WO2016167057A1

WO2016167057A1 - Trouble information utilization assist device and trouble information utilization assist method

Info

Publication number: WO2016167057A1
Application number: PCT/JP2016/057388
Authority: WO
Inventors: 清水　勇喜; 野中　紀彦
Original assignee: 株式会社日立製作所
Priority date: 2015-04-16
Filing date: 2016-03-09
Publication date: 2016-10-20

Abstract

This trouble information utilization assist device is provided with a storage unit in which are stored trouble information written about trouble with an apparatus by a sentence for each item, and a causal model having a plurality of words included in the trouble information. This trouble information utilization assist device is also provided with a control unit in which: candidates for elements on the cause side of the causal model and candidates for elements on the effect side of the causal model are extracted, in accordance with the category to which a word belongs and the part of speech for the word, from the plurality of words included in the trouble information; candidates that have an expression trend as a component and candidates that have an expression trend as a phenomenon are determined as cause-of-fault occurrence components and stress, respectively, from among the extracted candidates for the elements on the cause side; candidates that have an expression trend as a component and candidates that have an expression trend as a phenomenon are determined as fault components and fault modes, respectively, from among the extracted candidates for the elements on the effect side; and a causal model having the determined cause-of-fault occurrence components, stress, fault components, and fault modes is stored in the storage unit.

Description

Defect information utilization support device and defect information utilization support method

The present invention relates to a defect information utilization support device and a defect information utilization support method.

FMEA (FailureailMode and Effects Analysis) is a method for preventing device failures. In FMEA, failure modes that can occur in parts are predicted at the time of device design, problems are identified by analyzing and evaluating possible causes and effects, and countermeasures are taken. The failure mode is a classification of a failure state of the device, such as disconnection, short circuit, breakage, wear, deterioration of characteristics, and the like. The failure parts, failure modes, causes, effects, countermeasures, and the like, which are the results of the study by FMEA, are summarized in a table format. This table is called FMEA table. The FMEA table shows failure modes and causes that may occur when the same parts are used for other equipment, not only for FMEA practitioners to organize the examination results but also for third parties to check the examination results. It is also used for confirmation.

For example, if the failed part is a coil, “coil” is entered in the failed part column of the FMEA table. In the failure mode column, “disconnection”, “short circuit”, “abnormal resistance value”, etc. are described, and then in the cause column, the cause of the failure mode is “high load”, “immersion”, “ Indicate “Wire failure”. Next, the examination result is written in the influence column as the influence on the failure mode, and the examination result is written in the countermeasure column as a countermeasure against the failure mode. Thus, the failure mode becomes a reference point for work. Therefore, if a failure mode selection is missed, subsequent causes, effects, and countermeasures cannot be examined accurately. In other words, it is important to identify the failure mode without omission when creating the FMEA table.

多大 It takes a lot of effort to identify the failure mode without omission. When all failure modes are not grasped, it is necessary to extract a part corresponding to the part from an example of a problem that has occurred in the past (defect information). In order to efficiently perform this extraction, it is effective to organize defect information in the form of a causal model. The causal model is, for example, a combination of a failed part, a failure mode, and stress. Among them, the stress is a load imposed on the component at the time of manufacture or use of the component, such as disconnection, short circuit, and abnormal resistance value, and is a cause of malfunction. If there is such a causal model, the failure mode can be identified without omission using the failed part as a search key.

The design support apparatus described in Patent Document 1 stores a “fault record” (FIG. 11 of Patent Document 1) which is an example of such a causal model. Then, the component item (failed part), the control attribute, and the stress attribute are received, all defect records related to the received information are acquired and displayed, and the causal chain relationship between the defect records is also displayed (Patent Document 1). Paragraph “B: 1” last part).

International Publication No. 2005/101253

If the design support apparatus of Patent Document 1 is used, a failure record is searched using a failed part as a search key, and the FMEA table can be filled with a failure mode (failure mode) included in the failure record. However, in practice, a causal model must be extracted from a large amount of defect information generated in the past and prepared in advance as a database. In order to avoid the database becoming obsolete, a causal model created from new defect information must be added to the database. That is, a great deal of labor is required to create and update the causal model (defect record).

However, the design support apparatus of Patent Document 1 is based on the assumption that a sufficient number of defect records already exist, and is not particularly aimed at creating / updating defect records. Therefore, a separate measure is required to reduce such labor while using the design support apparatus of Patent Document 1.
Therefore, an object of the present invention is to efficiently extract a causal model from defect information.

The defect information utilization support apparatus of the present invention includes a storage unit that stores defect information in which a defect of a device is described in a sentence for each item, a causal model having a plurality of words included in the defect information, and the word belongs Depending on the item and the part of speech of the word, the cause-side element candidate of the causal model and the result-side element candidate of the causal model are extracted from a plurality of words included in the defect information, and the extracted cause-side element Of the candidates, those having an expression tendency as a component and those having an expression tendency as a phenomenon are determined as a failure factor component and a stress, respectively, and among the extracted result side element candidates as a component And those having an expression tendency as a phenomenon are determined as a failure part and a failure mode, respectively, and the determined failure cause component and stress are determined respectively. A control unit for storing the failed component and causal model with a failure mode in the storage unit, characterized in that it comprises a.
Other means will be described in the embodiment for carrying out the invention.

According to the present invention, a causal model can be efficiently extracted from defect information.

It is a figure explaining the structure of a malfunction information utilization assistance apparatus. It is a flowchart of a malfunction information and causal model registration process procedure. It is a flowchart of a FMEA table creation processing procedure. It is a flowchart of a causal model candidate extraction process procedure. 10 is a flowchart of a part / phenomenon determination processing procedure. It is a figure explaining an example of a malfunction information registration screen. It is a figure explaining an example of a causal model registration / edit screen. It is a figure explaining an example of a FMEA table creation screen. It is a figure explaining an example of the result of morphological analysis. (A) And (c) is a figure explaining an example of the calculation result of IEPH. (B) And (d) is a figure explaining an example of the calculation result of IEPA.

Hereinafter, two embodiments for carrying out the present invention will be described in detail with reference to the drawings and the like. The two embodiments consist of the first embodiment as a basic type and the second embodiment as an application type. Although details will be described later, the difference between them is whether or not the defect information utilization support apparatus has an account management unit and a payment / billing management unit. First, a first embodiment that does not have an account management unit and a payment / billing management unit will be described, and then a second embodiment that has an account management unit and a payment / billing management unit will be referred to as the first embodiment. This will be explained by paying attention to the differences.

(First embodiment)
(Equipment configuration)
The configuration of the defect information utilization support apparatus 2 will be described with reference to FIG. The defect information utilization support system 1 includes a defect information utilization support device 2 and a terminal device 3. These can be connected via the network 4. The defect information utilization support device 2 is a general computer, and includes a central control device 11, an input device 12, an output device 13, a main storage device 14, and an auxiliary storage device 15. These are connected to each other by a bus. The auxiliary storage device 15 stores a defect information database 31 and a causal model database 32 (details will be described later). Hereinafter, these are abbreviated as a defect information DB (Data Base) 31 and a causal model DB 32.
The auxiliary storage device 15 is an external storage device that is independent from the defect information utilization support device 2, and both can be connected via the network 4.

FMEA table creation unit 21, causal model search unit 22, defect information registration unit 23, public defect information collection unit 24, causal model extraction unit 25, causal model registration / editing unit 26, account management unit 27, and the like in main storage device 14 The payment / billing management unit 28 is a program. Thereafter, when the subject is described as “XX section”, the central control device 11 reads out each program from the auxiliary storage device 15 and loads it into the main storage device 14, and then the function of each program (detailed later). Shall be realized. The causal model extraction unit 25 includes a causal model candidate extraction unit 25a and a part / phenomenon determination unit 25b.

The terminal device 3 is also a general computer and includes a central control device, an input device, an output device, a main storage device, and an auxiliary storage device (not shown). These are connected to each other by a bus. For convenience of explanation, FIG. 1 also shows the terminal device 3, the account management unit 27, and the payment / billing management unit 28 that exist only in the second embodiment (description of the maximum configuration).

(Bug information)
The defect information is a report that describes the malfunction of the device that has occurred in the past as a sentence for each item (see reference numeral 131 in FIG. 7). Specifically, the defect information includes items such as a defect name, a progress until the occurrence of the defect, a (defect) phenomenon, a (defect) cause, and a countermeasure (for the defect). Each item has a sentence (character string).

(Causal model)
As described above, the causal model is a combination of a failed part, a failure mode, and stress. However, since it is necessary to distinguish between a failure part and a failure factor part, the causal model of this embodiment is a combination of a failure factor part, stress, a failure part, and a failure mode (reference numeral 141 in FIG. 7). reference). Incidentally, the failure-causing component is a component that is directly stressed. A faulty part is a part that actually caused a fault, that is, a fault mode. Although the failure-causing component itself is stressed, it does not always have a failure. Failure cause component, stress, failure component, and failure mode are called “elements” of the causal model.

Of the four types of elements, failure-causing components and stress are causal elements of causality. Of the elements, the failed part and the failure mode are elements on the result side of the causal relationship. Each element is, for example, a word such as “base” or a composite word composed of a plurality of consecutive words such as “electronic system”, and is suitable for being searched at high speed. On the other hand, since the above-described defect information is originally a report, it is not in a model suitable for a high-speed search.

(FMEA table)
The FMEA table is a table in which failure parts, failure modes, failure occurrence cause parts, stress, influence, and countermeasures are associated with each other (see reference numerals 151 to 155 in FIG. 8). Among these, the failure part, the failure mode, the failure factor part, and the stress are also elements of the causal model. Note that, as described above, the countermeasure 136 in FIG. 7 is a countermeasure against the failure mainly from the viewpoint of the field maintenance, whereas the countermeasure 155 in FIG. 8 is a countermeasure against the failure mode mainly from the part design aspect. .

(Processing procedure)
Hereinafter, four processing procedures will be described. These are (1) defect information / causal model registration processing procedure, (2) FMEA table creation processing procedure, (3) causal model candidate extraction processing procedure, and (4) part / phenomenon discrimination processing procedure. (3) and (4) are the subroutine of (1). In order for (2) to start, it is assumed that (1) has ended.

(Problem information / causal model registration procedure)
The defect information / causal model registration processing procedure will be described with reference to FIG.
In step S201, the defect information registration unit 23 displays a defect information input screen 51 (FIG. 6). Specifically, the defect information registration unit 23 displays a defect information input screen 51 on the output device 13. The defect information input screen 51 has a direct input field 101 and a file designation field 111. The direct input column 101 further includes a defect name column 102, a progress column 103 until a defect occurs, a phenomenon column 104, a cause column 105, and a countermeasure column 106.

The user may have already created the defect information and stored the created defect information in an arbitrary storage device. On the other hand, the user may not have created the defect information yet. In the former case, the user inputs information indicating the location of the created defect information in the external address input field 112 of the file specification field 111. When the user presses the reference field 113, the created defect information can be viewed. In the latter case, the user inputs sentences (character strings) about the items in the respective fields 102 to 106 of the direct input field 101 via the input device 12. The user presses the cancel button 122 when interrupting the operation without inputting.

The user presses the registration button 121 to register the input data. If the user does not want to register the input data, the user presses the cancel button 122. The defect information registration unit 23 accepts these operations by the user.

In step S202, the defect information registration unit 23 determines whether or not the registration button 121 has been pressed. Specifically, the defect information registration unit 23 proceeds to step S204 when accepting that the user presses the registration button 121 (step S202 “Yes”), and accepts that the user presses the cancel button 122. If so (step S202 "No"), the process proceeds to step S203.

In step S203, the defect information registration unit 23 closes the defect information input screen 51. Thereafter, the defect information / causal model registration processing procedure is terminated.

In step S204, the defect information registration unit 23 registers defect information. Specifically, the defect information registration unit 23 registers the defect information received in step S201 in the defect information DB 31.

In step S205, the causal model candidate extraction unit 25a extracts the causal model element candidates. Although details of step S205 will be described later, as a result, the causal model candidate extraction unit 25a causes the cause-side element candidate (stress candidate or failure-causing component candidate) and the result-side element candidate (failure). Mode candidates or faulty part candidates) are extracted.

In step S206, the part / phenomenon determination unit 25b determines a final element from the candidates. Although details of step S206 will be described later, as a result, the component / phenomenon determination unit 25b determines a final stress and a final failure-causing component from the cause-side element candidates. Furthermore, the final failure mode and the final failure component are determined from the candidate elements on the result side. “Final” means selected from the candidates.

In step S207, the causal model registration / editing unit 26 displays the causal model registration / editing screen 52 (FIG. 7). Specifically, first, the causal model registration / editing unit 26 displays the causal model registration / editing screen 52 on the output device 13, and displays the defect information registered in step S204 in the defect information column 131. To do. The configuration of the defect information column 131 is the same as that of the direct input column 101 in FIG. 6, but specific sentences are displayed in the respective columns 132 to 136 of the defect information column 131.

Secondly, the causal model registration / editing unit 26 displays the final elements determined in step S206 in the respective columns 142 to 145 of the causal model column 141. The user visually recognizes the defect information and the causal model at the same time, and confirms whether or not the causal model sufficiently expresses the characteristic of the defect information as an accident example.
Third, the causal model registration / editing unit 26 accepts that the user edits the elements displayed in the respective columns 142 to 145 of the causal model column 141 as necessary.
Fourth, the causal model registration / editing unit 26 accepts that the user presses either the registration button 146 or the discard button 147.

In step S208, the causal model registration / editing unit 26 determines whether or not the discard button 147 has been pressed. Specifically, when the causal model registration / editing unit 26 accepts that the user presses the discard button 147 (step S208 “Yes”), the process proceeds to step S209, and the user presses the registration button 146. If accepted (step S208 “No”), the process proceeds to step S210.

In step S209, the causal model registration / editing unit 26 discards the causal model. Specifically, the causal model registration / editing unit 26 discards the causal model displayed in the causal model column 141 of the causal model registration / editing screen 52, and then ends the defect information / causal model registration processing procedure.

In step S210, the causal model registration / editing unit 26 determines whether or not the causal model has been edited. Specifically, when the causal model registration / editing unit 26 accepts editing in “third” in step S207 (step S210 “Yes”), the causal model registration / editing unit 26 proceeds to step S211 and does not accept editing (step S210 “ No "), go to step S212.

In step S211, the causal model registration / editing unit 26 registers the edited causal model. Specifically, the causal model registration / editing unit 26 registers (stores) the causal model after the editing is accepted in “third” in step S207 in the causal model DB 32. Thereafter, the defect information / causal model registration processing procedure is terminated.

In step S212, the causal model registration / editing unit 26 registers a causal model that has not been edited. Specifically, the causal model registration / editing unit 26 registers (stores) the causal model displayed in “second” in step S207 in the causal model DB 32. Thereafter, the defect information / causal model registration processing procedure is terminated. There are multiple pieces of defect information. Therefore, the defect information / causal model registration process procedure is repeated for each defect information, and the causal model DB 32 stores a number of causal models equal to the number of defect information.

(FMEA table creation processing procedure)
The FMEA table creation processing procedure will be described with reference to FIG.
In step S221, the FMEA table creation unit 21 displays the FMEA table creation screen 53 (FIG. 8). Specifically, the FMEA table creation unit 21 displays an MEA table creation screen 53 on the output device 13.

In step S222, the FMEA table creation unit 21 determines whether or not the cancel button 163 is pressed. Specifically, when the FMEA table creation unit 21 accepts that the user presses the cancel button 163 (step S222 “Yes”), the process proceeds to step S223, and otherwise (step S222 “No”). Proceed to step S224.

In step S223, the FMEA table creation unit 21 closes the FMEA table creation screen 53 and ends the FMEA table creation processing procedure.

In step S224, the causal model search unit 22 receives a failed part. Specifically, the causal model search unit 22 accepts that the user inputs a faulty part (name) in the faulty part column 151 of the FMEA table creation screen 53 and then presses the causal model search button 161.

In step S225, the causal model search unit 22 searches for a causal model. Specifically, the causal model search unit 22 searches the causal model DB 32 using the faulty part received in step S224 as a search key, and acquires all corresponding causal models.
Here, an example has been described in which the causal model search unit 22 receives “failed parts” and uses the “failed parts” as a search key. However, the causal model search unit 22 may receive elements other than “failed parts” among the elements of the causal model and use them as search keys.

In step S226, the causal model search unit 22 displays the causal model. Specifically, first, the causal model search unit 22 uses the failure mode column 152, the failure mode column 152 of the FMEA table creation screen 53 to indicate the failure mode, the failure factor component, and the stress of the causal model acquired in step S225, respectively. It is displayed (output) in the failure factor component field 153a and the stress field 153b. When a plurality (n) of causal models are acquired, n rows are displayed in association with the failed part that is the search key. The causal model search unit 22 leaves the influence column 154 and the countermeasure column 155 of each line blank.

Secondly, the causal model search unit 22 accepts that the user edits at least one of the failure mode, the failure-causing factor component, and the stress of each row displayed in “first” as necessary.
Third, the causal model search unit 22 allows the user to input the influence and the countermeasure in the influence column 154 and the countermeasure column 155 of each line displayed in “first”, and then press the registration button 162. Accept. Here, the cause-and-effect model search unit 22 receives the influence and the countermeasure are merely examples. However, the present invention is not limited to this, and the causal model search unit 22 may accept that the user inputs general additional information for the causal model.

In step S227, the causal model search unit 22 registers the FMEA table. Specifically, the causal model search unit 22 associates the causal model edited as necessary in “second” in step S226 with the influence and countermeasure (more generally, input in “third”). Is registered (stored) in the auxiliary storage device 15.
Thereafter, the FMEA table creation processing procedure is terminated.

(Modification of FMEA table creation processing procedure)
An example in which the user manually inputs a failed part in step S224 has been described. However, failure documents are often included in any document or data that the user has already created. For example, existing design specifications, existing three-dimensional CAD (Computer Aided Design) data, and the like often include a failed part (name). Therefore, in step S224, the causal model search unit 22 accesses a document or data designated in advance without receiving a manual input from the user, and a faulty part (name) included in the document or data. May be obtained. In this case, it is desirable that the location where the failed part (name) is described in the document or data is specified by a tag or the like. The causal model search unit 22 may acquire an element other than the “failed part” among the elements of the causal model instead of the failed part (name).

(Causal model candidate extraction processing procedure)
The causal model candidate extraction processing procedure will be described with reference to FIG. The causal model candidate extraction processing procedure is details of step S205 of the defect information / causal model registration processing procedure.
In step S241, the causal model candidate extraction unit 25a unifies the notation of defect information. Specifically, first, the causal model candidate extraction unit 25a acquires defect information from the defect information DB 31.

Secondly, the causal model candidate extraction unit 25a performs the following unified processing on the sentence of each item of defect name of the acquired defect information, progress until the occurrence of the defect, phenomenon, cause, and countermeasure.
-Delete line breaks.
・ Delete spaces (both full-width and half-width).
・ Convert katakana that is not full-width to full-width.
・ Convert English characters that are not full-width or capital letters to full-width or capital letters.

In step S242, the causal model candidate extraction unit 25a performs morphological analysis. Specifically, first, the causal model candidate extraction unit 25a breaks down sentences of each item of defect information subjected to the unification processing into words.
Secondly, the causal model candidate extraction unit 25a acquires a part of speech for each decomposed word using a dictionary (not shown). FIG. 9 shows a result of disassembling the sentence “signal failure of electronic system due to whisker” into words, and obtaining a part of speech (column 172) for each word (column 171). The text often includes technical terminology such as “whiskers”. Since the word “whisker” does not exist in the dictionary, the causal model candidate extraction unit 25a acquires “unknown word” as the part of speech of “whisker”.

Thirdly, when the nouns (including unknown words) are continuous in the original sentence, the causal model candidate extraction unit 25a combines them into one word (synthetic word). For example, “electronic” 173a and “system” 173b are combined into an “electronic system”. Further, the “signal” 174a and the “failure” 174b are combined to form a “signal failure”.
The causal model candidate extraction unit 25a repeats the “second” and “third” processes for each item of inconvenience information.

In step S243, the causal model candidate extraction unit 25a extracts failure mode candidates and failure component candidates from the defect information. Specifically, the causal model candidate extraction unit 25a acquires a word whose part of speech is a noun from the sentence of the defect information item “defect name” 102 and sets it as a failure mode candidate or a failure part candidate. The nouns here include unknown words, and the words include synthesized words. And the word made into a candidate here comprises the candidate of the above-mentioned result side element. The process of step S243 is based on an empirical rule that the defect name sentence often includes a failure mode or a failure part so that the content of the defect can be immediately understood by looking at the sentence.

In step S244, the causal model candidate extraction unit 25a extracts a failure mode candidate and a failure part candidate from the progress and phenomenon until the failure occurs. Specifically, the causal model candidate extraction unit 25a acquires words (including synthesized words) that satisfy all of the following conditions, and sets them as failure mode candidates or failure part candidates.
(Condition 1) The item belongs to the item “Progress until failure” 103 of the failure information.
(Condition 2) It belongs to the item “phenomenon” 104 of defect information.
(Condition 3) The part of speech is a noun (including an unknown word).

The words that are candidates here also constitute candidates for the result-side elements described above. The progress until the occurrence of a problem describes the progress until the phenomenon occurs. Due to this, the processing in step S244 is based on an empirical rule that the passage of failure and the passage of phenomenon often include the same failure mode or the same failed part in common.

In step S245, the causal model candidate extraction unit 25a extracts stress candidates and failure occurrence factor component candidates from the causes and countermeasures. Specifically, the causal model candidate extraction unit 25a acquires a word (including a synthesized word) that satisfies all of the following conditions and sets it as a stress candidate or a failure cause component candidate.
(Condition 11) It belongs to the item “Cause” 105 of the defect information.
(Condition 12) It belongs to the item “Countermeasure” 106 of the defect information.
(Condition 13) The part of speech is a noun (including an unknown word).

The words that are candidates here constitute candidates for the cause-side elements described above. The countermeasure describes the countermeasure against the stress applied to the component described in the cause. Due to this, the processing in step S245 is based on an empirical rule that the cause sentence and the countermeasure sentence often include the same stress or the same failure cause component in common.
Thereafter, the causal model candidate extraction processing procedure is terminated (proceeds to step S206).

(Part / phenomenon discrimination processing procedure)
Before explaining the details of the part / phenomenon discrimination processing procedure, the concept of the procedure will be described. The part / phenomenon determination unit 25b calculates a value of “IEPH” and a value of “IEPA” for each word (including a synthesized word) extracted in the causal model candidate extraction processing procedure. The value of “IEPH” indicates the possibility that the word belongs to the following “expression tendency 1” (possibility that seems to be a phenomenon). The value of “IEPA” indicates the possibility that the word belongs to the following “expression tendency 2” (possibility that seems to be a part).

(Expression tendency 1) An expression tendency determined to be a phenomenon when a word such as “occurs” or “occurs” immediately follows the word. For example, such words often follow immediately after phenomena such as “residual stress” and “high temperature degradation”.
(Expression tendency 2) An expression tendency determined to be a part by immediately following the word followed by a word such as “no failure” or “purchase of”. For example, such words often follow immediately after parts such as “impeller” and “capacitor”.

Among the causal element candidates, the larger the “IEPH” value, the higher the possibility of stress. Among the cause-side candidate elements, the larger the value of “IEPA”, the higher the possibility that it is a failure-causing component. Of the candidate elements on the result side, the higher the value of “IEPH” is, the higher the possibility of the failure mode is. Of the candidate elements on the result side, the larger the value of “IEPA”, the higher the possibility of a failed part. Hereinafter, calculation formulas of “IEPH” and “IEPA” will be described.

(IEPH)
・ IEPH = PHL × TF × IDF
・ PHL = PHDF / TDF
・ IDF = log (K / DF)
The PHDF has at least one case in which the word appears with a word “occurs”, “occurs”, or “occurs” immediately after the defect information stored in the defect information DB 31. It is the number of defect information.
TDF is the number of pieces of defect information having at least one case where the word appears among the pieces of defect information stored in the defect information DB 31.
TF is the number of occurrences of the word in the cause and countermeasure items of the defect information.
K is the total number of defect information stored in the defect information DB 31.
DF is the number of pieces of defect information including the word among the defect information stored in the defect information DB 31.
“Log (K / DF)” is the natural logarithm (or common logarithm) of “K / DF”.

(IEPA)
・ IEPA = PAL × TF × IDF
・ PAL = PALDF / TDF
・ IDF = log (K / DF)
-PALDF uses the defect information stored in the defect information DB 31 immediately after the word “defect”, “purchase”, “acceptance”, “development” or “development”. It is the number of pieces of defect information having at least one occurrence of accompanying information.
TF is a value obtained by multiplying the number of appearances of the word in the defect name item of the defect information by a predetermined weight, the number of appearances of the word in the item of progress until the defect occurs in the defect information, This is the total value of the number of appearances of the word in the defect information phenomenon item. Since the defect name often includes a word representing the characteristic of the defect, it is multiplied by a weight (weight> 1).
TDF, K and DF are the same as described for “IEPH”.

(Part / phenomenon discrimination processing procedure)
The part / phenomenon determination processing procedure will be described with reference to FIG. The part / phenomenon determination processing procedure is the details of step S206 of the defect information / causal model registration processing procedure.
In step S251, the part / phenomenon determination unit 25b calculates an IEPH value for each of the stress candidate and the failure cause component candidate.

In step S252, the part / phenomenon determination unit 25b determines the word having the maximum IEPH value as the final stress. That is, the most probable phenomenon is determined as a “phenomenon” from the causal element candidates, and the phenomenon is defined as “stress”. Note that the component / phenomenon determination unit 25b excludes words that are not clearly phenomena as exemplified below from determination targets.
・ The word string at the end is a part name such as “part”, “device”, “product”, “machine”, “container”, etc. “change”, “production”, “use”, “assembly”, Words indicating human actions such as “execution” and “driving” ・ Words indicating time and frequency such as “morning”, “hour”, “year”, “month”, “day” “local”, “base” ”,“ Nearby ”,“ Place ”, etc., a word representing a place,“ Customer ”,“ Customer ”,“ Manufacturer ”,“ User ”, etc.

The calculation result of IEPH in step S252 will be described with reference to FIG. The part / phenomenon discriminating unit 25b calculates “2.5 × 10 ⁻³ ” as the IEPH value of the candidate “foundation” and “2.3” as the IEPH value of the candidate “whisker”. Therefore, the part / phenomenon determination unit 25b determines “whisker” as the final stress.

In step S253, the component / phenomenon determination unit 25b calculates an IEPA value for each of the stress candidate and the failure cause component candidate.

In step S254, the component / phenomenon determination unit 25b determines the word having the maximum IEPA value as the final failure factor component. In other words, the most likely part of the causal element candidates is determined as a “part”, and the part is designated as a “failure-causing part”. The part / phenomenon determination unit 25b excludes words that are clearly not parts from the determination target. Examples of words that are clearly not parts are:
-Words extracted as the final stress-The word string at the end represents a phenomenon such as "failure", "failure", "failure", "occurrence", "stop", "bad"

The calculation result of IEPA in step S254 will be described with reference to FIG. The part / phenomenon discriminating unit 25b calculates “1.1 × 10 ⁻¹ ” as the IEPA value of the candidate “base” and “0” as the IEPA value of the candidate “whisker”. Therefore, the part / phenomenon determination unit 25b determines “base” as the final failure factor part. “Whisker” has already been determined as the final stress, but even if this is not the case, the “base” will be determined as the final failure factor component.

In step S255, the part / phenomenon determination unit 25b calculates an IEPH value for each of the failure mode candidate and the failure part candidate.

In step S256, the part / phenomenon determination unit 25b determines the word having the maximum IEPH value as the final failure mode. That is, the most likely phenomenon among the candidate elements on the result side is determined as a “phenomenon”, and the phenomenon is designated as a “failure mode”. The part / phenomenon determination unit 25b excludes words that are clearly not phenomena from the determination target.

The calculation result of IEPH in step S256 will be described with reference to FIG. Excluding “whiskers” already determined as the final stress, the IEPH value of “signal failure” is maximized. Therefore, the part / phenomenon determination unit 25b determines “signal failure” as the final failure mode.

In step S257, the part / phenomenon determination unit 25b calculates an IEPA value for each of the failure mode candidate and the failure part candidate.

In step S258, the part / phenomenon determination unit 25b determines the word having the maximum IEPA value as the final failed part. That is, among the candidate elements on the result side, the most likely part is determined as a “part”, and that part is set as the “failed part”. The part / phenomenon determination unit 25b excludes words that are clearly not parts from the determination target.
Thereafter, the part / phenomenon determination processing procedure is terminated (proceeds to step S207).

The calculation result of IEPA in step S258 will be described with reference to FIG. When “signal failure” including “failure” at the end is excluded, the IEPA value of “electronic system” is maximized. Therefore, the part / phenomenon determination unit 25b determines “electronic system” as the final failed part.

(Comparison with conventional technology)
As a conventional technique for extracting a specific word from a sentence, a method using a “clue word” is known. For example, suppose there is a sentence that “there is a crack due to repeated stress in a gear”. In this case, “gear” described immediately before “in” is extracted as a faulty part, “repetitive stress” described immediately before “for” is extracted as stress, The “crack” described immediately before can be extracted as a failure mode. Here, “in”, “because”, and “reached” are clues about the failed part, stress, and failure mode, respectively.

However, this conventional technique is easily affected by fluctuations in expression. For example, it is assumed that there is a sentence having the same meaning, “a crack occurs in a gear and the cause is considered to be repeated stress”. The sentence does not have clue words such as “in”, “for”, and “reached”, and as a result, a word such as a faulty part cannot be extracted.

Also in this embodiment, clue words are used when obtaining the above-described PHDF and PALDF. However, the prior art simply searches for a sentence using a clue word as a search key. On the other hand, IEPH and IEPA of the present embodiment are the results of statistical and quantitative arithmetic processing. Therefore, the IEPH and IEPA of the present embodiment can appropriately determine the probability of phenomenon and the likelihood of parts without being directly affected by fluctuations in the expression of individual sentences.

(Modification 1 of the first embodiment)
In the above description, the example in which the defect information registration unit 23 receives input of defect information from the user has been described. However, it is more convenient if failure information published on the network can be collected without waiting for user input. Therefore, the public defect information collection unit 24 accesses a server of an arbitrary company or the like via the network 4 and acquires defect information. And the acquired defect information is memorize | stored in defect information DB31.

(Modification 2 of the first embodiment)
When the user is viewing the causal model as a search result on the FMEA table creation screen 53 (FIG. 8), it is more convenient if the inconvenience information from which the causal model is extracted can also be viewed at the same time. Therefore, immediately after step S211 or S212, the causal model registration / editing unit 26 creates a model / inconvenience relationship list in which the causal model and the inconvenience information from which the causal model is extracted are stored in association with each other, and stores them in the auxiliary storage device 15. Store. When the causal model registration / editing unit 26 accepts that the user selects a certain causal model row on the FMEA table creation screen 53, for example, the causal model registration / editing unit 26 refers to the model / inconvenience relation list and identifies corresponding inconvenience information. Then, the specified inconvenience information is acquired from the auxiliary storage device 15 and displayed (output) on the output device 13.

(Second Embodiment)
In the first embodiment, there is a premise that the user of the defect information utilization support apparatus 2 is a parts manufacturer and a device manufacturer. In the second embodiment, it is assumed that there are separate parts manufacturers, equipment manufacturers, and service providers. In FIG. 1, the service provider operates the defect information utilization support apparatus 2. The parts manufacturer operates the terminal device 3. The device manufacturer also operates the (other) terminal device 3.

(Division of roles)
The parts manufacturer is a business partner (cooperator) of the service provider, and provides many examples of defect information to the service provider. The service provider pays the price to the parts manufacturer. The device manufacturer is a customer of the service provider. The service provider causes the device manufacturer to use the defect information utilization support apparatus 2 and receives the price from the device manufacturer.

More specifically, the component manufacturer accesses the defect information utilization support device 2 from the terminal device 3 via the network 4. The component manufacturer provides defect information (corresponding to steps S201 to S204). The service provider extracts the causal model from the defect information, and edits and registers the causal model (corresponding to steps S205 to S212). Of course, the part manufacturer may perform the processing. The device manufacturer creates an FMEA table (corresponding to steps S221 to S227).

The account management unit 27 creates an account list for storing the access authority to each function of the defect information utilization support device 2 and the settlement account for each component manufacturer and device manufacturer, and stores them in the auxiliary storage device 15. The payment / billing management unit 28 creates and stores in the auxiliary storage device 15 a payment list that stores the number and unit price of the provided defect information and the number and unit price of the extracted causal models for each component manufacturer. Furthermore, the payment / billing management unit 28 creates a billing list for storing the number and unit price of the created FMEA table or the access time (usage time) and unit price to the defect information utilization support device 2 for each device manufacturer. And stored in the auxiliary storage device 15.

When the account management unit 27 receives an access request from the component manufacturer or device manufacturer, the account management unit 27 refers to the account list and permits access to a predetermined process. The payment / billing management unit 28 constantly monitors the processing performed by the component manufacturer and the device manufacturer, and maintains the payment list and the billing list in the latest state according to the processing contents.
At a predetermined cycle such as once a month, the payment / billing management unit 28 calculates a payment amount for each component manufacturer based on the payment list, and calculates a charging amount for each device manufacturer based on the charging list. The account management unit 27 deducts the payment amount from the service provider's payment account, and instructs the component manufacturer's payment account to be charged, and the charge amount is deducted from the device manufacturer's payment account, and stores it in the service provider's payment account. Send an instruction to deposit money to the financial institution server.

(Effect of embodiment)
The defect information utilization support device 2 of the present embodiment has the following effects.
(1) The user can efficiently generate a causal model indicating what kind of failure occurs in which component as a result of the failure when what stress is applied to which component as the cause of the failure. Can be created.
(2) The user can accurately search the causal model when creating the FMEA table.
(3) The user can use the name of a faulty part included in an existing document or the like as a search key for searching for a causal model.
(4) The user can use the disclosed defect information.
(5) The user can purchase defect information from another person.
(6) The user can cause another person to use the defect information utilization support apparatus for a fee.
(7) The user can determine a part or phenomenon from the candidate elements with high accuracy.

(8) The user can edit the causal model created by the defect information utilization support apparatus.
(9) The user can know the defect information that is the basis of the causal model.
(10) The user can use the inconvenience report as inconvenience information as it is.
(11) The user can use the synthesized word as an element of the causal model.
(12) The user can determine the part of speech of the word included in the inconvenience information with high accuracy.

In addition, this invention is not limited to an above-described Example, Various modifications are included. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. Further, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment. Further, it is possible to add, delete, and replace other configurations for a part of the configuration of each embodiment.

Each of the above-described configurations, functions, processing units, processing means, and the like may be realized by hardware by designing a part or all of them with, for example, an integrated circuit. Each of the above-described configurations, functions, and the like may be realized by software by interpreting and executing a program that realizes each function by the processor. Information such as programs, tables, and files that realize each function can be stored in a recording device such as a memory, a hard disk, or an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.
In addition, the control lines and information lines are those that are considered necessary for the explanation, and not all the control lines and information lines on the product are necessarily shown. In practice, it may be considered that almost all the components are connected to each other.

DESCRIPTION OF SYMBOLS 1 Defect information utilization support system 2 Defect information utilization support apparatus 3 Terminal apparatus 11 Central control apparatus (control part)
12 Input device 13 Output device 14 Main storage device (storage unit)
15 Auxiliary storage device (storage unit)
DESCRIPTION OF SYMBOLS 21 FMEA table preparation part 22 Causal model search part 23 Defect information registration part 24 Public defect information collection part 25 Causal model extraction part 25a Causal model candidate extraction part 25b Parts / phenomenon discrimination part 26 Causal model registration / edit part 27 Account management part 28 Payment / Billing Management Department 31 Defect Information Database 32 Causal Model Database

Claims

A storage unit that stores the defect information in which the defect of the device is described in sentences for each item, and a causal model having a plurality of words included in the defect information;
In accordance with the item to which the word belongs and the part of speech of the word, the cause-side element candidate of the causal model and the result-side element candidate of the causal model are extracted from the plurality of words included in the defect information. And
Among the extracted cause-side element candidates, those having an expression tendency as a part and those having an expression tendency as a phenomenon are determined as a failure factor part and a stress, respectively.
Among the extracted result side element candidates, those having an expression tendency as a part and those having an expression tendency as a phenomenon are determined as a failure part and a failure mode, respectively.
A controller that stores the causal model having the determined failure factor component, stress, failure component, and failure mode in the storage unit;
A defect information utilization support device characterized by comprising:
The controller is
Accept input of the element,
Obtaining a causal model including the received element from the storage unit;
Outputting the acquired causal model,
Accepting input of additional information for the output causal model,
Storing the received additional information in the storage unit in association with the output causal model;
The defect information utilization support apparatus according to claim 1, wherein:
The controller is
When receiving input of the element,
Accessing a pre-designated document or data and obtaining the elements contained in the document or data;
The defect information utilization support apparatus according to claim 2, wherein:
The controller is
Obtaining the disclosed defect information via a network;
The defect information utilization support apparatus according to claim 3, wherein:
The controller is
When acquiring the defect information via a network, calculating an amount to be paid to a person providing the defect information according to the number of the acquired defect information;
The defect information utilization support apparatus according to claim 4, wherein:
The controller is
When allowing any person to use the defect information utilization support apparatus via a network, calculating the amount to be charged to the arbitrary person according to the usage time;
The defect information utilization support apparatus according to claim 5, wherein:
The controller is
Calculating the number of times that the candidate for the element appears immediately after a predetermined word as the expression tendency;
The defect information utilization support apparatus according to claim 6.
The controller is
The memorized causal model is displayed on the screen,
Accepting edits to the causal model displayed on the screen;
The defect information utilization support apparatus according to claim 7, wherein:
The controller is
In association with the defect information, the causal model created from the defect information is stored in the storage unit,
Accept that any one of the output causal models is selected,
Obtaining the defect information corresponding to the selected causal model;
Outputting the acquired defect information;
The defect information utilization support apparatus according to claim 8, wherein:
Said item is
Including the name of the defect, the process up to the occurrence of the defect, the phenomenon of the defect, the cause of the defect, and the countermeasures for the defect,
The defect information utilization support apparatus according to claim 9.
The part of speech is
A noun,
The word is
Contain a compound word consisting of multiple consecutive words,
The defect information utilization support apparatus of Claim 10 characterized by these.
The controller is
Determining the part of speech by performing a morphological analysis;
The defect information utilization support apparatus according to claim 11, wherein:
A defect information utilization support device that is connectable to an external storage device that stores defect information in which device defects are described in sentences for each item, and a causal model having a plurality of words included in the defect information. There,
In accordance with the item to which the word belongs and the part of speech of the word, the cause-side element candidate of the causal model and the result-side element candidate of the causal model are extracted from the plurality of words included in the defect information. And
Among the extracted cause-side element candidates, those having an expression tendency as a part and those having an expression tendency as a phenomenon are determined as a failure factor part and a stress, respectively.
Among the extracted result side element candidates, those having an expression tendency as a part and those having an expression tendency as a phenomenon are determined as a failure part and a failure mode, respectively.
A controller that stores the causal model having the determined failure factor component, stress, failure component, and failure mode in the external storage device;
Defect information utilization support device characterized by.
The storage unit of the defect information utilization support device
The defect information in which the defect of the device is described in a sentence for each item, and the causal model having a plurality of words included in the defect information are stored,
The control unit of the defect information utilization support device,
In accordance with the item to which the word belongs and the part of speech of the word, the cause-side element candidate of the causal model and the result-side element candidate of the causal model are extracted from the plurality of words included in the defect information. And
Among the extracted cause-side element candidates, those having an expression tendency as a part and those having an expression tendency as a phenomenon are determined as a failure factor part and a stress, respectively.
Among the extracted result side element candidates, those having an expression tendency as a part and those having an expression tendency as a phenomenon are determined as a failure part and a failure mode, respectively.
Storing the causal model having the determined failure factor component, stress, failure component, and failure mode in the storage unit;
The defect information utilization support method of the defect information utilization support apparatus characterized by this.