Neuro-parity pattern recognition system and method

• US 6,119,111 A
• Filed: 06/09/1998
• Issued: 09/12/2000
• Est. Priority Date: 06/09/1998
• Status: Expired due to Term

• Current Assignee: ARCH Development Corporation (University of Chicago)
• Original Assignee: ARCH Development Corporation (University of Chicago)
• Inventors: Yue, Yong, Van Alstine, Rollin G., Wegerich, Stephan W., Gross, Kenneth C., Singer, Ralph M.
• Primary Examiner(s): Hafiz, Tariq R.
• Assistant Examiner(s): PENDER JR, MICHAEL J
• Application Number:
  US09/094,413
• Time in Patent Office: 826 Days
• Field of Search: 706/15, 706/26
• US Class Current: 706/15
• CPC Class Codes: G21D 3/04 Safety arrangements emergency protection of reactor G21C9/00
• CPC Class Codes: Y02E 30/00 Energy generation of nuclear origin
• CPC Class Codes: Y02E 30/30 Nuclear fission reactors

Abstract

A method and system for monitoring a process and determining its condition. Initial data is sensed, a first set of virtual data is produced by applying a system state analyzation to the initial data, a second set of virtual data is produced by applying a neural network analyzation to the initial data and a parity space analyzation is applied to the first and second set of virtual data and also to the initial data to provide a parity space decision about the condition of the process. A logic test can further be applied to produce a further system decision about the state of the process.

Claims

1. A method of monitoring a process and a data source and determining their condition, comprising the steps of:

collecting initial data from a data source characteristic of a process;

producing a first set of virtual data including the step of applying a system state analyzation to the initial data from the data source;

producing a second set of virtual data including the step of applying a neural network analyzation to the initial data from the data source;

analyzing the first and second set of virtual data and also the initial data to provide a decision about at least one of the condition of the process and condition of the data source; and

using a computer to perform a parity space analyzation of the first and second set of virtual data and also of the initial data to provide a parity space decision information set about at least one of the condition of the process being monitored and the condition of the data source, thereby determining whether at least one of the process and the data source is deviating from a particular state.

2. The method as defined n claim 1 further including the step of performing a sequential probability ratio test on the first set of virtual data to produce system state sequential probability ratio test decision information.

3. The method as defined in claim 1 further including the step of performing a logic test on the first set of virtual data and the second set of virtual data to produce a system state decision.

4. The method as defined in claim 1 wherein the step of producing a first set of virtual data comprises performing a vector similarity evaluation analyzation.

5. The method as defined in claim 4 wherein the step of producing the first set of virtual data further includes the step of performing a sequential probability ratio test.

6. The method s defined in claim 5 further including the step of performing a logic test on the parity space decision information set to produce a system state decision about status of the process.

7. The method as defined in claim 1 wherein the process is selected from the group consisting of an industrial, chemical, nuclear, biological, vehicular, financial, avionics and nautical process.

8. A method of monitoring a process and determining its condition, comprising the steps of:

collecting initial data characteristic of a process;

producing a first set of virtual data including the step of applying a system state analyzation to the initial data;

using a computer to produce a second set of virtual data including the step of applying a neural network analyzation to the initial data; and

using a computer to apply a parity space analyzation of the first and second set of virtual data and also to the initial data to provide a parity space decision about the condition of the process thereby determining whether the process is deviating from a particular state.

9. The method as defined in claim 8 further including the step of performing a sequential probability ratio test on the first and second set of virtual data to produce system state, sequential probability ratio test decision information about the status of the process.

10. The method as defined in claim 8 further including the step of performing a logic test on the first and second set of virtual data to produce a system state decision.

11. The method as define din claim 8 wherein the step of producing a first set of virtual data comprises performing a vector similarity evaluation analyzation.

12. The method as defined in claim 8 further including the step of performing a sequential probability ratio test on the first and second set of virtual data.

13. The method as defined in claim 12 further including the step of performing a parity space analyzation of the first and second set of virtual data and also of the initial data to produce a parity space decision information set about the condition of the process being monitored.

14. The method as defined in claim 8 wherein the process is selected from the group consisting of an industrial, chemical, nuclear, biological, vehicular, financial, avionics ad nautical process.

15. A method of monitoring a process, comprising the steps of:

collecting initial data characteristic of a process;

using a computer to apply a system state analyzation to the initial data to produce system state estimated values which are compared to the initial data to provide system state residual data;

using a computer to apply a sequential probability ratio test to the system state residual data to produce system state sequential probability ratio test decision information;

using a computer to apply a parity space analyzation to the system state estimated values to provide system state, parity space decision information;

applying a neural network analyzation to the initial data to produce neural network estimated values which are compared to the initial data to provide neural network residual data;

using a computer to apply a sequential probability ratio test to the neural network residual data to produce neural network, sequential probability ratio test decision information;

using a computer to apply a parity space analyzation to the initial data to provide initial state, parity space decision information; and

using a computer to perform a logic test on the system state parity space decision information, the neural network parity space decision information and the initial state parity space decision information to produce a system state decision about the condition of the process, thereby determining whether the process is deviating from a particular state.

16. The method as defined in claim 15 wherein the step of applying a system state analyzation comprises performing a vector similarity evaluation analyzation.

17. The method as defined in claim 16 wherein the process is selected form the group consisting of an industrial, chemical, nuclear, biological, vehicular, financial, avionics and nautical process.