JP4923638B2 - Knowledge and know-how systematized database construction device and method - Google Patents

Description

The present invention relates to a system and database construction apparatus and method for knowledge and know-how in a production system.

Conventionally, problems that occur during production in production systems, especially processing and assembly type production systems, are identified early in the design, production preparation (preparation) and production technology stages, and the problems that occur during production are minimized. It is hoped that There is also a need to improve loss costs, delays in market launch, and product brand image down due to defects after market launch. The main causes of problems are, for example: • Adjustment with the manufacturing site is not possible at the time of design • Problems are not sufficiently identified at the time of design.
・ The process capability at the time of design does not match the actual situation.
・ In the case of trial production, problems / problems during mass production have not been fully studied.

  In order to solve these problems, total management technology and support systems for knowledge and know-how in production systems are required. However, the existing technology and support system do not provide an environment that supports the former in an integrated manner, and most systems perform only a partial function in quality control of production systems.

  Regarding total management of knowledge and know-how in production systems, for example, the current production system process is roughly divided into a design stage and a production stage, and the design stage is further divided into a planning / design stage and a prototype / process design stage. The manufacturing stage is further divided into an installation / adjustment stage, a production preparation stage, and a mass production stage. In general, the company has knowledge and know-how such as design know-how in design and factory production know-how and defect improvement know-how in manufacturing. In order to solve the above-mentioned problems, it is conceivable to approach to accumulate defect improvement know-how, feedback to the design stage, and further to accumulate design / prototype / factory production know-how and promote reuse.

Here, FIG. 26 shows a flow of information in design, production preparation, and mass production.
In the figure, at the product design stage, the design of the function, part configuration, shape, material, reliability, etc. of an arbitrary product is performed based on the market information, product planning information, etc. stored in the database 401. The design result is stored in the database 402. In the illustrated example, for example, product drawings, materials, product model data, E-BOM, design FMEA (design process reliability design), and the like are stored.

  Here, the BOM (Bill of Material) of E-BOM refers to a bill of materials, and the bill of materials presents a list of parts constituting one product as a list. E-BOM (Engineering BOM) is a BOM used for design, showing what kind of parts the product is composed of, and describing the functions, specifications and shapes of each part in detail. . There are other M-BOMs (Manufacturing BOMs), which will be described later, in order to make E-BOM information available in actual production activities such as product planning and scheduling. Knowledge of how to manufacture is described.

  Referring again to FIG. 26, subsequently, in the production preparation stage, design result information stored in the database 402 and various information stored in advance in the database 403, that is, equipment capacity (C / T) and process capacity (C / T). Based on information such as workability (inspection, tool change), tool capability, equipment / tool cost, various constraints, etc., process / equipment / tool / tool design is performed. Process / Equipment / Jig / Tool Design: Manufacturing configuration / Flow design, M-BOM creation, Equipment / Jig tool selection, Layout design, Work drawing (Processing part / Standard / Dimension / Method, Tool path) creation, Manufacturing Process reliability design (process FMEA) or the like. This result is stored in the database 404.

  In the example shown in the figure, the database 404 includes equipment / tool / tool specifications, equipment / tool / tool layout, work drawing, layout diagram, program / parameter, manufacturing condition / process flow / work procedure / inventory standard, M -BOM, process FMEA, etc. are stored.

In the mass production manufacturing stage, product manufacturing is performed using information stored in the database 404.
If a problem occurs in the production preparation stage or mass production stage, it is fed back to the upstream stage (design or production preparation). In addition, when a defect occurs in the market after the product is introduced to the market, defect information in the market (failure location and its phenomenon (for example, abnormal vibration)) is fed back to the design / production site.

FIG. 27 shows a business flow of such defect handling processing. That is, the business flow of the trouble recurrence prevention work is shown.
Problems in the production preparation stage / mass production stage or defects in the market are fed back in the form of a defect notification form.

In the business department of defect handling processing, etc., first, the phenomenon is confirmed based on the defect notification form. This is, for example,
・ Identify the primary cause ・ Determine the importance (decide whether to respond immediately or only later)
・ Determining the priority, checking the frequency of occurrence, and determining the destination.

  Next, a measure corresponding to the decision item in the above-described phenomenon confirmation is performed, and subsequently, the cause is investigated (the true cause is examined). In addition, impact analysis is performed. This examines the range of affected parts and processes, grasps the degree of influence, and determines the destination.

Then, consider countermeasures. This includes planning alternatives (primary countermeasures, permanent countermeasures), estimating effects, and determining the execution order.
After executing the above work, implement countermeasures. This is done by selecting a plan to be executed from among alternatives, making an implementation decision, and notifying the execution department. If the executing department is in the product design stage, review the specifications, etc., if it is a process design, review the standards / standards, work procedures, processing methods, etc. Review the environment and layout, etc. If mass production, review the countermeasures for pokayoke.

  Then, the execution department feeds back the execution result and the execution result (execution confirmation). In response to this, the processing results and measures to prevent recurrence are reported to the complaint issuing department (quality answer).

Here, there are the following two problems related to the above-described trouble recurrence prevention work.
1. At the time of preliminary examination, examination in product design, production technology, and manufacturing is insufficient / leaked / delayed.
2. The review of specifications, standards and standards is delayed or leaked.

The following problems can be considered as causes of the above problems.
(1) Problems with quantity and quality of know-how ・ There is insufficient knowledge and know-how accumulated about defects.

・ There is too much knowledge and know-how accumulated about defects, so useful information cannot be extracted.
・ It is difficult to use because information varies greatly between individuals.
(2) Problems in utilizing knowledge and know-how ・ It is difficult to pursue the root cause / understand the range of influence.

・ Past measures are not being used.
・ The review of specifications, standards and standards is not timely.
・ Knowledge and know-how have been used, but it has not been checked whether countermeasures have been implemented.
(3) Sharing problems ・ Similar failures at other factories and other lines are not shared.

・ If you are an experienced person, you can find out where other similar defects occur.
Here, consider (1) dealing with the problem of quantity and quality of know-how.

In order to solve this problem, it is desirable to realize the following mechanism.
-A mechanism that facilitates the registration of knowledge and know-how.
・ A mechanism that eliminates individual differences in expression and description of knowledge and know-how.
-A mechanism for easily extracting necessary items from a large amount of knowledge and know-how.

  Here, for example, in the invention described in Patent Document 1, conventionally, the knowledge and data of each department (design department, production department) are centrally managed, and a database is made public to other departments. In response to the need for a system that utilizes the data of the company to improve the quality of operations in its own department, the product database that records product information, which is information about the product, and the content of troubles that occur in the plant formed from the product, It has a trouble database that records a plurality of inspection methods that should be inspected after a trouble occurs. Based on the above product information, it is estimated which product is causing the trouble, and in order from that product. Create an inspection procedure to inspect. At that time, if there is a difference between the design value and the actual measurement value, it is generally considered that a trouble is often generated.

  In the above-mentioned conventional technology, there is an idea to centrally manage the knowledge and data of each department (design department, production department), but (1) to a mechanism sufficient to deal with the problem of quantity and quality of know-how Not considered.

Next, an existing method related to the prevention of defects will be described.
First, “design trouble prevention method SSM (Stress Strength Model)” by Tamura et al. Will be described. This method is intended for design for preventing defects in the manufacturing industry, and the mechanism of the causal chain of defects is represented by a connected structure in units of models generalized by conventional SSM. Then, from general-purpose defect knowledge accumulated in the SSM format, knowledge that matches the design specifications is extracted at the design stage using the hierarchical relationship between the defect concepts, and the knowledge is used to prevent design troubles. (For example, see Non-Patent Documents 1, 2, and 3).

  In addition, FTA is targeted for analysis of causes of failure events in systems (including human factors and software factors), with failure phenomena (events) as top events, and intermediate events that cause failures as logic gates. In this way, the reliability is graphically analyzed using a fault tree that is sequentially traced from top to bottom and expanded to all the causal events. In this analysis method, the failure tree is used to specify the occurrence path of the top event and the fatal event, so that the reliability weakness of the related part can be pointed out and the countermeasures can be recommended. A clear understanding of what the weaknesses are and how they can lead to failures.

  However, it is assumed that the event is in a normal / abnormal state, and cannot be analyzed if it is in an intermediate or transitional state. Further, intermediate / causal events unrelated to the top event cannot be analyzed (see, for example, Non-Patent Document 4).

  FMEA is designed to prevent malfunctions in the system, analyze the effects of failure events, and recommend target proposals (single hardware failure, process design is the main target), and the functional connection relationships of system components. And systematically study how the lower-level failure mode affects the upper level in a bottom-up manner while creating a worksheet (table) that takes into account failure grades and countermeasures. This is to evaluate the reliability incorporated. In this technique, the failure priority of failure modes is ranked based on the frequency of occurrence, the degree of influence, and the degree of detection difficulty, so that the weaknesses of the design can be ranked and understood, and a review of the design and proposal of countermeasures are recommended. it can. In addition, it is possible to examine the failure of all the configuration items (events).

However, the failure (failure) of the complex factor cannot be logically analyzed using Boolean algebra (suitable for the influence analysis of a single failure) (see Non-Patent Document 4).
Failure Mandala (adopted in the failure knowledge database led by Hatamura of the Japan Science and Technology Agency) covers all failure cases involving humans and a general analysis of failures centered on human errors, and all failure cases are human. Based on the recognition that it is an error, the context of failure is structured as a flow of human causes → human behavior → results. A major feature is that each phase of "cause", "behavior", and "result" is muffled, and the mandales are classified into layers and systematized / standardized. In this technique, similar cases can be searched by using patterns and keywords of causes, actions, results, and lessons can be obtained.

  However, the outline of the cause and effect of failure can be expressed, but the cause and effect mechanism cannot be expressed precisely. In addition, it is impossible to predict a newly assumed failure case by combining each failure case. In this technique, failure cases are systematized using hierarchical classification defined by each cause, action, and result, so failure (especially failure due to human factors) is systematized and standardized. (Refer nonpatent literature 5).

  The staged reliability design support system (Koga-Aoyama model) is targeted for reliability evaluation in staged design (top-down design). Product models are "substance", "attribute", and "interface". In each stage of top-down design, which is expressed by three elements and features step-by-step refinement, the possible failure groups correspond to each stage of the design model from the product model information and hazard information assumed when using the product. By extracting and using a separately created hierarchical failure database (hierarchical failure DB) having a hierarchy, it supports the reliability evaluation of FMEA, FTA, and the like. In the hierarchical failure DB, past failure cases are collected widely and converted into failure units that are added to the SSM with the expression of hazards that are potential hazards and the expression of the mechanism of stress propagation through the interface between entities. Accumulate.

  In this technique, it is possible to extract failure groups assumed in the design model at each stage of the design, and using them, FMEA provides a list of failure modes and their effects, and FTA provides a primary plan for the development of top events. By providing it to the designer, comprehensiveness can be examined from the viewpoint of past cases in reliability evaluation. In addition, reliability can be improved sequentially from the upstream stage of design.

  However, although the product model is configured based on E-BOM, it is irrelevant to M-BOM, and therefore, it is not possible to analyze defects caused by production technology / manufacturing information (see Non-Patent Documents 6 and 7).

 Here, M-BOM and E-BOM will be described. First, BOM (Bill of Material) is a bill of materials. The parts list is a list of parts constituting one product and is used in various departments such as planning, design, manufacturing, and sales.

  There are a summary type and a structure type in the bill of materials. The summary-type parts table displays only part names and quantities necessary for each final product in a form corresponding to the final product, and is used for parts development. The summary bill of materials is not very useful except for purchasers.

  The structure-type parts table displays the mutual structural relationship of each part from the material to the product, and it can be seen at a glance which parent part is composed of which child part. The structure type BOM can also be used in the design department and the manufacturing department.

E-BOM is a BOM used for design and is called Engineering BOM. It shows what kind of parts a product is composed of, and the functions, specifications and shapes of individual parts are described in detail. ing. M-BOM is a BOM used for manufacturing. It is called Manufacturing BOM and is manufactured to make E-BOM information available in actual production activities such as manufacturing planning and scheduling. Knowledge of how to do is described.
JP 2003-44546 A Yasuhiko Tamura and two others, “Development of a system to utilize structured knowledge about defects” Hideo Saito and 5 others, “Practical use of structured knowledge about defects” Yasuhiko Tamura, “Practice of design trouble prevention system by using structured intrinsic knowledge”, Justsystem Knowledge Management Forum 2003, Junjiro Suzuki et al. “FMEA / FTA Implementation Method” published by Nihon Kagakuren, 1982 Director Yotaro Hatamura, “Structure and Expression of Failure Knowledge Database” (Explanation of “Failure Mandara”), December 2002, Science and Technology Promotion Agency, Failure Knowledge Database Development Project Kaoru Koga and two others, “Proposal and Construction of Failure Information Management System that Supports Reliability Design in Steps”, December 6, 2001, 10th Annual Conference of the Japan Society of Mechanical Engineers, TRANSLOG 2001 Satoshi Koga and 3 others, “Construction of Product Reliability Management System in Top-down Oriented Information Model”, November 29, 2002, The 12th Design Engineering and System Division Lecture Meeting of the Japan Society of Mechanical Engineers

An object of the present invention is to facilitate the registration of knowledge and know-how, eliminating the individual differences of expression, description of knowledge and know-how, organized database building apparatus of knowledge and know-how can be extracted easily any needed registered knowledge and expertise And providing a method .

This onset Ming, in organized database building apparatus of knowledge and know-how to systematically accumulated knowledge and know-how for all to manufacturing process from design in the manufacturing system, information representation model in product structure based with respect to the knowledge and expertise And process structure-based information expression models were integrated to model product design defects, defect preparation in production preparation / manufacturing, and defect propagation mechanisms based on the causal chain of defects in product manufacturing processes. As a data structure for describing information, an entity code uniquely assigned to an entity such as a unit or a part, a status code uniquely assigned to various states, and a state before transition to the state indicated by the state code Uniquely assigned pre-status code and one entity manufacturing process State transition order number consisting of serial indicating the state transition order in the state, state transition code indicating the state transition condition (type), and item (state) used (input) in each step of the actual manufacturing process Input item codes that are uniquely assigned, output item codes that are uniquely assigned to items (states) that are manufactured (output) in each step of the actual manufacturing process, and parts that make up the entity such as blocks and units A means for constructing a product process database by storing each of the data in the data structure, and obtaining an information expression model of the products and processes accumulated in the constructed product process database On the basis of the configuration flag in the acquired information representation model. Product configuration information about entities such as a network, means for hierarchically extracting each entity and elements (sub-entities) constituting each entity, and corresponding to each extracted entity (including sub-entities), Means for acquiring an entity data model template defining various target events related to the entity as attributes from a storage means created and stored in advance, the extracted product configuration information, and each acquired entity data model template And a means for creating a data model corresponding to a product model, and various states related to each state corresponding to each state of each entity based on the state code in the acquired information expression model. A means for acquiring a state data model template in which a target event is defined as an attribute from a storage means that has been created and stored in advance, and the created product model pair Product process by sequentially adding the acquired state data model template in accordance with the order of the state transition order in the acquired information representation model for each entity data model template in the response data model Means for creating and storing a framework of a data model corresponding to the model, and attributes corresponding to each attribute developed in the framework of the data model by reading out the stored data model framework corresponding to the product / process model And a method for performing various determinations using the attribute value and means for constructing a quality knowledge database by describing hazard propagation .

  In the conventional stepwise reliability design support system (Koga-Aoyama model), as described above, the product model is configured based on E-BOM, but is not related to M-BOM.・ Failure caused by manufacturing information cannot be analyzed.

On the other hand, in the system for constructing a knowledge and know-how systematic database according to the present invention, an information expression model based on a product structure (E-BOM base) and an information expression model based on a process structure (M-BOM base) are used. As a data structure for describing information expressing the failure propagation mechanism based on the causal chain relationship of defects in the product manufacturing process, modeling of product design defects, manufacturing preparation / manufacturing defects, and product manufacturing processes, An entity code uniquely assigned to an entity such as a unit or a part, a status code uniquely assigned to various states, a previous state code uniquely assigned to a state before transition to the state indicated by the state code, State transition sequence number consisting of serial indicating the state transition sequence in the manufacturing process of a single entity A state transition code indicating a state transition condition (type), an input item code uniquely assigned to an item (state) used (input) in each step of the entity manufacturing process, and an entity manufacturing process It has an output item code that is uniquely assigned to an item (state) that is manufactured (output) in each process, and a configuration flag that indicates a part that constitutes an entity such as a block or a unit, and stores the data structure as the data structure. Since the product process database is constructed , such problems can be solved. Further, an information representation model of products and processes is acquired from the constructed product process database, and product configuration information about each entity, each entity, and each entity are configured based on the configuration flag in the acquired information representation model. Extracting elements (sub-entities) hierarchically, corresponding to each extracted entity (including sub-entities), create in advance a entity data model template that defines various target events related to each entity as attributes. A data model corresponding to a product model is acquired using the extracted product configuration information and each acquired entity data model template, and acquired from the storage means stored, and the acquired information representation model Corresponding to each state of each entity based on the state code in A state data model template defined as sex is acquired from storage means that has been created and stored in advance, and for each entity data model template in the created data model corresponding to the product model, in the acquired information representation model By sequentially adding the acquired state data model templates according to the order of the state transition order, a framework of the data model corresponding to the product / process model is created and stored, and the stored product / process model compatible Read out the data model framework and build the quality knowledge database by describing the attribute values corresponding to each attribute developed in the data model framework, various judgment methods using the attribute values, and hazard propagation can do. In other words, it is possible to construct a quality knowledge database that systematically describes and accumulates knowledge and know-how about everything from design to manufacturing processes in production systems. This facilitates, for example, the accumulation and sharing of defect improvement know-how.

According to organize database construction apparatus and method of the knowledge and know-how of the present invention, a mechanism to facilitate the registration of knowledge see and know, a mechanism to eliminate individual differences in expression, description of knowledge and know-how, a large amount of knowledge has been registered, It is possible to provide a mechanism for easily extracting necessary items from know-how.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a functional block diagram of the knowledge / knowhow systematization apparatus 10 according to this example.
The knowledge / knowhow systematization apparatus 10 according to this example comprises a systematization / accumulation function unit 11, a registration function unit 12, and a reference function unit 13, and the data accumulated in the systematization / accumulation function unit 11 is an coverage table DB. (Database) 21, product / process model DB (database) (PPMDB) 22, quality knowledge DB (database) (QA2-KDB) 23, and knowledge data DB (database) 24 are stored in each database.

PPMDB is an abbreviation for Product and Process Model Data Base, and QA2-KDB is an abbreviation for Quality Analysis / Question Answering Knowledge Data Base.
Although not shown in particular, the systematization / accumulation function unit 11 creates a database of an integrated model obtained by integrating an information expression model based on a product structure and an information expression model based on a process structure, and a product / process model DB (database) (PPMDB) ) The integrated model storage function unit stored in 22 and, based on this integrated model, at least various target events related to each entity or state prepared in advance for each entity and state indicated by the integrated model It has an entity / state data model creation function unit that creates a framework of an entity / state data model using a template defined as an attribute, and an attribute corresponding to each attribute based on the framework of this entity / state data model Data model that describes values, methods for making various decisions using the attribute values, and hazard propagation So made, it is stored in quality knowledge DB (database) (QA2-KDB) 23. The attributes and attribute values will be described later.

  The registration function unit 12 causes registration of attribute values and the like to the knowledge data DB 24 and the quality knowledge DB 23 using the coverage table based on the design information examined in advance. As a result, the description contents of the attribute value are standardized, and the conventional problem of “information expression and description are difficult to use because of large individual differences” is solved. In addition, since registration can be performed by selecting an attribute value or the like from a hierarchically organized coverage table, the labor of registration can be reduced.

  The product / process model DB (PPMDB) 22 is a product structure information model that is an information expression model in the product structure base (E-BOM), and a process configuration information that is an information expression model in the process structure base (M-BOM). It is a database that stores an integrated model (product-process model (PPM)) that integrates models. In addition to modeling the product design model, this integrated model also models each state and state transition in production and manufacturing. And since this integrated model expresses a process by a state transition, it can also express the malfunction occurrence mechanism based on the causal chain relation of the malfunction in a production system.

The knowledge data DB 24 stores design specifications, standards / reference documents, knowledge on manufacturing methods, and knowledge files (raw data) on knowledge and know-how in production systems.
The quality knowledge DB (QA2-KDB) 23 is created based on the entity (unit / part) expressed by the integrated model, its configuration relationship, and the state / state transition (process), and further uses knowledge and know-how. It is an object-oriented knowledge database that stores multiple entity / state data models that describe attributes, methods, hazard propagation, interaction (that an entity applies stress to another entity via an interface), and so on. Further, as will be described in detail later, the basis (framework) of each entity / state data model stored in the quality knowledge DB (QA2-KDB) and the configuration relationship between the models are the product / process model DB (PPMDB) 22. It is automatically created based on the integrated model information stored in. And based on the automatically created framework, manually add and change attributes, enter attribute values based on the coverage table, enter arbitrary attribute values, select / set or create new methods, and propagate hazards and hazards Then, the object-oriented knowledge database is completed by describing the interaction and the like.

  The coverage table DB 21 stores various coverage tables in which parts, phenomena, results, causes, etc. related to defects are systematized as attributes and attribute values, and is used when inputting attribute values based on the coverage table. Is done. The coverage table is a table in which various target events are hierarchically and comprehensively organized.

  The reference function unit 13 can refer to data stored in the systematization / accumulation function unit 11 from each department (product design, process design, facility design, mass production), and can be used for each business. Can do.

FIG. 2 shows a system for processing knowledge and know-how according to this example and the processing flow of the entire system related to this.
In this figure, the knowledge / know-how systematization device 10 is as described in FIG. As shown in FIG. 2, using the data stored in the system 10 for knowledge and know-how, phenomenon confirmation, priority determination support, cause investigation support, impact analysis support, countermeasure plan consideration measure decision, implementation department notification However, since it is not a feature of the apparatus of this example, it is not particularly described here.

  The characteristics of the device 10 in this example are related to the knowledge / know-how systematization device 10 itself. As shown in FIG. 3, (1) the process can be expressed by state transition, and (2) the knowledge / know-how is systematized. (3) Accumulate and evolve knowledge and know-how. Furthermore, when knowledge and know-how are systematized, the description of attribute values and the like can be standardized and facilitated using an coverage table.

  In FIG. 3, the knowledge and know-how systematization apparatus 10 has various existing techniques, that is, if it fails, SSM (Stress-Strength Model), FTA (Fault Tree Analysis), FMEA (Failure Model and Effects Analysis). The Koga-Aoyama model is partly used to construct various databases. At that time, M-BOM (manufacturing / manufacturing information) is taken from the knowledge data DB 24 or the like and used for database construction (not shown, but E-BOM is also used naturally).

First, based on the idea of failure, a person comprehensively organizes failure modes, phenomena, causes, results, measures, etc., creates the above-mentioned coverage table, and stores it in the coverage table DB 21.
In addition, general-purpose expression of failure causal mechanism is realized based on SSM and Koga-Aoyama model. This expresses a cause-and-effect relationship of occurrence of a failure by SSM. In other words, the mechanism of the causal chain of defects is expressed by a connection structure with a unit model of a generalized SSM. From the general-purpose defect knowledge accumulated in the SSM format, knowledge that matches the design specifications is extracted at the design stage using the hierarchical relationship between the defect concepts. As a result, it is possible to extract knowledge of defects that may occur in the product at the time of product design, and support the implementation of FMEA and FTA.

  In the Koga-Aoyama model, first, an abstraction of a danger source that can cause a failure is expressed as a hazard. In addition, product structure information (E-BOM) is expressed by three elements: entity, attribute, and interface. The hazard acts on the substance as a load, expresses the occurrence of the failure mode by SSM, changes the product state due to the occurrence of the failure, and thinks that a new hazard is generated. We think that it acts on other entities via the interface (representation of hazard propagation) and causes new hazards on other entities.

  Further, based on FTA, a connection of a plurality of events can be expressed by a logic gate, and a failure due to a complex cause can be handled using the logic gate. Furthermore, the probabilistic analysis of the contribution degree of the causal event that causes the result event (top event) can be performed from the probabilistic description of the tree diagram.

Further, based on FMEA, the risk priority of the target event can be ranked based on the occurrence frequency, the influence degree, and the detection difficulty level.
And knowledge and know-how can be systematized by the attributes of the coverage table stored in the coverage table DB 21 and the entity / state data model stored in the quality knowledge DB 23. Further, knowledge and know-how can be accumulated and evolved by adding / updating attributes / attribute values of the entity / state data model stored in the quality knowledge DB 23.

  Here, using some of the above existing methods will be described in more detail below. Here, in order to solve the above conventional problems, “general expression of failure causal mechanism”, “analysis of multiple failures”, “systematization of failure (including human factors) / standardization”, “risk source ( To realize the functions of each item (evaluation item) of "abstract abstraction expression", "defect expression using product structure information" and "expression of defect using production technology (production technology) / manufacturing information". Think.

First, an outline of each evaluation item will be described.
“Generic expression of failure causal mechanism” realizes the following functions.
-Evaluate whether the target technology / technique can model knowledge / know-how information related to defects with a generalized expression that abstracts the causal mechanism of defects.

  ・ By utilizing knowledge and know-how related to defects that have been modeled and accumulated using general-purpose expressions, the cause of a specific phenomenon can be traced back correctly when performing cause analysis or impact analysis of a newly generated defect.

  ・ By abstracting / modeling the cause-and-effect mechanism, it is possible to search for similar events. In other processes for manufacturing a product similar to the own process or in other processes possessed by a manufacturing method similar to the own process. Knowledge and know-how can be used.

“Multiple failure analysis” realizes the following functions.
-Evaluate whether multiple failures (failures caused by multiple causes) can be analyzed using the target technology / technique.

  ・ Accumulated past knowledge and know-how information, even when there are multiple causes of defects, in the cause-and-effect relationship, in order for workers to identify possible causes when performing cause analysis and impact analysis (Including phenomenon and cause) can be expressed logically correctly.

“Systematization / standardization of defects” realizes the following functions.
-Evaluate whether the target technology / technique can standardize the systematic expression and description of defects.

  -Systematize and standardize the expression and description method when describing a failure phenomenon, so that anyone can describe the failure phenomenon equally even if there is a difference in the experience of operators or differences in the expression of words.

  ・ When entering search terms for cause analysis, systematize and standardize the search terms, so that anyone can search for the cause of failure even if there is a difference in the experience of the workers or differences in the expression of words. (Search) is possible.

“Abstract expression of hazards (hazards)” realizes the following functions.
-Evaluate whether the target technology / technique can abstract the hazard.
・ When workers perform cause analysis and impact analysis, in order to extract the failure mechanism, the event that seems to cause the failure is reduced to a hazard that is a potential hazard. The hazard acts on the entity as a load, and clarifies the failure mode occurrence mechanism. Furthermore, a new hazard that occurs due to a change in the product state due to the occurrence of a failure acts on another entity via the interface, thereby causing a new hazard to the other entity.

In this way, the failure propagation path can be modeled through hazards.
“Failure expression using product structure information” can realize the following functions.
-Evaluate whether the target technology / technique can express defects using product structure information.

  ・ When causal analysis / impact analysis / subject examination / knowledge know-how is accumulated, the cause of the defect is a design defect (standard, standard, function, material, tolerance, shape, dimension, accuracy, positional relationship, etc.) By associating knowledge and know-how (including defect information) in the design business process with function information corresponding to product structure information, when a defect occurs in “XX function”, the cause is “XX function” The “representation of defects using production technology (production technology) / manufacturing information” that can specify that the process is in the business process for designing can implement the following functions.

-Evaluate whether the target technology / technique can express defects using production technology / manufacturing information.
・ When performing cause analysis, impact analysis, countermeasure review, and know-how accumulation, the cause of defects is defects in production technology (standards, work standards, check standards, manufacturing methods, procedures, conditions, equipment, jigs / tools, environment, Layout, etc.) or manufacturing defects (pochamis, carelessness, non-compliance with standards, etc.), know-how and know-how (including defect information) in the production / manufacturing business process -By associating with manufacturing information, it is possible to identify in which part of the business process of production / manufacturing the cause of the defect.

  Regarding the application of each of the above existing methods to realize each of the above functions, first, the “design trouble prevention method SSM (Stress Strength Model)” by Tamura et al. Therefore, “general expression of failure causal mechanism” is possible, and defect knowledge that can occur in a product can be extracted at the time of product design. It can also support the implementation of FMEA and FTA (Fault Tree Analysis). However, the propagation of a plurality of defects cannot be expressed only by the existing SSM method.

In the FTA method, a failure due to multiple causes can be handled using a logic gate, so that "multiple failure analysis" is possible.
In the above-mentioned stepwise reliability design support system (Koga-Aoyama model), hazards act on the substance as a load, express failure mode occurrence by SSM, change the product state due to failure occurrence, and introduce new hazards I think that occurs. And since it further acts on another entity via the interface and causes a new hazard in the other entity, a “general expression of failure causal mechanism” is possible. In addition, since an abstraction of danger sources that can cause defects is expressed as a hazard, "abstract source expression of danger sources" is possible. In addition, product structure information (E-BOM) can be expressed by three attributes: entity, attribute, and interface, and product structure information can be used to express and analyze defects. Can be realized.

  However, as described above, in this existing method, the product model is configured based on E-BOM, but since it is irrelevant to M-BOM, it is possible to analyze defects caused by production / manufacturing information. Can not.

  On the other hand, in the method of this example, an integrated model obtained by integrating the information representation model based on the product structure (E-BOM) and the information representation model based on the process structure (M-BOM) is stored as a database. Can solve such problems. Details will be described later.

The requirements necessary for realizing the function of each evaluation item are listed below.
The requirements necessary for the realization of “general expression of failure causal mechanism” are “general expression of single event propagation mechanism” and “general expression of occurrence mechanism”. A requirement necessary for the realization of “multiple failure analysis” is “expression of multiple phenomenon relationship”. The requirements necessary to realize “systematic / standardization of defects” are “systematization (segmentation)”, “leveling of expression”, “systematization of concepts”, and “completeness”. Requirements necessary for the realization of the "abstract representation of the danger source", is the "abstraction of things elephant". A requirement necessary for the realization of “problem expression using product structure information” is “use of production technology / manufacturing information”.

FIG. 4 shows an overall input / output image.
The knowledge and know-how systematization device 10 integrates and manages design, production technology and manufacturing information, and systematizes, accumulates and evolves knowledge and know-how. The knowledge and know-how of the knowledge and know-how systematization device 10 The information used for storage of information is roughly classified into product design information, process design information, and production manufacturing information.

  Product design information includes structure / configuration information (E-BOM), drawings, link information of specifications such as materials, link information of standards / standards, design FMEA information, and the like. Furthermore, design knowledge / know-how information as knowledge / know-how in the human brain is manually input from, for example, an arbitrary information terminal.

  As process design information, various information such as process configuration / order (M-BOM), manufacturing method, manufacturing conditions, etc. stored in the database, link information for work procedures / inventory standards, equipment / tool / tool specifications Link information of equipment, link information of equipment / tool / tool layout drawings, link information of work drawings, layout drawings, process FMEA information, and the like. Furthermore, process design knowledge / know-how information as knowledge / know-how in the human brain is manually input from, for example, an arbitrary information terminal.

  As the production / manufacturing information, information such as manufacturing knowledge / know-how, poka-yoke measures, cause of defects, countermeasures, etc., which are knowledge / know-how in the human brain, is manually input from an arbitrary information terminal, for example.

  According to the knowledge and know-how systematization apparatus 10 of this example, knowledge and know-how can be accumulated and evolved systematically. In addition, necessary information can be easily browsed and shared. Furthermore, it is possible to check whether or not the knowledge and know-how to be used when making a design change using existing information is used by using a check target (concept division) and a check requirement described later.

Although not specifically described here, it can also be used for investigating the cause of a defect, grasping the influence of the defect, examining countermeasures for the defect, and the like.
The coverage table stored in the coverage table DB 21 will be described with reference to FIGS.

The coverage table is basically defined and created manually by humans.
In the example shown in FIG. 5, the types of the coverage table are, for example, the failure site coverage table 31, the failure mode coverage table 32, the failure phenomenon coverage table 33, the failure result coverage table 34, the failure cause coverage table 35, the failure countermeasure coverage table 36, etc. There is.

  The failure part coverage table 31 and the failure phenomenon coverage table 33 are defined based on data stored in, for example, a product / process model DB (database) (PPMDB) 22. The failure result coverage table 34, the failure cause coverage table 35, and the failure countermeasure coverage table 36 are defined based on, for example, the concept of failure failure.

Specific examples of these various coverage tables are shown in FIGS.
In these examples, the product is an automobile.
FIG. 6 is a specific example of the failure part coverage table 31.

  In the example of the failure part coverage table 31 shown in the figure, parts related to the trouble (here, various parts constituting the automobile), for example, a rim part, a hub part, a steering main shaft part, etc. are classified and arranged hierarchically. ing. For example, it is classified into a steering device, a braking device, a driving device, etc. as a large classification, and further, when a braking device is taken as an example, it is classified as a disc brake, drum brake, ABS device as a middle classification, and when a disc brake is taken as an example, it is classified as a small classification It is classified into disk rotor, retainer, pad, piston and caliper. Parts related to the above-described defects are described according to these hierarchical classifications. These descriptions of the rim portion, hub portion, steering main shaft portion, etc. are used when describing attribute values corresponding to the attribute “part” in the QA2-KDB entity / state data model 200 shown in FIG. Is done. Further, as will be described later, the entity / state data model 200 shown in FIG. 15 and the like exists for each part / unit, but the above-mentioned subclasses are parts, the middle classification is a unit combining these parts, and the large classification. Corresponds to a (superordinate) unit that combines these units. The entity / state data model 200 shown in FIG. 15 and the like corresponds to the part “handle”. Therefore, the attribute value corresponding to the attribute “part” is “rim part”, as shown in FIG. “Hub”, “Spoke” etc.

  The other coverage table described below is not a classification directly related to the configuration between the plurality of entity / state data models 200 as the failure site coverage table 31 described above, but relates to the failure mode, phenomenon, cause, result, and countermeasure. The attribute values are described hierarchically classified and organized.

FIG. 7 is a specific example of the failure mode coverage table 32.
Failure modes are roughly classified into human error bases shown in FIG. 7A and physical phenomenon bases shown in FIG. 7B.

  In the example shown in FIG. 7A, the failure mode attribute values related to the human error are described hierarchically as shown in the figure. The attribute values of the failure mode in this example are, for example, missing, wrong number of counts, wrong count, miss of danger, wrong position, and the like.

  In the example shown in FIG. 7B, the attribute values of the failure mode related to the physical phenomenon are classified and described as electrical / mechanical equipment failure, processing failure, assembly failure, etc., as shown. For example, the attribute values of failure modes classified as electrical system failures are open, short circuit, overheating, leakage, poor contact, and the like. For example, the attribute values of the failure mode classified as a machining failure are a machining surface chatter pattern, a workpiece scratch, excessive wear, and the like. In the example of the entity / state data model 200 shown in FIG. 15 and the like, attribute values “work surface chatter pattern”, “work scratch”, and the like are described for the attributes “failure mode 1” and “failure mode 2”.

FIG. 8 is a specific example of the failure phenomenon coverage table 33.
The example of the failure phenomenon coverage table 33 is also roughly divided into the human error base shown in FIG. 8A and the physical phenomenon base shown in FIG. And classified. In both cases, as shown in the drawing, various malfunction phenomena and the malfunction modes (attribute values) corresponding to the malfunction phenomena are described. For example, it can be seen from FIG. 8B that the failure phenomenon corresponding to the attribute value “machined surface chatter pattern” of the failure mode is “has a chatter pattern” and “makes an abnormal noise”. In the example of the entity / state data model 200 shown in FIG. 15 and the like, the attribute value of the attribute “phenomenon” related to the failure mode whose attribute value is “machined surface chatter pattern” is “chatter pattern, abnormal sound”. ing.

FIG. 9 is a specific example of the failure result coverage table 34.
As shown, the large-scale breakage, fracture and damage, electrical failures, system failure, various failure result of injury, such as (attribute value) have been described are hierarchically classified and organized. In the entity / state data model 200 shown in FIG. 15 and the like, these attribute values are described for the attribute “result”.

FIG. 10 is a specific example of the failure cause coverage table 35.
Similar to the failure mode coverage table 32 and the like, the example of the failure cause coverage table 35 is roughly divided into a human error base shown in FIG. 10 (a) and a physical phenomenon base shown in FIG. 10 (b). are categorized. As shown in the figure, various failure causes (attributes) such as insufficient environmental investigation, misrecognition, lack of knowledge, poor management, thermal fluctuation, vibration, arc, processing and manufacturing method failure, positioning accuracy failure, structural deficiency, inappropriate material, etc. Values) are described hierarchically classified and organized. Note that, as in the example based on the physical phenomenon shown in FIG. 10B, a layer structure having two layers and a layer structure having three layers may be mixed.

  Here, in the example of the entity / state data model 200 shown in FIG. 15 and the like, the attribute with the attribute name “cause of failure (cause)” does not exist, but for example, “work manufacturing method (defect)” shown on the left side of FIG. The attribute values of the failure cause coverage table 35 are described in association with attributes (hereinafter referred to as cause attributes) that can be input to FTA logic methods, such as “maximum handled weight (excess)”.

FIG. 11 is a specific example of the failure countermeasure coverage table 36.
The failure countermeasure coverage table 36 is roughly divided into the human error base shown in FIG. 11A and the physical phenomenon base shown in FIG. 11B, like the failure mode coverage table 32 and the like. are categorized. In both cases, as shown in the figure, specific countermeasures against failures such as integration, removal of variation, instruction / recording, labeling, operation detection / display, suspension bush spring change, welding, machining method correction, cutting condition correction, etc. Attribute values) are described hierarchically classified and organized. In the entity / state data model 200 shown in FIG. 15 and the like, these attribute values (processing method correction, detachable member quality correction, etc.) are described for the attribute “measure” as shown.

Next, the product / process model DB (database) (PPMDB) 22 will be described below.
FIG. 12 is a diagram for explaining an expression model of a failure-causal chain related to PPMDB22.

  Here, for the sake of easy understanding, the data stored in the PPMDB 22 will be described in the form of an expression model, but the configuration of the actually stored data is as shown in FIG. 13, for example.

  FIG. 12 shows an example of a unit expression model based on PPMDB22. In the failure-causal chain expression model 50 shown in the figure, the entity is indicated by a rectangle (box) and the state is indicated by an ellipse. The state transition is indicated by an arrow from the entity (or state) before the transition to the entity (or state) after the transition. This notation is the same in FIG. 14 described later.

  FIG. 12 shows a representation model 50 of an entity 51 (unit C), and product structure information (E-BOM) that the entity 51 (unit C) is composed of an entity 52 (component A) and an entity 53 (component B). Then, the parts are received 66 and become the state 61 (initial (preparation)), and then the entity 52 (part A) is assembled to become the state 62 (intermediate product C1). (Part B) is assembled into state 63 (intermediate product C2), and inspection of state 63 (intermediate product C2) results in state 64 (finished product), that is, entity 51 (unit C) is completed. Manufacturing process information (M-BOM) is expressed.

  The expression model 50 represents the product structure information and the manufacturing process information of the entity 51 (unit C) with the following description. Between the entity 51 (unit C) and the entity 52 (component A) and between the entity 51 (unit C) and the entity 53 (component B), arrows 71 and 72 indicate state transitions toward the entity 51 (unit C), respectively. Write to connect. This description expresses the product structure information of the entity 51 (unit C) that the entity 51 (unit C) is composed of the entity 52 (component A) and the entity 53 (component B).

In addition, a description that the entity 51 (unit C) is completed through manufacturing steps of state 61 (initial (preparation)), state 62 (intermediate product C1), state 63 (intermediate product C2), and state 64 (finished product). , An arrow 66 indicating state transition is input to state 61 (initial (preparation)), and a state transition between state 61 (initial (preparation)) and state 62 (intermediate product C1) toward state 62 (intermediate product C1) is performed. Connected by the arrow 67 shown, and the state 62 (intermediate product C1) and the state 63 (intermediate product C2) are connected by the arrow 68 toward the state 63, and the state 63 (intermediate product C2) and the state 64 (finished product) are connected. It is described so as to be connected by an arrow toward the state 64 (finished product). Further, description is made such that the arrow (state transition) 67 is associated with the arrow (state transition) 71 and the arrow (state transition) 68 is associated with the arrow (state transition) 72. Furthermore, the condition for performing the state transition 66 (state transition condition (in this case, the process)) is “part arrival”, the condition for performing the state transitions 67 and 68 is “assembly”, and the condition for performing the state transition 68 is “inspection”. Describe that it is. According to these descriptions, the parts are first received and are in the state 61 (initial (preparation)) state, and the entity 52 (part A) is first assembled to the state 62 (intermediate product C1), and the state 62 (intermediate product C1). ), The entity 53 (part B) is assembled into a state 63 (intermediate product C2). This state 63 (intermediate product C2) is inspected and the entity 51 (unit C) is completed. Manufacturing process information is expressed.

For the entity 52 (part A) and the entity 53 (component B), the manufacturing process information is described in the same manner as the entity 51 (unit C).
For the entity 52 (part A), state 81 (raw material), state 82 (intermediate product A1), state 83 (intermediate product A2), state 84 (intermediate product A3) and state 85 (finished product) are shown in the order of arrows. (State transition) Connected at 86 to 90, the condition for the arrow (state transition) 86 to be performed is “raw material arrival”, the condition for the arrow (state transition) 87 to be performed is “cutting”, and the arrow (state transition) 88 is performed It is described that the condition to be performed is “polishing”, the condition for performing the arrow (state transition) 89 is “surface treatment”, and the condition for performing the arrow (state transition) 90 is “inspection”. According to this description , raw material is received in state 81 , state 81 ( raw material) is cut to state 82 (intermediate product A1), and state 82 (intermediate product A1) is polished to state 83 (intermediate product A2). Then, the state 83 (intermediate product A2) is surface-treated to obtain a state 84 (intermediate product A3), and the state 84 (intermediate product A3) is inspected to obtain a state 85 (finished product), that is, the entity 52 (part A The manufacturing process information that is completed) is expressed.

Regarding the entity 53 (part B), state 91 (raw material), state 92 (intermediate product B1), state 93 (intermediate product B2), and state 94 (finished product) are sequentially indicated by arrows (state transitions) 96 to 99. The conditions under which the arrows (state transitions) 96 are connected are “raw material arrival”, the conditions under which the arrows (state transitions) 97 are performed are “forging”, the conditions under which the arrows (state transitions) 98 are performed are “heat treatment” and arrows (State transition) describes that the condition under which 99 is performed is “inspection”. This description, raw material stock against state 91, forged its state 91 (raw) in a state 92 (intermediate product B1), in its state 92 state 93 by heat treatment (intermediate product B1) (intermediate product B2), the state 93 (intermediate product B2) is inspected, and the state 94 (finished product), that is, the manufacturing process information that the entity 53 (part B) is completed is expressed.

FIG. 13 is a diagram showing a data structure when the representation model shown in FIG. 12 is stored in the memory or storage of the information processing apparatus.
The table (PPM information) 100 shown in FIG. 13 includes an entity code 101, a state code 102, a previous state code 103, a state transition order 104, a state transition code 105, an input item code 106, an output item code 107, and a configuration flag 108. An item is stored for each entity shown in FIG. The definition of each code above is as follows.

Entity code 101: Code uniquely assigned to an entity such as a unit or part Status code 102: Code uniquely assigned to various states Previous state code 103: Before transition to the state indicated by the state code 102 Status code of the state of the state State transition order: Serial number indicating the order of state transition in the manufacturing process until an entity such as one unit or part is completed. State transition code assigned in order from 1 ... Code indicating the condition (state transition condition (process)) under which the state transition is performed Input item code ... Used in each manufacturing process of each entity such as a unit or part (input) Code uniquely assigned to the item (state) to be output Output item code: code uniquely assigned to the item (state) to be manufactured (output) in each manufacturing process of each entity such as a unit or part ... The entity code 101 indicating sub-entities (parts) constituting the entity such as a unit
The table 100 stores information on each state (process) in the state transition (manufacturing process) for each entity of the part A (entity 52), the part B (entity 53), and the unit C (entity 51). ing. For part A, raw material (state 81), intermediate product 1 (state 82), intermediate product 2 (state 83), intermediate product 3 (state 84), and finished product (state 85) status code 102 are stored. The previous state code 103 is stored for other states except the raw material (state 81). From the state code 102 and the previous state code 103, the state change due to the state transition can be known.

Further, the state transition order 104 is set for each state of the part A, and the order of state transition until the part A is completed can be known based on the state transition order 104. Furthermore, the state transition code 105 for each status code 10 2 parts A are set. The state transition code 105 has a code indicating work (step) performed on condition indicated by the corresponding status code 10 2. An input item code 106 is set for the state transition code 105 other than the “raw material arrival” of the part A.

  The state indicated by the input item code 106 is a state that is a target of the work (process) indicated by the state transition code 105. An output item code 107 is set for each state transition code 105 of the part A. The state indicated by the output item code 107 is normally the same as the state indicated by the state code 101 when the corresponding input item code 106 is not set, and when the corresponding input item code 106 is set. Is a state created by the operation (process) indicated by the state transition code 105 in the state indicated by the input item code 106.

  Also for the part B, items from the entity code 101 to the output item code 107 are set based on the expression model 50 of FIG. Since the contents of items such as each code set in the table 100 of the part B have the same format as that of the part A, detailed description is omitted here.

  In the unit C (entity 51), necessary items are set in the table 100 based on the expression model 50 in FIG. In the case of unit C, the components A and B flags are further set in the configuration flag 108. In other words, the “component A” and “component B” are displayed in the configuration flags 108 of the two rows in which the unit C is set in the entity code 101 and the state transition code 105 is set to “assembly”. An entity code 101 is set. By such setting in the table 100, it is possible to know the product structure information of the unit C that “the unit C is composed of the parts A and B”.

  As described above, the PPM information defines the structure of the product and the structure, relationship, and sequence (sequence) of the process of manufacturing the entity (product or part). As a result of executing the process (state transition condition), the state transitions to the next state.

FIG. 14 is a diagram illustrating a representation image of a failure mechanism based on PPMDB22.
The representation model 110 shown in the figure is a description in which the hazard and its propagation path and direction are further added to the description in the representation model 50 shown in FIG. However, data corresponding to such an expression model is not stored, and is merely an image. The data actually stored in the PPMDB 22 is, for example, data shown in FIG. 13, and data corresponding to hazards and propagation of hazards as shown in FIG. 14 is not stored. The hazard and hazard propagation and interface as shown in FIG. 14 are described manually at the stage of creating the stored data of the quality knowledge DB 23 based on the PPMDB 22, and this will be described later. This is true not only for hazards and their propagation, but also for possible failure modes and interactions.

  Here, a hazard is a potential or tangible danger source that causes a malfunction of an entity. Examples of hazards include those related to the environment, resources, and human resources. Examples of environmental hazards include narrow working space, vibration, and high temperatures.Examples of resource hazards include inadequate work procedures and excessive tool wear. Mistakes in work, neglect of procedures, lack of sleep, etc.

FIG. 15 shows an example of the entity / state data model created based on the product / process model DB 22.
The entity / state data model 200 is created for each entity (unit, part).

  The illustrated entity / state data model 200 is roughly composed of an input unit 201, an output unit 202, and various determination units (methods) 203 (methods 203a to 203e). Or, when roughly classified from another viewpoint, the entity data model 204 corresponding to the completed form of each entity (unit, part) and each manufacturing process / process (each state) until each entity (unit, part) is completed And the state data model 205 corresponding to.

  The input unit 201 and the output unit 202 describe, for example, attributes such as stress, strength, failure mode, site, phenomenon, result, countermeasure, and the attribute values. As described above, attributes such as failure modes, parts, phenomena, results, countermeasures, and attribute values related to the cause attribute can be described with reference to the above-mentioned various coverage tables, and standardization of expression can be achieved. This solves the problem of “it is difficult to use because there are large individual differences in the expression and description of information”. Stress and strength are mainly used for input of the SSM logic, and express the occurrence condition of the failure mode.

  Moreover, not all description items of the input unit 201 and the output unit 202 are described based on the above-described coverage table. For example, know-how information on design and manufacturing may be described using a design document, a standard document, a standard document, and the like. Or, to the storage location of related documents with knowledge and know-how (such as absolute path for electronic files, shelf location for paper files, etc.) You can also link This makes it easy to refer to these related documents later as reference information. In the illustrated example, the related documents include design specifications, design know-how, equipment standards, jig standards, tool standards, manufacturing guidelines, material standards, and the like.

  For example, the notation “processing manufacturing method (defect)” can mean two pieces of information. One is information that the attribute “machining manufacturing method” is the attribute value “cutting” or the like (from other than the coverage table). The other attribute is information that the cause attribute is the attribute value “defective machining / manufacturing method” (when described from the coverage table).

The framework of the entity / state data model 200 is automatically created using a template (described later) created by making use of know-how such as various operations.
The determination unit (method) 203 is a method logic created by making use of analytical work know-how. Examples of the method include a method 203a based on the SSM logic and methods 203b to 203e based on the FTA logic, as well as other methods based on the FMEA logic, which will be described in detail later.

  The entity / state data model 200 generally represents how to hold various information, knowledge, and know-how in production activities. The entity / state data model 200 provides a way of holding information, knowledge, and know-how, and by storing various information, knowledge, and know-how in production activities as attributes, attribute values, or method logic of the model 200, information / It is possible to systematically accumulate and use knowledge and know-how in association. For example, looking at the inputs and outputs related to the method 203c, failure mode 1 “machined surface chatter pattern” occurs on the rim due to a defect in the processing manufacturing method and excess weight that can be handled, and the phenomenon when the failure occurs is the chatter pattern. It can be seen at a glance that there is abnormal noise, resulting in “destruction / damage”, and “processing method correction” as a countermeasure against defects. Furthermore, it can be seen at a glance that failure mode 1 “machined surface chatter pattern” that occurs during the manufacturing process of the part / unit affects the completed part / unit (hazard propagation). In the expression in the figure, this is indicated by an arrow that is an output from the method 203c and that is input to the method 203b. Similarly, the propagation of hazard from one state (manufacturing process) to the next state can also be described.

In the quality knowledge DB 23, a data model using a plurality of the entity / state data models 200 is stored. This will be described below with reference to FIG.
Here, in the drawings and the description after FIG. 16 below, the entity / state data model 200 does not show the attribute name, attribute value, and others (links etc.) in detail as in the example shown in FIG. It shall be shown in a simplified manner and only the main attribute names shall be shown.

  The example shown in FIG. 16 is based on the failure causal chain representation model stored in “Failure mechanism unit representation example using product / process model DB 22” 25. The basic (framework) and these relationships (structural relationships between entities, the order of the process configuration) are automatically created, and methods are created manually, attribute values are entered, and hazards and hazard propagation are described. The image which completes a data model is shown, and the state after completion is shown in quality knowledge DB23 on the right side of the figure of FIG. Each arrow (for example, an arrow heading from the attribute “(failure) mode” of the part A to the attribute “stress 3” of the part A) indicates the propagation of the hazard and the like, and these are also described manually. In particular, the propagation path between components is based on an “interface”, which is established by a person defining an “interface”. In FIG. 16, the stored contents of “unit expression example of failure mechanism using product / process model DB 22” 25 also show the image of hazard and hazard propagation as in FIG. 14. Is not actually stored in the product / process model DB 22, and is described by human beings when creating the data model of the quality knowledge DB 23.

  For example, an additional input / change operation as shown in FIG. 17 may be performed on the data model created as shown in FIG. 16 by manual judgment / input. That is, analysis logic (method) change, FMEA parameter change, failure (failure) mode may be added as an attribute, and an influence attribute value may be input. Of course, it is not limited to such an example. These additional input / change operations are performed with reference to, for example, various information stored in the database in advance (such as past failure causes / measures, process design knowledge / know-how, design FMEA information, manufacturing knowledge / know-how, etc.). Also good.

The basic (framework) of the entity / state data model 200 described above and a method for automatically creating these relationships will be described below with reference to FIGS.
Here, an example of data stored in the product / process model DB 22 will be described with reference to the example (PPM information 100) shown in FIG.

FIG. 18 is a flowchart of information creation processing in the quality knowledge DB 23 based on the PPMDB information 100.
In FIG. 18, first, the PPM information 100 is acquired from the product / process model DB 22 (step S11).

  Next, based on the acquired PPMDB information 100, first, product configuration information, each entity, and elements (sub-entities) constituting each entity are hierarchically extracted (block, unit, component, etc.) (configuration block 108) ( Step S12). In the examples of FIGS. 12 and 13, the sub-entities constituting the unit C are the part A and the part B. Although simple examples are shown in FIGS. 12 and 13, actually, there are higher layers, that is, blocks and the like, and the blocks are constituted by a plurality of units or parts.

  Subsequently, an entity data model template (template) corresponding to each extracted entity (including sub-entities) is acquired from a database (not shown) (step S13). That is, an entity data model template corresponding to each entity is created in advance and stored in a database (not shown). In the example of FIG. 13, the entity data model template corresponding to each of the part A, the part B, and the unit C is acquired from a database (not shown). 18 shows an example of the entity data model template 211 for unit C and the entity data model template 212 for component A. Although not shown, naturally, there is also an entity data model template for part B, which is acquired.

  Then, a data structure 214 corresponding to the product model is created using the product configuration information and the acquired various entity data model templates (step S14). FIG. 19A shows a data structure 214 corresponding to the product model corresponding to the example of FIG. As shown in FIG. 19A, the data structure 214 corresponding to the product model arranges various actual data model templates acquired in step S13, and product configuration information, that is, the unit C is composed of the parts A and B. Information (indicated by an arrow in the figure).

Furthermore, various state data model templates (templates) corresponding to the various states of each entity are created in advance and stored in a database (not shown). From this, from the PP status code 102 of M information 100 acquired in step S11, the status data model template 213 corresponding to each status code 102, it acquires from a database (not shown) (step S15). Then, the state data model template acquired in step S15 is sequentially added according to the order of the state transition order 104 to each entity data model template in the data structure 214 corresponding to the product model created in step S14. By doing so, the data structure 215 corresponding to the product / process model (product model + state data model correspondence) is created (step S16).

  An example of the data structure 215 corresponding to (product model + state data model) is shown in FIG. As shown in FIG. 19B, (product model + state data model) is created for each entity. That is, (product model + state data model) 221 corresponding to unit C, (product model + state data model) 222 corresponding to part A, and (product model + state data model) 223 corresponding to part B are created. .

  Here, as described above, the various templates (the entity data model template and the state data model template) are created in advance. The creation method will be briefly described.

The various templates are created by a person based on the characteristics of the business for each business related to each part (or unit) or each manufacturing process.
In the case of an entity data model template, organize and create from product design information (including knowledge and know-how). This is information, knowledge, and know-how related to product design work, such as unit or part functionality, performance, material (material), geometric shape, specification information, business standards, quality standards, etc. In addition, information on defects due to past design defects (phenomenon, cause, countermeasures, etc.) (these correspond to knowledge and know-how), etc.

  In the case of the state data model template, it is organized and created from process design information and production manufacturing information (including knowledge and know-how). These are information, knowledge, and know-how related to process design operations and manufacturing operations for manufacturing products, and process specification information such as unit or component manufacturing methods, conditions, process order, equipment, jigs, machine specifications, layouts, etc. Business standards, quality standards (these correspond to know-how), and defect information (phenomenon, causes, countermeasures, etc.) due to past design defects (these correspond to knowledge / know-how), etc. . It also includes various knowledge and know-how from production sites and countermeasures against poka-yoke.

  As for the actual implementation, if the template is conscious of input / output, the left attribute group corresponds to the input and the right attribute group corresponds to the output, but when used as a local variable, the left and right attributes There is no particular need to distinguish it from. In other words, it can be used as input or output on either the left or right side.

  After the frameworks (initial states) 221 to 223 of the entity / state data model 200 are automatically created by the processing described above, the entity / state data model 200 is thereafter completed by manual determination and additional input operations.

First, processing for adding the determination unit 33 and its input / output relationship will be described with reference to FIG.
First, FIG. 20 (c) shows a result of performing an operation of adding the determination unit (method) 203 and its input / output relationship. The figure shows entity / state data models 231, 232, and 233 in a state where two units of method d and method e are added as the determination unit 203. Taking method d as an example, the input / output relationship is that stress 1, stress 2, strength 1 and strength 2 are output, and output is mode or failure 3.

  A specific example of the method d is shown in FIG. This is an example corresponding to SSM logic. That is, it is determined whether the corresponding failure mode is true using the if-the rule for the stress and strength to be input.

  As is well known, for the time being, the strength is a comprehensive characteristic representing the tolerance and functionality of the failure mode to be designed and evaluated, and stress is given to an object with a certain strength. It is a driving force that causes a failure mode. The failure mode relates to a causal chain, and is an undesirable phenomenon, state, characteristic change, or the like that occurs in a target object.

  In addition, method e includes equipment (equipment failure (probability)), tool (tool failure (probability)), and manufacturing method (manufacturing method) in the state data model corresponding to state 2 (intermediate product C1 in FIG. 12). Defective), manufacturing conditions, and work procedure are input, and output is performed for defect 3 in the finished product and defect 2 in the intermediate product (state 2). A specific example of the method e is to hierarchically express a failure or damage mechanism by a combination of AND logical product and OR logical product corresponding to FTA logic as shown in FIG. 20B, for example.

  Various method logics such as methods d and e are selected and used if they are created in advance and registered in the method library, etc. create.

Further, the method is not limited to the method based on the SSM logic and the FTA logic, and there is also a method corresponding to, for example, FMEA logic.
FIG. 21 shows an example of a method corresponding to FMEA logic.

  As described above, in the FMEA method, the risk priority (number of risk priority) in the failure (failure) mode is ranked based on the occurrence frequency, the influence degree, and the detection difficulty level. As shown in the figure, the risk priority number is calculated by occurrence frequency × effect level × detection difficulty level. A risk priority number is obtained for each of the failure modes 1 to n, and these are compared in size to rank them.

Further, in the knowledge / know-how systematization apparatus 10 of this example, the registration function unit 12 in FIG. 1 can perform the attribute value registration support processing described in FIG.
In the framework (initial state) 221 to 223 of the automatically created entity / state data model 200, for example, attributes such as mode (failure mode), phenomenon 3, cause 2, and countermeasure 2 are stored. The specific attribute value has not been registered yet.

  Registration of this attribute value is not an automatic process but is performed manually by a human. However, in this example, as shown in FIG. 22, various master tables stored in the master table 241 based on the various coverage tables created in advance and stored in the coverage table DB 21, that is, the part master table, the phenomenon master By listing the contents of the table, result master table, failure mode master table, cause master table, countermeasure master table, etc. on the display of the user's information processing terminal, for example, and letting the user select the necessary items, the attribute value Can be registered. For example, when registering an attribute value for “phenomenon” among the attributes on the input side in the illustrated entity / state data model 200, a phenomenon master table is displayed. The phenomenon master table stores specific phenomena, for example, heat generation, vibration, noise, and the like, and by registering attribute values corresponding to the attribute “phenomenon” by displaying them in a list and allowing the user to select them. Can do.

  As a result, it is possible to save the trouble of data registration and to level the description of the attribute value regardless of whether the input person is a veteran or a beginner. Furthermore, by managing the maintenance authority of the master table, it is possible to ensure data consistency by preventing the contents of various master tables from being changed arbitrarily.

The registration function unit 12 is not limited to the above function, and may have a function for registering the coverage table (its master table) itself.
Further, the attribute values selected from the various master tables may be described in the knowledge / know-how sheet stored in the knowledge data DB 24.

  Furthermore, not only the above-described attributes but also other attributes (impact level, frequency, probability data, etc.) are included to update the attribute values of the entity / state data model 200 or add new attributes・ Knowledge is updated and the system can evolve.

For example, the attribute value can be updated by
・ Update the attribute value related to the occurrence frequency (attribute) from 2 times / month to 3 times / month ・ Change the attribute value related to tolerance (strength) from -30 ℃ to + 50 ℃ to -20 ℃ to +55
The attribute value related to update / degree of influence (attribute) is updated from 2 to 5, etc.

  Here, evolution means that the result of analysis changes due to the update of the attribute value or the like. For example, since the impact analysis depends on the greater impact level, if the impact level changes, the result will be led in a direction different from the previous analysis. Moreover, the flow of analysis may change due to the addition of attributes. Since addition, deletion, and change of processes (states) are easy, it goes without saying about evolution due to addition, deletion, and change of processes.

  Further, the master maintenance support function unit shown in FIG. 22 allows the user to select and display an arbitrary master table, and the user can arbitrarily change, add, delete, etc. the contents of the master table on this display screen. The master table can also be maintained by performing the input operation.

FIG. 23 is a diagram for explaining a reference function by the reference function unit 13 of FIG.
The above-described knowledge and know-how systematization apparatus 10 is connected to a network 240 such as a LAN, and is connected to a network 240 from an information processing terminal (not shown) provided at each site (product design / production preparation / mass production). The knowledge and know-how systematizing device 10 can be accessed via the information and the knowledge and know-how extracted from the knowledge data DB 24 and the quality knowledge DB 23 can be referred to. For example, the person in charge or the like refers to the work for quality improvement. Alternatively, in the business of developing a new product with reference to the existing design information 402 or the existing biotechnology information, the development staff and the like refer to the knowledge and know-how described above for development, and the design information 402-1 after the change Post-production technique information 404-1 is created. Alternatively, in the development work, a developer or the like uses FMEA and FTA information extracted from the knowledge data DB 24 and the quality knowledge DB 23 for reliability design of a new product.

Further, based on the entity / state data model 200, a check sheet for checking the use of knowledge and know-how may be generated.
FIG. 24 shows an example of check sheet generation.

The illustrated check sheet 250 includes a check target 251 and check essentials 252. The check target 251 performs concept division, and in the example shown in the figure, the check target 251 is divided into three layers, a large division, a medium division, and a small division. In the check requirement 252, each check item related to each of the subcategories classified hierarchically by this concept category is identified and stored as item 1, item 2, item 3, and the like. A specific example of each check item is not particularly shown, but in the figure, for example, check items related to the subdivision “ aluminum (metal)” are indicated as “A, U, E, O” and the like.

  The check target 251 and the check requirement 252 are described based on the entity / state data model 200. The check target 251 is described based on the description content of the input unit 201 of the entity / state data model 200. Among the description contents of the input unit 201, the attribute (attribute name) is described in the major category and the middle category, and the attribute value is described in the minor category. The attribute is a structured design / manufacturing item, and the attribute value is the content of this item. This is obtained, for example, by searching the entity / state data model 200 using an arbitrary attribute (name) as a key. In the illustrated example, for example, as a result of searching the quality knowledge DB 23 for the attribute “material”, the attribute values “iron (metal)”, “aluminum”, and “plastic” are obtained. The check essentials 252 describe defect countermeasures, design / manufacturing know-how and procedures, etc., which are extracted from the description contents of the output unit 202 of the entity / state data model 200 corresponding to each attribute value. It is.

The check sheet can be used, for example, at the time of design change, when checking a changed part or checking for omission of change.
FIG. 25 is a diagram illustrating an example of a hardware configuration of a computer that realizes the knowledge / know-how systematizing apparatus 10 described above.

  A computer 300 shown in the figure includes a CPU 301, a memory 302, an input unit 303, an output unit 304, a storage unit 305, a recording medium drive unit 306, and a network connection unit 307, which are connected to a bus 308. It has become. The configuration shown in the figure is an example, and the present invention is not limited to this.

The CPU 301 is a central processing unit that controls the entire computer 300.
The memory 302 is a memory such as a RAM that temporarily stores a program or data stored in the storage unit 305 (or the portable recording medium 309) during program execution, data update, or the like. The CPU 301 uses the program / data read out to the memory 302 to execute the various processes described above (processing of each functional unit in FIG. 1, processing shown in FIG. 18, etc.).

The input unit 303 is, for example, a keyboard or a mouse.
The output unit 304 is a display, for example.
The network connection unit 307 is configured to connect to a network such as an intranet or the Internet and perform command / data transmission / reception with another information processing apparatus.

  The storage unit 305 is, for example, a hard disk or the like, and is a program / data for causing the computer 300 to execute the various processes and functions described above (a program for executing the process of FIG. 18 and the like, and data of various databases shown in FIG. 17). Etc.) are stored.

  Alternatively, these programs / data may be stored in the portable recording medium 309. In this case, the program / data stored in the portable recording medium 309 is read by the recording medium driving unit 306. The portable recording medium 309 is, for example, an FD (flexible disk) 309a, a CD-ROM 309b, a DVD, a magneto-optical disk, or the like.

  Alternatively, the program / data may be downloaded from another device via a network connected by the network connection unit 307. Or you may further download what was memorize | stored in the external other apparatus via the internet.

  In addition, the present invention can be configured not only as a portable storage medium recording a program for realizing the various processes of the present invention on a computer, but also as the program itself.

It is a functional block diagram of the systematization apparatus of knowledge and know-how. The systematization device of knowledge and know-how and the processing flow of the entire system related to this are shown. It is a figure for demonstrating roughly the systematization apparatus of knowledge and know-how, and the existing method. It is a figure which shows the image of the whole input / output. It is a figure which shows the kind of coverage table. It is a specific example of a failure part coverage table. It is a specific example of a failure mode coverage table. It is a specific example of a failure phenomenon coverage table. It is a specific example of a failure result coverage table. It is a specific example of a failure cause coverage table. It is a specific example of a failure countermeasure coverage table. It is a figure for demonstrating the expression model of a failure causal chain. It is a figure which shows the data structure in the case of storing the representation model shown in FIG. 12 in the memory and storage of information processing apparatus. It is a figure which shows the unit representation image of a failure mechanism. An example of the entity / state data model created based on the product / process model DB is shown. It is the figure (the 1) which shows the data construction image of quality knowledge DB based on product and process model DB. It is FIG. (2) which shows the data construction image of quality knowledge DB based on product and process model DB. It is a creation processing flowchart figure of the information of quality knowledge DB based on PPMDB information. (A), (b) is a figure which shows an example of the data structure corresponding to the product model produced | generated as a processing result in the middle of the process of FIG. 18, and the data structure corresponding to (product model + state data model). (A), (b) is a specific example of the method by SSM and FTA logic, (c) is a figure which shows the example which described the method additionally with respect to the automatic creation result shown in FIG.19 (b). It is a figure which shows the specific example of the method by FMEA logic. It is a figure for demonstrating the registration assistance function of the attribute value based on an coverage table. It is a figure for demonstrating the reference function of a database. It is a figure which shows the production | generation of a check sheet based on an entity / state data model, and its example. It is a computer hardware block diagram. It is a figure which shows the flow of the information in the conventional design, production preparation, and mass production manufacture. It is a figure which shows the business flow of the conventional malfunction handling process.

Explanation of symbols

10 Knowledge and Know-how Systematization Device 11 Systematization / Storage Function 12 Registration Function Unit 13 Reference Function Unit 21 Coverage Table DB 21
22 Product / Process Model DB (PPMDB)
23 Quality Knowledge DB (QA2-KDB)
24 Knowledge Data DB
25 Unit Representation Example of Failure Mechanism Using Product / Process Model DB 31 Failure Part Coverage Table 32 Failure Mode Coverage Table 33 Failure Phenomenon Coverage Table 34 Failure Result Coverage Table 35 Failure Cause Coverage Table 36 Failure Countermeasure Coverage Table 50 Failure Causal Chain Representation model 51 entity (unit C)
52 entity (part A)
53 entity (part B)
61 state (initial (preparation))
62 state (intermediate product C1)
63 State (intermediate product C2)
64 state (completed product)
66-69 arrow (state transition)
71, 72 arrows (state transition)
81 state (raw material)
82 state (intermediate product A1)
83 state (intermediate product A2)
84 state (intermediate product A3)
85 state (finished product)
86-90 arrow (state transition)
91 State (raw material)
92 state (intermediate product B1)
93 state (intermediate product B2)
94 State (Completed)
96-99 arrow (state transition)
100 tables (PPM information)
101 entity code 102 state code 103 previous state code 104 state transition order 105 state transition code 106 input item code 107 output item code 108 configuration flag 110 unit representation 200 of failure mechanism entity / state data model 201 input unit 202 output unit 203 determination unit (Method)
204 entity data model 205 state data model 211, 212 entity data model template 213 state data model template 214 data structure corresponding to product model 215 data structure corresponding to product / process model 221 corresponding to unit C (product model + state data) model)
222 Corresponding to part A (product model + state data model)
223 corresponding to part B (product model + state data model)
231 to 233 entity / state data model 241 master table 300 computer 301 CPU
302 Memory 303 Input unit 304 Output unit 305 Storage unit 306 Recording medium drive unit 307 Network connection unit 308 Bus 309 Portable recording medium

Claims (2)

In a system for building a database of knowledge and know-how that systematically accumulates knowledge and know-how on everything from design to manufacturing processes in production systems,
Integrate information representation model based on product structure and information representation model based on process structure for the above knowledge and know-how to model product design defects, model defects in production preparation / manufacturing, and product manufacturing process As a data structure to describe information that expresses the failure propagation mechanism based on the cause-and-effect chain relationship,
An entity code uniquely assigned to an entity such as a unit or a part ;
A status code uniquely assigned to each status;
A previous state code uniquely assigned to a state before transition to the state indicated by the state code;
A state transition order number consisting of serial indicating the order of state transition in the manufacturing process of one entity;
A state transition code indicating the condition (type) of state transition;
An input item code uniquely assigned to an item (state) used (input) in each step of the actual manufacturing process;
An output item code uniquely assigned to an item (state) manufactured (output) in each step of the actual manufacturing process;
A configuration flag indicating a part constituting an entity such as a block or a unit;
Means for building a product process database by storing each of the data structures, and
Means for obtaining information representation models of products and processes accumulated in the constructed product process database;
Based on the configuration flag in the acquired information representation model, means for hierarchically extracting product configuration information about entities such as blocks and units, each entity and elements (sub-entities) constituting each entity;
Corresponding to each extracted entity (including sub-entities), means for acquiring an entity data model template in which various target events related to each entity are defined as attributes from a storage means that is created and stored in advance,
Means for creating a data model corresponding to a product model using the extracted product configuration information and each of the acquired entity data model templates;
Based on the state code in the acquired information representation model, corresponding to each state of each entity, a state data model template defining various target events related to each state as attributes is created and stored in advance. Means for obtaining from storage means,
The acquired state data model template is sequentially added to each entity data model template in the created data model corresponding to the product model according to the order of the state transition order in the acquired information expression model. Means to create and store a data model framework for product and process models,
Methods and hazards that read the stored data model framework corresponding to the product / process model, perform attribute values corresponding to each attribute developed in the framework of the data model, and perform various determinations using the attribute values A means to describe the propagation and build a quality knowledge database;
A systematic database construction device for knowledge and know-how characterized by having at least . In the method of creating a systematic database of knowledge and know-how that systematically accumulates knowledge and know-how about everything from design to manufacturing processes in production systems,
Integrate information representation model based on product structure and information representation model based on process structure for the above knowledge and know-how to model product design defects, model defects in production preparation / manufacturing, and product manufacturing process As a data structure to describe information that expresses the failure propagation mechanism based on the cause-and-effect chain relationship ,
An entity code uniquely assigned to an entity such as a unit or a part;
A status code uniquely assigned to each status;
A previous state code uniquely assigned to a state before transition to the state indicated by the state code;
A state transition order number consisting of serial indicating the order of state transition in the manufacturing process of one entity;
A state transition code indicating the condition (type) of state transition;
An input item code uniquely assigned to an item (state) used (input) in each step of the actual manufacturing process;
An output item code uniquely assigned to an item (state) manufactured (output) in each step of the actual manufacturing process;
A configuration flag indicating a part constituting an entity such as a block or a unit;
And building a product process database by accumulating these data structures,
Obtaining an information representation model of products and processes accumulated in the constructed product process database;
Based on the configuration flag in the acquired information representation model, a process of hierarchically extracting product configuration information about entities such as blocks and units, each entity and elements (sub-entities) constituting each entity;
Corresponding to each extracted entity (including sub-entities), a process of acquiring an entity data model template that defines various target events related to each entity as attributes from a storage means that has been created and stored in advance,
Using the extracted product configuration information and each acquired entity data model template to create a data model corresponding to a product model ;
Based on the state code in the acquired information representation model, corresponding to each state of each entity, a state data model template defining various target events related to each state as attributes is created and stored in advance. The process of obtaining from the storage means
The acquired state data model template is sequentially added to each entity data model template in the created data model corresponding to the product model according to the order of the state transition order in the acquired information expression model. The process of creating and accumulating the framework of the data model corresponding to the product / process model,
Methods and hazards that read the stored data model framework corresponding to the product / process model, perform attribute values corresponding to each attribute developed in the framework of the data model, and perform various determinations using the attribute values The process of describing the propagation and building the quality knowledge database;
To create a systematic database of knowledge and know-how that includes at least