Failure-recovering sequential procedure teaching system and method thereof

• US 5,590,036 A
• Filed: 06/01/1995
• Issued: 12/31/1996
• Est. Priority Date: 06/01/1994
• Status: Expired due to Fees

• Current Assignee: Nippondenso Co., Ltd. (DENSO Corporation)
• Original Assignee: Nippondenso Co., Ltd. (DENSO Corporation)
• Inventors: Maeda, Noboru
• Primary Examiner(s): Envall, Jr., Roy N.
• Assistant Examiner(s): PRASS, RONALD E
• Application Number:
  US08/457,281
• Time in Patent Office: 579 Days
• Field of Search: 364/184, 364/185, 364/186, 364/187, 364/468, 364/550, 364/551.01, 364/551.02, 371/48, 371/67.1, 371/69.1, 395/181, 395/182.04
• US Class Current: 700/79
• CPC Class Codes: G05B 9/02 electric

Abstract

When a piece of failure occurs, the associated-portion state data items of a relevant equipment and the relevant recovery sequence operated by an operator are recorded as a set of data. At this time, the associated-portion state data items which fall within a plurality of sets of data and which correspond to the same recovery sequence are grouped as the same recovery sequence group. Simultaneously, the common elements to the associated-portion state data items which fall within each recovery sequence group are extracted and recorded as the elements paired with the corresponding recovery sequence. Thereby, when a new piece of failure occurs, relevant associated-portion state data items are compared with the common elements. Then, regarding the associated-portion state data item which coincides with the common element, the corresponding recovery sequence is outputted to teach it to the operator as an available recovery sequence. The failure state data items recoverable with the use of the same recovery sequence are learned by the system per se while analysis is being sequentially performed on the common elements, thereby producing accurate teach sequences with respect to a large number of failure state items by using a small number of data items base.

Claims

1. A failure-recovering sequential procedure teaching system comprises:

a failure removal data recording section which records as a set of failure removal data an associated-portion state data representing prescribed associated-portion states at the time when the equipment to be controlled is in failure and a recovery sequence data representing a recovery sequencial procedure with which an operator performs his recovering operation;

a failure removal data analysis section which selects one or a plurality of associated-portion state data items corresponding to the same recovery sequence data among a plurality of sets of failure removal data and groups them as data of the same recovery sequence data, analizes said associated-portion state data in each recovery sequence group to produce failure-characterizing data typically representing the characteristics of the relevant failure state, and records the failure-characterizing data as the data paired with the corresponding recovery sequence data; and

a recovery sequence determination section which compares the corresponding associated-portion state data with each item of the failure-characterizing data as stored to determine the degree of match between an item of said corresponding associated-portion state data when a new piece of failure occurs and outputs the recovery sequence data corresponding to a high match-degree item of the failure-characterizing data.

2. A failure-recovering sequential procedure teaching system as set forth in claim 1, wherein said failure removal data recording section records said recovery sequence data which have been normalized in accordance with a normalization rule representing as a single sequential-order operation-step chain formed by commonizing a plurality of operation steps eventually bringing about the same operation effects even when such operation steps are different from each other.

3. A failure-recovering sequential procedure teaching system as set forth in claim 2, wherein said failure removal data analysis section adds a normalization rule for converting another recovery sequence to a certain recovery sequence and merges said another recovery sequence into said certain recovery sequence in case said failure-characterizing data corresponds to said certain recovery sequence includes the failure-characterizing data corresponding to said another recovery sequence.

4. A failure-recovering sequential procedure teaching system as set forth in claim 1, wherein said failure removal data analysis section converts said associated-portion state data into vectors, and generates failure state vectors including as its elements common data items of the associated-portion state data falling within the recovery sequence group and mask vectors having data items for excluding data item elements other than such common data-item elements, so that said failure state vector and the mask vector paired therewith are processed as the failure-characterizing data while automatically learning from the failure occurrence and recovery operations.

5. A failure-recovering sequential procedure teaching system as set forth in claim 4, wherein said failure removal data analysis section processes as said failure characterizing data a total number of all failure state vectors corresponding to a same recovery sequence and a mean value thereof and a set of data composed of variance covariance matrix or a set of data necessary for deriving those data.

6. A method of teaching failure-recovering sequential procedure comprising steps of:

recording as a set of failure removal data an associated-portion state data representing prescribed associated-portion state at time when an equipment to be controlled is in failure and a recovery sequence data representing a recovery sequencial procedure with which an operator performs his recovering operation;

selecting one or a plurality of associated-portion state data items corresponding to same recovery sequence data among a plurality of sets of said failure removal data;

grouping said associated-portion state data items as data of said same recovery sequence data;

analyzing said associated-portion state data in each group of said recovery sequence data to produce failure-characterizing data typically representing characteristics of a relevant failure state; and

recording said failure-characterizing data as data paired with corresponding data of said recovery sequence data;

comparing said corresponding associated-portion state data with each item of said recorded failure-characterizing data to determine degree of match between an item of said corresponding associated-portion state data when a new piece of failure occurs; and

outputing said recovery sequence data corresponding to a high match-degree item of said failure-characterizing data.

7. A method as set forth in claim 6, wherein said step of recording said recovery sequence data comprises steps of:

commonizing a plurality of operation steps included in said recovery sequence data eventually bringing about the same operation effects even when such operation steps are different from each other; and

standarding said recovery sequence data in accordance with a normalization rule representing as a single sequential-order operation-step chain.

8. A method as set forth in claim 7, wherein said step of standardizing said recovery sequence data comprises steps of:

adding a normalization rule for converting another recovery sequence to a certain recovery sequence; and

merging said another recovery sequence into said certain recovery sequence in case said failure-characterizing data corresponds to said certain recovery sequence includes the failure-characterizing data corresponding to said another recovery sequence.

9. A method as set forth in claim 6, wherein

said step of recording a set of failure removable data further comprises steps of comverting said associated-portion state data into vectors, and generating failure state vectors including as its elements common data items of the associated-portion state data falling within the recovery sequence group and mask vectors having data items for excluding data item elements other than such common data-item elements;

and

said step of recording failure characterizing data comprises a step of processing said failure state vector and said mask vector paired therewith as failure-characterizing data.

10. A method as set forth in 9, wherein said step of failure state vetor comprises a step of processing as said failure characterizing data a total number of all failure state vectors corresponding to a same recovery sequence and a mean value thereof and a set of data composed of variance covariance matrix or a set of data necessary for deriving those data.