domain difference function characteristic of the difference pairwise for the given time between said frequency domain transformed first signal and said frequency domain transformed second signal;

operating on the frequency domain difference function using computer means for applying a statistical analysis technique to determine a combined statistical frequency spectrum data set for the process and/or the data set;

repeating the steps for a new incremental time position of providing frequency domain transformation data, obtaining a frequency domain difference function and operating on the frequency domain difference function by applying said statistical analysis technique until generation of a three-dimensional surface characteristic of a combined frequency and time domain function for a particular condition of the process and/or the data set; and

establishing an alarm condition by observing a characteristic interconnected point geometry along a plane intersecting points of said three-dimensional surface.

2. The method as defined in claim 1 wherein said at least a first source of data comprises a signal from a system being monitored and said at least a second source of data comprises a reference signal characteristic of a desired operational state.

3. The method as defined in claim 1 wherein said statistical analysis technique comprises a pattern recognition methodology.

4. The method as defined in claim 1 wherein said statistical analysis technique comprises a sequential probability ratio test technique.

5. The method as defined in claim 1 wherein the process comprises one or more of a chemical process, a mechanical process, an electrical operational process, a biological process, a manufacturing process and a financial process.

6. The method as defined in claim 5 further including the step of generating an alarm upon detecting an alarm condition upon deviating from a desired condition of the process.

7. The method as defined in claim 6 wherein the process of collecting data from said first source of data is comprised of detecting light spectra having interacted with a product being manufactured and rejecting the product upon detecting the alarm condition.

8. The method as defined in claim 1 further including at least one of the steps of modifying the process or changing the data set responsive to sensing an alarm condition.

9. The method as defined in claim 8 wherein the step of modifying the process comprises ceasing operation of the process.

10. A method of operating a system to test a process and/or a data set for determining a state of the system, comprising the steps of:

monitoring at least one source of data of the system to detect at least one variable of the process to provide a real signal from said at least one source of data;

generating a standard signal which is characteristic of the at least one variable;

generating frequency domain transformation data for said real signal and said standard signal for a given time;

obtaining a frequency difference function characteristic of the difference pairwise at the given time between said frequency domain transformed real signal and said frequency domain transformed standard signal;

operating on the frequency domain difference function using a computer for applying sequential probability ratio test analysis technique to determine a sequential probability ratio test frequency spectrum data set for the process and/or the at least one source of data;

repeating the steps for a new incremental time position of generating frequency domain transformation data, obtaining a frequency domain difference function and operating on the frequency domain difference function by applying said sequential probability ratio test analysis technique until generation of a three-dimensional surface, characteristic of a combined frequency and time domain function for a particular condition of the process and/or the at least one source of data; and

establishing an alarm condition by observing a characteristic interconnected point geometry along a plane intersecting points of said three-dimensional surface.

11. The method as defined in claim 10 further including the step of providing a signal to said system allowing modification of the process and/or the at least one source of data when an alarm condition is detected.

12. The method as defined in claim 11 wherein the step of generating a frequency domain transformation comprises performing a Fourier transformation.

13. The method as defined in claim 10 further including the step of generating an alarm upon detecting an alarm condition upon deviating from a desired condition of the system.

14. The method as defined in claim 13 wherein the step of collecting said at least one source of data is comprised of detecting a spectrum of electromagnetic radiation having undergone interaction with a product being manufactured by the system and rejecting the product upon detecting the alarm condition.

15. A method for testing a process and/or a data set of a system for determining the condition thereof, comprising:

collecting a first signal from at least a first source of data over a monitoring time period from at least one variable of the process and/or the data set;

collecting a second signal from at least a second source of data from the process and/or the data set over the monitoring time period for comparison with said first signal from said at least a first source of data;

providing a frequency domain transformation at a given time for said signals from said first source of data and said at least a second source of data;

obtaining a frequency domain difference function characteristic of the difference pairwise at the given time between said frequency domain transformed signals from said at least a first source of data and said at least a second source of data for the given time in the monitoring time period;

continuing said frequency domain transformation for a time increment for a next time in the monitoring time period to provide a three-dimensional surface of power spectrum data over time and frequency space;

applying a pattern recognition methodology to determine a three-dimensional spectrum data set characteristic of an operating condition of the at least a first data source and at least a second data source; and

establishing an alarm condition by observing a characteristic interconnected point geometry along a plane intersecting points of said three-dimensional surface.

16. The method as defined in claim 15 wherein said first signal is derived from said process and/or said data set and said second signal is derived from a characteristic model system.

17. The method as defined in claim 16 wherein said data set comprises a spectroscopy system data set.

18. The method as defined in claim 17 wherein said spectroscopy system data set is selected from the group consisting of data from infrared spectrometry, near infrared spectrometry, mass spectrometry, electron paramagnetic spectrometry, nuclear magnetic resonance spectrometry, raman spectrometry, electron spin resonance spectrometry, fluorescence spectrometry, thermogram spectrometry, ultraviolet spectrometry, chromatographic spectrometry, gamma spectrometry, absorption spectrometry, electronic emission spectrometry, bioluminescence spectrometry, acoustic resonance spectrometry, chlorophyll fluorescence spectrometry.

19. The method as defined in claim 15 wherein said pattern recognition methodology comprises a sequential probability ratio test technique.

20. The method as defined in claim 15 wherein said pattern recognition methodology comprises a statistical analysis technique and said frequency domain transformation comprises a Fourier transformation.