Diagnostic expert system

• US 5,107,500 A
• Filed: 04/19/1989
• Issued: 04/21/1992
• Est. Priority Date: 04/20/1988
• Status: Expired due to Term

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Kim, Moo W., Fukazawa, Mitsunori, Wakamoto, Masaaki, Aramaki, Takahiro
• Primary Examiner(s): Smith, Jerry
• Assistant Examiner(s): Cass, Russell
• Application Number:
  US07/340,243
• Time in Patent Office: 1,098 Days
• Field of Search: 364/200 MS File, 371/15.1, 395/50, 395/916, 395/922
• US Class Current: 714/26
• CPC Class Codes: G06F 11/2257 using expert systems
• CPC Class Codes: Y10S 706/916 Electronic or computer, internal or network, circuit

Abstract

A diagnostic expert system is set up with a first diagnostic subsystem based on a maintenance technician'"'"'s knowledge and a second diagnostic subsystem based on a design engineer'"'"'s knowledge. If a fault cannot be located under the first diagnostic subsystem, the second diagnostic subsystem, which operates by using the result obtained from the first diagnostic subsystem is started so as to determine the location of the fault.

Claims

1. A diagnostic expert system for diagnosing a data processing apparatus having a control unit, memory unit, and input/output unit, according to a fault message issued from the apparatus due to a fault occurring therein, comprising:

diagnostic memory means for storing therein both diagnostic data and structure data;

diagnostic control means for maintaining the diagnostic data;

first diagnostic means for performing automated reasoning to determine a questionable object, based on both said fault message and said diagnostic data and for diagnosing the questionable object; and

second diagnostic means for hypothesizing the questionable object based on the structure data regarding the data processing apparatus, a related faulty object being specified by said first diagnostic means when possible and, if the related faulty object cannot be specified by said first diagnostic means, said second diagnostic means specifies the related faulty object,

wherein in said second diagnostic means, the questionable object is identified, and then data of a processing procedure corresponding to the questionable object is selected, and the processing procedure is executed from last to first in a reverse direction to search for the related faulty object.

2. A diagnostic expert system as set forth in claim 1, wherein the diagnostic data is defined by each said fault message in correspondence with a related questionable object.

3. A diagnostic expert system as set forth in claim 1, wherein the structure data define a connection relationship between the control, memory and input/output units comprising the data processing apparatus in accordance with design information for constructing the data processing apparatus.

4. A diagnostic expert system as set forth in claim 3, wherein the design information includes definition of each combination of both a connection relationship between the control, memory and input/output units comprising the data processing apparatus and a processing procedure between the control, memory and input/output units for every processing procedure.

5. A diagnostic expert system as set forth in claim 1, wherein the questionable object is determined by the automated reasoning of said first diagnostic means based on both the fault message and the diagnostic data, and then the questionable object is subjected to execution of a self diagnostic program to carry out a diagnosis.

6. A diagnostic expert system as set forth in claim 1, wherein the data processing apparatus also has a bus connecting the control, memory and input/output units, and

wherein said first and second diagnostic means are provided in a diagnostic apparatus connected via the bus to receive the fault message from the data processing apparatus and start a related diagnosis in response thereto.

7. A diagnostic expert system as set forth in claim 1, w herein said first and second diagnostic means are provided in a work station separate from and connected to the data processing apparatus to receive the fault message.

8. A diagnostic expert system as set forth in claim 1, wherein the data processing apparatus is a duplex data processing apparatus having an active part and a stand-by part, and a self-diagnostic program is executed for the questionable object in the stand-by part.

9. A diagnostic expert system as set forth in claim 8, wherein, if the fault is found in the active part of the questionable object, the related faulty object is replaced by a related functioning object in the stand-by part.

10. A diagnostic expert system as set forth in claim 1, wherein said first and second diagnostic means identify the questionable object based on a diagnostic rule contained in said diagnostic data and, when the questionable object is in a duplex system having an active part and a stand-by part, a diagnosis is made by starting a self-diagnostic program executed by the stand-by part and, when the questionable object is not a duplex system but a single system not having a standby part, a test is achieved under control of said diagnostic control means and, when no fault is found in the questionable object, the automated reasoning is performed by using a corresponding processing procedure for the questionable object with reference to the structure data to finally specify the related faulty object with reference to the test.

11. A diagnostic expert system as recited in claim 1, further comprising a standby part capable of performing self-diagnosis.

12. A diagnostic expert system for diagnosing a data processing apparatus having a control unit, memory unit, and input/output unit, according to a fault message issued from the apparatus due to a fault occurring therein, comprising:

diagnostic memory means for storing therein both diagnostic data and structure data;

diagnostic control means for maintaining the diagnostic data;

first diagnostic means for performing automated reasoning to determine a questionable object, based on both said fault message and said diagnostic data and for diagnosing the questionable object; and

second diagnostic means for hypothesizing the questionable object based on the structure data regarding the data processing apparatus, a related faulty object being specified by said first diagnostic means when possible and, if the related faulty object cannot be specified by said first diagnostic means, said second diagnostic means specifies the related faulty object;

said first diagnostic means diagnosing the questionable object by using a self diagnostic program, and if the related faulty object is not specified, said second diagnostic means determines the related faulty object by performing automated reasoning based on a connection relationship between the control, memory and input/output units comprising the data processing apparatus, and a processing procedure for the questionable object is executed from last to first in a reverse direction.

13. A diagnostic expert system for diagnosing a data processing apparatus having a control unit, memory unit, and input/output unit, according to a fault message issued from the apparatus due to a fault occurring therein, comprising:

diagnostic memory means for storing therein both diagnostic data and structure data;

diagnostic control means for maintaining the diagnostic data;

first diagnostic means for performing automated reasoning to determine a questionable object, based on both said fault message and said diagnostic data and for diagnosing the questionable object; and

second diagnostic means for hypothesizing the questionable object based on the structure data regarding the data processing apparatus, a related faulty object being specified by said first diagnostic means when possible and, if the related faulty object cannot be specified by said first diagnostic means, said second diagnostic means specifies the related faulty object;

said first diagnostic means including

a first part for receiving the fault message, translating the fault message into a corresponding fault number and retrieving a diagnostic rule contained in the diagnostic data using the corresponding fault number, and

a second part for identifying the questionable object based on the diagnostic rule and starting execution of a test program corresponding to the questionable object to carry out a self-diagnosis, and

wherein said second diagnostic means also includes

a third part identifying an object corresponding to the fault message as the questionable object based on the automated reasoning performed by said first diagnostic means,

a fourth part performing a backtrace of a processing procedure step by step with respect to the questionable object, and

a fifth part specifying the questionable object identified as a result of the self-diagnosis achieved in said first diagnostic means.

14. A diagnostic expert system for diagnosing a data processing apparatus having a control unit, memory unit and input/output unit, according to fault messages issued from the apparatus due to faults occurring therein, comprising:

a first knowledge data base containing therein diagnostic rules, created by maintenance technicians, the diagnostic rules defining causes of each fault in connection directly with at least one of the fault messages;

a second knowledge data base, created by design engineers, containing therein data relating to structure and function of the data processing apparatus;

a first reasoning subsystem for performing automated reasoning with an objective of determining a questionable object liable to cause a corresponding fault based upon a receive fault message with reference to said first knowledge data base; and

a second reasoning subsystem for performing automated reasoning to determine the questionable object with reference to said second knowledge data base when the questionable object is not found by the automated reasoning performed by said first reasoning subsystem, a circumstance analysis is carried out by said second reasoning subsystem with reference to the automated reasoning performed by said first reasoning subsystem, and according to the circumstance analysis, a hypothesis regarding the questionable object is generated and verified to find a target faulty object,

wherein in said second reasoning subsystem, the questionable object is identified and a processing procedure corresponding to the questionable object is selected, and the processing procedure is executed from last to first in a reverse direction to search for the target faulty object.

15. A diagnostic expert system as set forth in claim 14,

wherein each of said diagnostic rules is defined by one of the fault messages in correspondence with a related questionable object, and

wherein the data in said second knowledge data base consists of data defining each combination of a connection relationship between the control, memory and input/output units comprising the data processing apparatus and a processing procedure between the control, memory and input/output units for every type of processing.

16. A diagnostic expert system as set forth in claim 14, wherein the questionable object is reasoned by said first reasoning subsystem based on both the received fault message and the diagnostic rules and the questionable object is subjected to execution of a self-diagnostic program to carry out a diagnosis.

17. A diagnostic expert system as set forth in claim 14, wherein said first and second reasoning subsystems identify the questionable object based on the diagnostic rules and, when the questionable object is constructed as a duplex system having a main part and a stand-by part, a diagnosis is made by starting a self-diagnostic program executed by the stand-by part thereof, when the questionable object is not a duplex system but a single system, a test is carried out under control of said diagnostic control means cooperating with said first and second reasoning subsystems, and when no fault is found in the questionable object, the automated reasoning is achieved by using a corresponding processing procedure for the questionable object with reference to the data to finally specify the target faulty object with reference to the automated reasoning of said first reasoning subsystem.

18. A diagnostic expert system as set forth in claim 14, wherein said first reasoning subsystem includes an expert shell reasoning unit to perform the automated reasoning with reference to said first knowledge data base.

19. A diagnostic expert system as set forth in claim 18, wherein said second reasoning subsystem includes a design knowledge reasoning unit to perform the automated reasoning with reference to the second knowledge base.

20. A diagnostic expert system as set forth in claim 19, further comprising a reasoning explanation unit for outputting explanations of the automated reasonings from said first and second reasoning subsystems to a maintenance technician.

21. A diagnostic expert system as set forth in cl aim 14, further comprising a knowledge base management unit for managing the diagnostic rules and the data in said first and second knowledge data bases, respectively.

22. A diagnostic expert system for diagnosing a data processing apparatus having a control unit, memory unit and input/output unit, according to fault messages issued from the apparatus due to faults occurring therein, comprising:

a first knowledge data base containing therein diagnostic rules, created by maintenance technicians, the diagnostic rules defining causes of each fault in connection directly with at least one of the fault messages;

a second knowledge data base, created by design engineers, containing therein data relating to structure and function of the data processing apparatus;

a first reasoning subsystem for performing automated reasoning with an objective of determining a questionable object liable to cause a corresponding fault based upon a received fault message with reference to said first knowledge data base; and

a second reasoning subsystem for performing automated reasoning to determine the questionable object with reference to said second knowledge data base when the questionable object is not found by the automated reasoning performed by said first reasoning subsystem, a circumstance analysis is carried out by said second reasoning subsystem with reference to the automated reasoning performed by said first reasoning subsystem, and according to the circumstance analysis, a hypothesis regarding the questionable object is generated and verified to find a target faulty object;

said first reasoning subsystem including

first means for receiving the fault messages, for translating each fault message into a corresponding fault number and for retrieving a diagnostic rule using the corresponding fault number; and

second means for identifying the questionable object based on the diagnostic rule and starting execution of a test program corresponding to the questionable object to carry out a self-diagnosis, and

wherein said second reasoning subsystem includes

third means for identifying an object corresponding to one of the fault messages as the questionable object based on the automated reasoning performed by said first reasoning subsystem,

fourth means for performing a backtrace of a processing procedure step by step with respect to the questionable object, and

fifth means for specifying the questionable object identified as a result of the self-diagnosis achieved in said first reasoning subsystem.

23. A method of diagnosing a fault occurring in an operating device having components, comprising the steps of:

(a) storing diagnostic data relating to the operating device;

(b) storing structure data defining relationships between the components of the operating device;

(c) evaluating the diagnostic data in a data processing apparatus with an objective of identifying a potential source of the fault;

(d) evaluating the structure data to hypothesize the potential source of the fault, in dependence upon said evaluating in step (c); and

(e) executing a processing procedure corresponding to the potential source in reverse direction to determine the source of the fault.

24. A method as set forth in claim 23, wherein said storing in step (a) includes the step of defining the diagnostic data by a correspondence between at least one fault indication generated by the operating device and a cause of the fault indication.

25. A method as recited in claim 23, wherein said storing in step (b) comprises the steps of:

(b1) defining each of the structure data by a physical connection relationship between at least two of the components of the operating device; and

(b2) defining for each of the structure data a processing procedure shared by the at least two of the components of the operating device.

26. A method as recited in claim 23, further comprising the step of (e) executing self-diagnostics in the operative device to obtain information on the potential source of the fault.

27. A method as recited in claim 26, wherein step (c) comprises the steps of:

(c1) retrieving a diagnostic rule from the diagnostic data based on a received fault message; and

(c2) executing self-diagnostics on a potentially faulty component corresponding to the potential source of the fault indicated by the diagnostic rule retrieved in step (c1).

28. A method as recited in claim 27, wherein said operating device is a duplex data processing device having stand-by parts, wherein said method further comprises the step of:

(f) replacing the potentially faulty part by one of the stand-by parts, and

wherein said executing of the self-diagnostics in step (c2) is performed on the potentially faulty part after said replacing in step (f).

29. A method as recited in claim 23,

further comprising the step of (e) executing self-diagnostics within the potential source, and

wherein said evaluating in step (d) evaluates the relationships defining connections between the components of the operating device when said executing of the self-diagnostics in step (e) does not identify the source of the fault.

30. A method as recited in claim 23, wherein the structure data is obtained from design engineering experience.