Method and system for controlling and / or monitoring production processes and / or production facilities within an industrial production chain

Abstract

A computer-implemented method for controlling and / or monitoring a production process and / or a production facility as part of an industrial production chain is disclosed, wherein different taxonomies of chemical, physical and / or technical entities within an ontology / taxonomy domain provided by at least two participants of the industrial production chain are automatically resolved to keep the taxonomies consistent along the production chain, wherein underlying information / data of chemical, physical and / or technical entities within one taxonomy domain provided by the at least two participants of the production chain are evaluated with respect to their semantics and / or technical meaning, and wherein based on the evaluation results of the semantics and / or technical meaning of the underlying information / data of the chemical, physical and / or technical entities, taxonomy relationships between the underlying information / data of the at least two participants are determined.

Description

Technical Field

[0001] This disclosure relates to industrial production chains or value chains, particularly industrial production chains or corresponding supply chains involving at least two production chain participants. Background Technology

[0002] In the realm of industrial production and corresponding production processes or underlying production chains, different companies use different classifications or ontology for segregants, products, intermediate products, or even production processes or process steps. Such classifications or ontology include, but are not limited to, different technical classifications, technical terms, and corresponding technical system classifications or structures. Such technical or scientific terms can be used for, for example, raw materials, products to be produced and their components, and even for production process steps and related services.

[0003] Currently, these different taxonomies or ontology are processed manually, which may negatively impact the control and / or monitoring of underlying production processes and / or facilities if there are differences or deviations in the underlying taxonomies or ontology.

[0004] US 2017 / 0278000 A1 discloses a method for digital automation of manufacturing bodies in the field of robotics, where products are defined according to their manufacturing methods. More specifically, it describes a computer-implemented method for automating manufacturing supply chain planning, which provides:

[0005] Computer processor, which is used to process data.

[0006] At least one input device,

[0007] At least one output device,

[0008] Computer-readable storage device

[0009] A first entity is used to define the product according to a manufacturing method of one product, or according to manufacturing methods of multiple products.

[0010] The second ontology, which defines the capabilities of multiple manufacturing facilities, and

[0011] A knowledge representation and reasoning system that executes on the computer processor.

[0012] The described technology integrates robot capability models with manufacturing ontology within an artificial intelligence (AI) inference system. This AI inference system supports the classification of robot capabilities and the development of product descriptions based on the manufacturing steps required to utilize these capabilities and produce various products. The system's output is a sequence of steps required to produce an underlying product that meets and optimizes given product specifications, enabling the development of a manufacturing plan for that product. This will allow consumers, application software, and manufacturers to design intelligent products, thereby automating the supply chain. Summary of the Invention

[0013] This disclosure relates to a computer-implemented method for controlling and / or monitoring production processes and / or facilities as part of an industrial production chain, wherein different classifications of chemical, physical, and / or technical entities within a technical or scientific ontology / taxonomic domain provided by at least two participants in the industrial production chain are automatically parsed to ensure consistency of these classifications along the production chain, wherein underlying information / data of chemical, physical, and / or technical entities within a technical or scientific ontology domain provided by the at least two participants in the production chain is evaluated with respect to their semantic and / or technical meaning, and wherein a taxonomic relationship between the underlying information / data of the at least two participants is determined based on the evaluation results of the semantic and / or technical meaning of the underlying technical or scientific information / data of the chemical, physical, and / or technical entities.

[0014] The disclosed methods for evaluating the semantic and / or technical meaning of the underlying information / data of these chemical, physical and / or technical entities may be based on predefined classes and / or subclasses of these entities.

[0015] According to the disclosed methods, these chemical, physical, and / or technical entities can be represented by corresponding databases, reference lists, semantic models, and / or digital twins.

[0016] According to the disclosed method, a technical / scientific ontology or classification template for these chemical, physical and / or technical entities can be provided based on a predetermined ontology / classification of the underlying processes and / or products.

[0017] According to the disclosed methods, classification can utilize artificial intelligence, machine learning, or rule-based processing methods to determine similarity based on the technical and / or semantic meaning of technical terms.

[0018] According to the disclosed method, in the case of machine learning methods, existing data related to production processes, materials, etc., of at least one participant in the production chain can be pre-classified using existing taxonomy methods.

[0019] According to the disclosed method, semantic text recognition based on natural language processing methods can be performed even when the technology or technology ontology / classification is unknown.

[0020] According to the disclosed method, data integration can be performed for production chain participants based on an extract-transform-load procedure to combine data from multiple data sources into a single consistent data store.

[0021] According to the disclosed methods, these chemical, physical, and / or technical entities can be key performance indicators of underlying industrial processes or products.

[0022] This disclosure also relates to an apparatus for automatically resolving different taxonomic classifications of chemical, physical, and / or technical entities within a technical or scientific ontology / taxonomic domain provided by at least two participants in an industrial production chain, so that these taxonomic classifications remain consistent along the production chain. The apparatus includes a processing unit for evaluating underlying information / data of a chemical, physical, and / or technical entity within an ontology / taxonomic domain provided by the at least two participants in the production chain with respect to its semantic and / or technical meaning, and for determining ontology / taxonomic relationships between the underlying information / data of the at least two participants based on the evaluation results of the semantic and / or technical meaning of the underlying information / data of the chemical, physical, and / or technical entity.

[0023] This disclosure also relates to a distributed manufacturing / production chain system for automatically resolving different taxonomic classifications of chemical, physical, and / or technical entities within a technical or scientific ontology / taxonomic domain provided by at least two participants in an industrial production chain, so that these taxonomic classifications remain consistent along the production chain. The system includes one or more computing nodes for evaluating the underlying information / data of chemical, physical, and / or technical entities within an ontology / taxonomic domain provided by the at least two participants in the production chain with respect to their semantic and / or technical meaning, and for determining the ontology / taxonomic relationship between the underlying information / data of the at least two participants based on the evaluation results of the semantic and / or technical meaning of the underlying information / data of the chemical, physical, and / or technical entities.

[0024] This disclosure also relates to a data structure stored in a computer-readable storage medium, which includes a data format for dynamically resolving different taxonomies, the data format being independent of the different taxonomies provided in a distributed manufacturing / production chain.

[0025] The disclosed data structure may include technical or scientific ontology / taxonomy templates for these chemical, physical and / or technical entities based on a predetermined ontology / taxonomy of the underlying processes and / or products.

[0026] This disclosure also relates to the use of the methods and / or data structures and / or digital twins disclosed herein for controlling and / or monitoring production processes and / or production facilities as part of an industrial production chain, wherein different taxonomic classifications of chemical, physical and / or technical entities within the domain of technology or scientific ontology / taxonomics provided by at least two participants in the industrial production chain are automatically and / or dynamically resolved to ensure that these taxonomic classifications or ontology remain consistent along the production chain.

[0027] This disclosure also relates to a computer program product for generating an ontology / taxonomy-based digital twin of an underlying industrial product or production process using a template based on a predetermined ontology or taxonomy of the underlying product. The digital twin can be semantically described, wherein corresponding elements or objects of the underlying production chain are classified, and wherein the digital twin is dynamically modified along the production chain according to different ontologies or taxonomy of the participants in the production chain.

[0028] This disclosure also relates to a computer element having instructions that, when executed on one or more computing nodes, are configured to perform the steps of the methods disclosed herein, or are configured to be executed by the apparatus or system disclosed herein. Attached Figure Description

[0029] Figure 1 A schematic diagram of the global automotive production chain is depicted to illustrate the different technological or scientific ontology / taxonomic areas during typical pre-manufacturing, manufacturing, and post-manufacturing processes;

[0030] Figure 2 A schematic diagram depicting a typical ontology definition used in the chemical industry is provided.

[0031] Figure 3 A combo box / flowchart is shown to illustrate exemplary digital process steps for automating the generation of taxonomic-based digital twins through classification (i.e., in the field of coating production in the automotive industry);

[0032] Figure 4 A flowchart is shown for demonstrating the digital recognition of KPIs (as entities) provided in different classification systems;

[0033] Figure 5 A schematic diagram depicts the production process steps for coatings used in the automotive industry, illustrating the basic aspects of the corresponding taxonomic transformation.

[0034] Figure 6a , Figure 6b Exemplary material codes used in the chemical industry are depicted to demonstrate how data can be semantically represented and connected based on the corresponding ontology. Figure 6b (and Figure 6c); and

[0035] Figure 7 An exemplary structure is shown to illustrate the ontological panorama of the European Commission’s (EC) “EMMC-CSA” project in the chemical industry. Detailed Implementation

[0036] It is generally known that "classification" requires describing, naming, and classifying objects or organisms.

[0037] In an industrial supply chain environment, taxonomy refers to the hierarchical classification of entities of interest within underlying business or industrial entities (such as businesses, organizations, or administrative departments) for categorizing documents, digital assets, and other information. Within such entities, taxonomy more specifically identifies hierarchical relationships within information categories and is used to organize information for use both internally and externally. Internally, taxonomy can be used to classify documents into categories such as proposals, contracts, correspondence, and briefings. Therefore, taxonomy is a crucial aspect of supply chain digitalization and related automation.

[0038] Taxonomy is designed to categorize items within only one dimension. In Product Information Management (PIM) systems, taxonomical relationships exist in the digital twins of production assets (e.g., product part numbers), but this taxonomy cannot be extended to other assets within the company or even outside the company.

[0039] On the other hand, ontology provides an additional layer to this taxonomic relationship, extending it from the product domain alone to other domains (such as customer data and digital assets).

[0040] According to existing technology, this classification of underlying business entities or corresponding industrial entities is processed, for example, through mapping tables and including data on underlying products, processes and services, in order to associate, for example, a part number mentioned by one company with a corresponding part number of another company.

[0041] In another example, within the technical field of coating materials, testing procedures are applied to determine the mechanical and optical surface properties of the produced coating. To determine these physical properties, a large amount of test data is generated, such as experimental data from salt water tests (German: "Salzlaugentest") or stone impact tests (German: "Steinschlagtest"). During such tests, a great deal of customer-specific, and therefore unstructured, data or datasets are produced.

[0042] According to this disclosure, it has been recognized that the above data is based on different local or proprietary classifications, but consistency is needed throughout the production chain to avoid serious production problems. Furthermore, these different classifications may also lead to product quality issues, such as "COA" (Certificate of Authenticity) problems, because different DIN standards are required to interpret the meaning of the underlying assessment results.

[0043] The underlying concept is based on entities, such as data objects or the attributes they contain, like the known "digital product passport." Specifically, it proposes utilizing such data objects along the entire production chain (i.e., "end-to-end" in a complete production chain, or even throughout the entire lifecycle of the underlying product or its corresponding intermediate product), where further attributes regarding the underlying taxonomies of different participants along the entire production chain are added. These further attributes can be added to the underlying data objects at logistics or product-related business interfaces, for example, via the corresponding program interfaces of the underlying Enterprise Resource Planning (ERP) systems located on one or both sides of the two corresponding participants in the production chain.

[0044] This disclosure includes the following aspects:

[0045] The proposed data objects can be implemented as reference lists based on digital twins of corresponding underlying products pre-built using an underlying template based on a predetermined ontology or a taxonomy concerning the underlying product. Such an ontology could relate to, for example, coating materials delivered to the automotive industry.

[0046] - Such a digital twin can be described semantically, in which corresponding elements or objects of the underlying production chain are classified, and in which the mentioned digital twin can be dynamically enriched along the production chain based on the different ontologies of the production chain participants.

[0047] - A new data format for dynamically parsing different taxonomic classifications, which is independent of the different taxonomic classifications.

[0048] - An automatically implemented classifier method that is based on AI, ML, or rules and is used to determine similarity based on the technical and / or semantic meaning of technical terms.

[0049] In the case of machine learning (ML) methods, existing data related to business processes, production processes, materials, etc., of at least one participant in the production chain are pre-classified with reference to existing taxonomy methods.

[0050] If the taxonomy is unknown, semantic text recognition can be performed, for example, using known Natural Language Processing (NLP) methods. The data integration process required by all participants can then be implemented using known Extract-Transform-Load (ETL) procedures, which combine data from multiple data sources into a single, consistent data store and load it into a data warehouse or other target system.

[0051] Using the automated digital classification parsing method proposed in this paper, any entity (i.e., any subject, object, process, material, product, etc.) in the underlying production chain can be represented along the entire production chain through ontology / classification (e.g., through a corresponding semantic model). Thus, different classifications of different production chain participants can be linked together in a fully automated manner.

[0052] Based on the proposed automated taxonomy parsing method, a given taxonomy can be horizontally extended (i.e., along the production chain) with additional semantic models, and to some extent "grows" along with the underlying knowledge domain or automatically generated knowledge base or system. In other words, new use cases for a given taxonomy can be dynamically and automatically generated.

[0053] Figure 1 The exemplary global automotive production chain shown includes different taxonomic areas along three exemplary manufacturing process stages (pre-manufacturing stage 100, manufacturing stage 105, and post-manufacturing stage 110). In particular, the production of plastic materials required for global automotive production chain 115 is schematically depicted. The underlying raw materials supplied by chemical companies are labeled by boxes 120, 125, 130, 135, and 140.

[0054] The pre-manufacturing phase 100 includes the research, design, and development of automotive products 145. The manufacturing phase 105 includes the production of materials and components 150 and system modules 155. Following the intermediate steps 160 of downstream integration and assembly of automotive components, post-manufacturing 110 includes after-sales service 165 and spare parts and recycling processes 170. During the manufacturing phase 106, basic components 180 are produced from starting materials 175 at "Level 3", subsystem components 185 at "Level 2", and system modules 190 at "Level 1".

[0055] As highlighted in the lower part of the diagram, in this production scenario, the starting (raw) materials 175 are rubber and plastic 120. From these starting materials 175, at "Level 3," basic components, such as processed rubber and plastic 125 and composite materials 130, are produced. From these basic components 125 and 130, at "Level 2," automobile body parts and interior parts 135 are manufactured. From these subsystem components 185, at "Level 1," automobile exterior and body, electronic equipment, and interior 140 are manufactured.

[0056] For all the described materials 175, basic components 180, subsystem components 185, and system modules 190, in most cases, suppliers and manufacturers use different material specifications and part numbers based on their respective proprietary ontology or classification system. These differences can lead to serious production-related problems.

[0057] exist Figure 2 The image illustrates different examples of ontology definitions in the chemical industry. In the lower part of the illustrated ontology and at the corresponding levels, “individuals” 200 representing things or objects that can be pointed to in the chemical environment are shown. These individuals 200 are linked 205, 215 to class 220 and subclass 210 via the predicate “rdf:type” 225 (i.e., a property used to declare that a resource is an instance of a class) and the predicate “rdfs.subClassof” 230 (describing that the individual is a subclass of a class). This RDF schema (=“Resource Description Framework” schema) can be based on the “Web Ontology Language” (OWL) defined by the World Wide Web Consortium (W3C).

[0058] The example of individual 200 shown in this article is the assumed company location of the underlying "Production Location of Chemical Company 1" 235, "Chemical Compound 1" 240, and "Chemical Reagent X" 245. These "individuals" 200 are linked to corresponding / corresponding subclasses "Chemical Company 1" 250, "Chemical Compound" 255, and "Chemical Reagent" 260, which are the corresponding higher-level terms. Further higher-level subclasses are "Chemical Company" 265, "Chemical Compound" 270, and "Chemical Reagent" 275. In this example, the corresponding top-level classes are "Company" 280, "Chemical Substance" 285, and "Product" 290.

[0059] In the diagram shown, another (horizontal) relationship between certain classes / subclasses is: "Chemical Company" Production 293 "Chemical substances" and "Chemical compounds" 255 For 295 "Chemical Reagents" 260.

[0060] exist Figure 3 In the combo box / flowchart shown, different taxonomic areas are depicted by dashed lines 300, 305, 310, and 315. The first company can be a coating material manufacturing company. The first two taxonomic areas are provided within the first company 302 (e.g., the coating material manufacturing company), while the other two taxonomic areas are provided outside the first company 302, i.e., by other participants 303 in the associated production chain or corresponding "production chain" (e.g., OEMs or companies that manufacture coating materials).

[0061] Based on chemical raw material 320, the company has developed 325 new coating materials for small-batch application to vehicles such as passenger cars or trucks. These new coating materials can be selected from electrophoretic coating materials, primer materials, paint and varnish materials, adhesive compositions, pigment pastes, topcoat materials, and hardener compositions.

[0062] The development and production of new coating materials can be carried out based on information 323 regarding these materials. The development and production of small batches of new coating materials can be associated with the first classification field 300.

[0063] In this taxonomy domain 300, the underlying entities are product data and corresponding reliability data considered as key performance indicators (KPIs) in this embodiment, which need to be used both inside and outside the first company, for example, in conjunction with the entire production chain or other participants in the production chain. Such KPIs / #1 can be coating specification data, such as physical property data (e.g., density, viscosity, solvent content, pH), application data, and / or appearance data (e.g., color data).

[0064] It is worth mentioning here that using KPI data as the basis for automatically processing taxonomic problems along the entire production chain is only a preferred option, because this data can be considered the most important data that needs to be shared by participants in the production chain and that all participants have the same understanding of the technical or business-related implications of this shared data.

[0065] Similarly, the second classification area 305 provided within the first company can be associated with two other technical areas 335 and 340, which also generate certain KPIs that need to be shared with external companies (e.g., the entire production chain / participants in the production chain).

[0066] The first technical field 335 of the second classification area 305 relates to scaling up and testing, wherein a larger volume of newly developed coating materials is produced, and prototypes of coated products are generated and tested. Such testing may include stress testing of the produced coated products to determine their durability under real-world conditions. This KPI / #2 may be coating reliability data, such as weather resistance, scratch resistance, blister resistance, appearance data, application data, curing data, and / or adhesion performance.

[0067] The second technical field 340 of the second classification field 305 relates to the large-scale production of coating materials that have successfully passed tests in the first technical field 335. This KPI / #3 can be physical property data, curing data, and / or appearance data.

[0068] Based on the underlying process used for automated ontology / classification resolution, the three types of KPIs described (i.e., KPI / #1, KPI / #2, and KPI' / #3) are provided to another processing step 330 (328, 337, 343), where the underlying KPI data is extracted and evaluated regarding its semantic and / or technical meaning. This technical meaning may also include semantic aspects as well as perspectives on the coating material and the environmental impact of the coating produced therefrom.

[0069] It is worth mentioning here that processing step 330 not only processes the KPI data provided by the two taxonomic domains 300 and 305 provided by the first company, but also processes the KPI data created in other taxonomic domains 310 and 315 provided by other production chain participants (such as supply chain partners or even end customers).

[0070] In this scenario, the ontology / taxonomy resolution process includes two additional taxonomic domains 310 and 315. These taxonomic domains 310 and 315 are processed by two other participants 303 in the entire production chain (i.e., the original equipment manufacturer (OEM) 345 and the end customer 350 of the underlying products produced by the participants in the production chain). OEM 345 provides additional KPIs / #4, and the end customer 350 provides additional KPIs / #5. These KPIs are provided 348 and 353 to the already described processing step 330, in which additional underlying KPI data are also extracted and evaluated in the manner described above, i.e., evaluated regarding their semantic and / or technical meaning.

[0071] To generate an additional KPI / #4, OEM 345 applies coating material 347, supplied by underlying company 302, to the vehicle or part. The supplied coating material may be used at least partially to generate a multilayer coating. Such a multilayer coating may include at least two coatings prepared from different coating materials. A typical multilayer coating may include, for example, an electrophoretic coating produced by an electrophoretic coating material, a primer layer produced by a primer coating material or a first colorant material, a colorant layer produced by a colorant or a second colorant material, and a clear coat layer produced by a clear coat material.

[0072] To generate additional KPI / #5, end customer 350 uses this coated vehicle manufactured by OEM 345, where the body of the vehicle manufactured by OEM 345 is coated only in the final production step (i.e., after the vehicle is completed). However, the coating can be applied by OEM 345, or OEM 345 can use pre-coated parts supplied by a previous supplier; that is, the vehicle can also be manufactured by another (not shown) automotive / vehicle manufacturer using the mentioned automotive parts / components as pre-coated parts (i.e., parts already painted with this coating by OEM 345).

[0073] Considering the underlying usage conditions, these KPIs / #5 related to the use of Customer 350's vehicle can primarily relate to the usage conditions of the car / vehicle and the corresponding signs of wear and cracking in the coating due to the use of the car / vehicle. These signs of wear and cracking can be assessed or measured, for example, based on the roughness of the coating, which should increase during the use of the car / vehicle.

[0074] Based on all the extracted and evaluated KPIs in step 357, there are two additional processing steps 355 and 360. Step 360 follows step 358 after step 355 and is repeated through program loop 359.

[0075] In processing step 355, the extracted KPIs are categorized using an automated classifier. Based on this categorized KPI data, a digital twin (DT) of the underlying production chain is created. This DT created in step 355 provides a well-structured reference list of all KPIs, based on which the automated allocation or mapping of different KPI data used by different participants in the production chain can be achieved. Through this semantic allocation or mapping, it can be ensured that all these KPIs can be used with the same semantic and / or technical meaning throughout the entire production chain.

[0076] In processing step 360, the DT already created (355) is continuously enhanced or improved based on the newly collected KPIs / #1, KPI / #2, KPI / #3, KPI / #4, and / or KPI / #5. Thus, for example, other aspects of the production process or newly added participants can be considered to obtain the most complete picture of the underlying production and supply chain.

[0077] It is also worth noting that some or all of processing steps 330, 355, and 360 can be implemented as centralized processing units, wherein all participants in the production chain are connected to and / or have access to these processing units. In this scenario, the KPIs of all participants can be... Figure 3 The collection and processing are performed as described in the embodiments. Alternatively, in a distributed / decentralized approach, each participant may implement a processing device for performing the first processing step 330, wherein only the two additional processing steps 355, 360 are implemented in a centralized manner, such that all participants are connected to and / or have access to the corresponding two additional processing devices.

[0078] Figure 4 An exemplary implementation of a method for creating semantic relationships between KPIs of different entities from a production chain that belong to different taxonomic domains is shown. This method can be implemented using common artificial intelligence (“AI”) or “(random) forest” approaches.

[0079] In step 400, semantic recognition is performed on the KPIs existing in various taxonomic domains, and the identified KPIs are stored in a reference list or database. This database can be structured based on all existing entities in the production chain that provide KPI data for the underlying manufacturing or production processes.

[0080] In subsequent process loop 405, the absence of taxonomic / ontology-based relationships between two or more entities is repeatedly checked for an empirically predefined number of loops. If such a relationship is missing, the missing taxonomic relationships between these entities and / or the underlying taxonomic relationships are stored (at least temporarily), and step 405 is repeated.

[0081] After identifying a certain number of business-related or technical / scientific entities with missing taxonomic relationships, in subsequent step 410, for newly identified entities with missing taxonomic relationships, possible taxonomic relationships are determined or proposed. This determination is preferably done automatically, i.e., based on an existing database or reference list that includes semantic relationships (e.g., an existing ontology database, a thesaurus database with existing technical synonyms, antonyms, and / or other relevant technical information).

[0082] According to step 415, a reasonableness check is performed on the determined / proposed taxonomic relationship. This reasonableness check can be performed by cross-checking the taxonomic relationships of such taxonomic relationships with the taxonomic relationships of other entities in the entire production chain, or it can be performed based on general or semantic language rules.

[0083] Figure 5 The diagram schematically depicts three different taxonomic domains 510, 520, and 530 associated with coating material manufacturer 500, and how taxonomic conversion can be achieved between coating material manufacturer 500 and another exemplary participant in the production chain 505 (in this example, a vehicle manufacturer).

[0084] Classification 510 of the first category pertains to coating materials (e.g., newly developed coating materials, see...). Figure 3 The production and supply of frames 325, 335, and 340. The coating material may be an electrophoretic coating material, a primer material, a paint material, a varnish material, an adhesive composition, a pigment paste, a topcoat material, or a hardener composition.

[0085] The technical terminology used by the coating material manufacturer 500 must be consistent with the corresponding terminology used by the other production chain participant 505. Therefore, the different technical terms used by the two sides 500 and 505 must be translated in terms of both the technical or semantic meaning of the terminology used for processing the coating material 515 purchased from the coating material manufacturer 500 and the technical or business-related terminology used for marketing and selling the final coated vehicle product 540. Exemplary and potentially different technical terms in the first classification domain 510 are coating material name, coating material number, and / or coating material ID.

[0086] The second classification domain 520 relates to coated products manufactured from coating materials supplied by coating material manufacturer 500. However, in this coating material manufacturing scenario and the corresponding underlying classification domain, the technical terminology used by coating material manufacturer 500 must be consistent with the corresponding terminology used by another production chain participant 505. Therefore, any different technical terminology must also relate to the application of the coating material as automotive / vehicle paint 525 by the automotive / vehicle manufacturer 505. Figure 3 The technical or semantic meaning of the terminology used in step 335 for testing and evaluating the coating is translated.

[0087] However, it is worth noting that Figure 3 The OEM 345 described in the document may never have seen the paint material being produced, making it impossible for OEM 345 to provide any data about the material. In this case, the two classification fields 510 and 520 can be merged into a single classification field.

[0088] Therefore, the terminology used for coatings provided through the marketing and sale of underlying end products 540 (i.e., personal vehicles, multi-purpose vehicles, trucks, etc.) must be semantically consistent with the corresponding terminology used on the side of coating manufacturer 500. Exemplary and potentially different technical terms in the second taxonomic domain 520 are color name data, such as color code, color name, and color ID.

[0089] The third classification field 530 relates to automobile / vehicle frames, which are coated with [a coating used for] Figure 3 The coating materials produced in the second technical field 340 described herein. Also within this taxonomic field, the technical terminology used by the coating manufacturer 500 must be consistent with the corresponding terminology used by another production chain participant 505. Therefore, any different technical terminology must be translated into technical or semantic meaning regarding the coating material applied to the vehicle / car frame 535 and the terminology used by the vehicle / car manufacturer 505 to market and sell the entire vehicle / car 540. Exemplary and potentially different technical terms in the third taxonomic field 530 are coating material name, coating material number and / or coating material ID, and color name data (e.g., color code, color name, color ID).

[0090] Figure 6a and Figure 6b The paper describes an exemplary knowledge-based approach that connects data based on a given ontology to allow the retrieval of alternative data from ontology-based graph structures.

[0091] The method shown is based on the following semantic data for this ontology:

[0092] -EC raw material code;

[0093] - The amount of a specific substance in the raw materials;

[0094] - Substance name;

[0095] - The CAS Registry Number of the substance.

[0096] Figure 6a and Figure 6b This illustrates exemplary class relationships (hereinafter referred to as "related / linked"), where the boxes shown represent "class," "individual," and "value." The relationships shown between these boxes are "object attribute," "subclass," and "type." The "value" box contains the value data of the attribute of the "individual" in the main class (see [link]). Figure 6b ).

[0097] exist Figure 6a In the ontology-based graph structure shown, the first category "raw materials" 600 is related to the second category "quantity" 605, i.e., as a mentioned object characteristic relationship. The second category "quantity" is further related to two other categories "matter" 610 and "unit" 615, i.e., as another object characteristic relationship. The category "quantity" 605 is also related to the subcategory "value" 620, where the category "matter" 610 is also related to the corresponding subcategory "value" 630. In addition, the category "unit" 615 has a "type" relationship with the individual characteristic "percentage" 625.

[0098] Figure 6b The second example of the ontology-based graph structure depicted also includes the same four classes: "raw materials" 600, "quantity" 605, "substance" 610, and "unit" 615. However, in this example, these four classes 600 to 615 are not linked to each other or related accordingly based on object property relationships.

[0099] Instead of this relationship, these four classes 600 to 615 have a "type" relationship with the corresponding four individual characteristics 625, 635, 640, and 645. In this case, class "raw materials" 600 is associated with individual characteristic "EC MAT code" 635, class "quantity" 605 is associated with individual characteristic "zinc content" 640, class "unit" 615 is associated with individual characteristic "percentage" 625, and class "substance" 610 is associated with individual characteristic "zinc" 645.

[0100] Therefore, the individual characteristic "EC MAT code" 635 has an object characteristic relationship with the individual characteristic "zinc content" 640. The individual characteristic "zinc content" 640 further has two other object characteristic relationships with the individual characteristic "percentage" 625 and the individual characteristic "zinc" 645.

[0101] In addition, the individual characteristic “Zinc content” 640 is further correlated with the value subclass “96” 620', and the individual characteristic “Zinc” 645 is further correlated with the value subclass “1234-56-7” 630'.

[0102] Figure 7 An exemplary structure is depicted in the chemical industry based on the European Commission (EC) ontological panorama or corresponding taxonomic panorama.

[0103] A correct ontology enables semantic interoperability between different technology systems. The Semantic Interoperability Expert Group of the Alliance for Internet of Things Innovation (AIOTI) Standardization Working Group has created an ontology panorama, which currently includes 30 ontologies from different application areas of IoT.

[0104] Chemical Entities of Biological Interest (ChEBI) is a free dictionary focusing on molecular entities of small chemical compounds. The molecular entities discussed are either natural products or synthetic products used to intervene in processes in living organisms. Genomically encoded macromolecules (nucleic acids, proteins, and peptides resulting from protein cleavage) are generally not included in ChEBI. In addition to molecular entities, ChEBI includes groups (parts of molecular entities) and entity classes. ChEBI includes ontological classifications, which specify the relationship between a molecular entity or entity class and its parents and / or offspring. ChEBI is available online. http: / / www.ebi.ac.uk / chebi / Access it online.

[0105] The EU Raw Materials Intelligence Capacity Platform (EU-RMCP) – Technical System Specification (2018) – is based on an ontology in the mineral resources domain. Users can navigate the ontology using a Dynamic Decision Graph (DDG), allowing them to discover solutions they are looking for without asking any questions. The system is coupled with an RDF triplet store (a database storing the ontology) and fact tables, document tables, and process tables (i.e., specifically formatted forms) related to methods and documents, scenarios, and metadata.

[0106] This technology system connects with existing knowledge data platforms (KDPs) such as IKMS (EURare), EU-MKDP (Minerals4EU), EU-UMKDP (ProSUM), EU-CRMKDP (SCRREEN), the European Geological Data Infrastructure (EGDI) developed by EuroGeoSurveys (EGS), and RMIS 2.0 (Raw Materials Information System) currently being developed by the European Commission's DG JRC in Isapura, enabling users of all these KDPs to benefit from the underlying expert system.

[0107] The MICA main body actually covers seven thematic areas: 'primary' and 'secondary mineral resources', 'industrial processing and conversion', 'raw materials economics' (including CRM), 'raw materials policy and legal framework', 'raw materials sustainability', and 'international reporting'. Figure 3 DDG provides unique access to most available data in a single location, including contextual access to resources such as European legislation, and access to several key studies such as “Minventory Research 1” (Parker et al., 2015) or “Materials Systems Analysis 2” (Deloitte, BIO, 2015). The primary function of DDG and its accompanying applications is as an ‘intelligent’ search engine, where data, information, and knowledge are tightly and cleverly connected, making it a powerful decision-making aid. Therefore, DGG is not a ‘pure’ search engine that simply generates pre-defined answers like many other search engines we use.

[0108] Now return to the reference Figure 7 In this example, the “Raw Materials” ontology 715 mentioned above includes entities 'Raw Materials', 'Substance', and 'Quantity'. The entity 'Raw Materials' relates to the “Product Numbering System” ontology 730, which includes the entity 'PNS Code' and can be accessed via the corresponding DB query 'Has PNS Code'. The entity 'Raw Materials' further relates to the “FIIRM” ontology 720, which includes the entity 'FIIRM' and can be accessed via the corresponding DB query 'Has FIIRM'. "FIIRM" is a well-known organization focused on “Comprehensive Business Transformation”.

[0109] Furthermore, the entity 'Quantity' relates to the 'Unit' ontology 725, which includes the entity 'Unit' and can be accessed via the corresponding DB query 'Has Unit'. Further, the entity 'Substance' relates to the aforementioned 'ChEBI' ontology 700, which includes the entity 'Chemical Entity' and can be accessed via the corresponding DB query 'dbxref'. Secondly, the entity 'Substance' also relates to the 'Chemicals' ontology 705, which also includes the entity 'Chemical Entity' and can be accessed via the same DB query 'dbxref'. Further, the entity 'Substance' refers to the 'Property' ontology 710, which includes the entity 'CASRN' and can be accessed via the DB query 'Has CASRN'. Additionally, the 'Property' ontology 710 further includes the entity 'Boiling Point', which can be accessed via the 'FIIRM' ontology 720 via the DB query 'Has Boiling Point'.

[0110] As shown in box 735, the selected class of this EC ontology can have further relationships with external ontology.

[0111] The above or similar ontological panoramas can be used for, for example Figure 4 The method described herein is for finding missing ontology / taxonomic relations. In particular, for any newly discovered entity with missing taxonomic relations, possible taxonomic relations can be automatically determined or proposed based on this ontology panorama.

[0112] An example used to find alternatives to a chemical substance is shown, illustrating substitutions for “Pluriol P900” (polypropylene glycol with a molar mass of 900 g / mol). An automated search in a chemical database would display up to 20 results, such as “Pluriol E1500,” “Pluriol E300,” “Caradol ED110,” or “Pluracol P710R.” Based on the classification of these substances according to their chemical properties, only two substances, “Pluriol E1500” and “Pluriol E300” (polyethylene glycol with molar masses of 1500 g / mol and 300 g / mol, respectively), would be automatically selected as valid alternatives to “Pluriol P900.”

[0113] Another example related to taxonomic knowledge acquisition is obtaining alternatives to "Pluriol P900" with 'good' or 'very good' wetting properties. This substance (i.e., polypropylene glycol with a molar mass of 900 g / mol) is classified as a chemical substance and a polyol. Therefore, one alternative would be "Uniol TG-1000," which, although an additive, is also polypropylene glycol, but with a molar mass of 1000 g / mol.

Claims

1. A computer-implemented method for controlling and / or monitoring production processes and / or production facilities as part of an industrial production chain, wherein, Automatically resolve different taxonomic classifications of chemical, physical, and / or technical entities within an ontology / taxonomic domain provided by at least two participants in the industrial production chain to ensure consistency of these classifications along the production chain. This involves evaluating the underlying information / data of chemical, physical, and / or technical entities within a taxonomic domain provided by the at least two participants in the production chain regarding their semantic and / or technical meaning, and determining the taxonomic relationship between the underlying information / data of the at least two participants based on the evaluation results of the semantic and / or technical meaning of the underlying information / data of the chemical, physical, and / or technical entities.

2. The method according to claim 1, wherein, The assessment of the semantic and / or technical meaning of the underlying information / data of these chemical, physical, and / or technical entities is based on predefined classes and / or subclasses of these entities.

3. The method according to claim 1 or 2, wherein, These chemical, physical, and / or technical entities are represented through corresponding databases, reference lists, semantic models, and / or digital twins.

4. The method according to claim 3, wherein, Provide ontology or taxonomy templates for these chemical, physical, and / or technical entities based on a predefined ontology / taxonomy of the underlying processes and / or products.

5. The method according to one or more of claims 2 to 4, wherein, Classification utilizes processing methods based on artificial intelligence, machine learning, or rules to determine similarity based on the technical and / or semantic meaning of technical terms.

6. The method according to claim 5, wherein, In the case of machine learning methods, existing classification methods are used to preclassify existing data related to production processes, materials, etc., for at least one participant in the production chain.

7. The method according to one or more of the preceding claims, wherein, In the absence of a classification method, semantic text recognition based on natural language processing methods is performed.

8. The method according to claim 7, wherein, The extract-transform-load process performs data integration for production chain participants, combining data from multiple data sources into a single, consistent data store.

9. The method according to one or more of the preceding claims, wherein, These chemical, physical, and / or technical entities are key performance indicators of underlying industrial processes or products.

10. An apparatus for automatically resolving different taxonomic classifications of chemical, physical, and / or technical entities within a taxonomic domain provided by at least two participants in an industrial production chain to ensure consistency of these taxonomic classifications along the production chain, the apparatus comprising a processing unit for evaluating underlying information / data of chemical, physical, and / or technical entities within a taxonomic domain provided by the at least two participants in the production chain with respect to their semantic and / or technical meaning, and for determining a taxonomic relationship between the underlying information / data of the at least two participants based on the evaluation results of the semantic and / or technical meaning of the underlying information / data of the chemical, physical, and / or technical entities.

11. A distributed manufacturing / production chain system for automatically resolving different taxonomic classifications of chemical, physical, and / or technical entities within a taxonomic domain provided by at least two participants in an industrial production chain, thereby ensuring consistency of these taxonomic classifications along the production chain, the system comprising one or more computing nodes configured to evaluate underlying information / data of chemical, physical, and / or technical entities within a taxonomic domain provided by the at least two participants in the production chain with respect to their semantic and / or technical meaning, and to determine taxonomic relationships between the underlying information / data of the at least two participants based on the evaluation results of the semantic and / or technical meaning of the underlying information / data of the chemical, physical, and / or technical entities.

12. A data structure stored in a computer-readable storage medium, the data structure including a data format for dynamically resolving different taxonomies, the data format being independent of the different taxonomies provided in a distributed manufacturing / production chain.

13. The data structure of claim 12, comprising an ontology or taxonomy template for these chemical, physical, and / or technical entities based on a predetermined ontology / taxonomy of the underlying processes and / or products.

14. A computer program product for generating ontology / classification-based digital twins of underlying industrial products or production processes using templates based on predetermined ontology or taxonomy of the underlying product.

15. The computer program product according to claim 14, wherein, The digital twin is described semantically, wherein corresponding elements or objects of the underlying production chain are classified, and wherein the digital twin is dynamically modified along the production chain according to different ontologies or taxonomy of the participants in the production chain.

16. The use of the method according to claims 1 to 9 and / or the data structure according to claims 12 and 13 and / or the digital twin according to claim 14 or 15 for controlling and / or monitoring production processes and / or production facilities as part of an industrial production chain, wherein, Automatic and / or dynamic resolution of different taxonomic classifications of chemical, physical, and / or technical entities within the taxonomic domain provided by at least two participants in the industrial production chain, so as to keep these taxonomic classifications consistent along the production chain.

17. A computer element having instructions which, when executed on one or more computing nodes, are configured to perform the steps of the method of any one of claims 1 to 9, or are configured to be executed by the device of claim 10 or the system of claim 11.

Images