EP4673799A1

EP4673799A1 - Enabling and optimizing production in the one or more plants

Info

Publication number: EP4673799A1
Application number: EP23719979.9A
Authority: EP
Inventors: Rafael Blumenfeld; Stephan Grimm; Ronald Lange; David Michaeli; Patrick Jon Milligan; Jörn PESCHKE; Wolfgang Schlögl
Original assignee: Siemens Industry Software Inc
Current assignee: Siemens Industry Software Inc
Priority date: 2023-03-30
Filing date: 2023-03-30
Publication date: 2026-01-07
Also published as: WO2024205587A1; CN121153014A

Abstract

A method of generating a master recipe including resources and describing a process flow for a specific production in a specific plant includes obtaining a general recipe including information related to the process flow without identifying the resources to be used to perform the process. The method includes: determining required capabilities of the general recipe in the form of a first semantic graph; determining provided capabilities of at least one plant in the form of a second semantic graph labeled with the capabilities of resources of the one or more plants and/or based on an equipment topology of the respective plant also in the industry specific language or ontology; semantically matching the required capabilities with the provided capabilities by identifying one or more sub-graphs in the second semantic graph comprising the resources for carrying out the process flow; and outputting the one or more sub-graphs as a master recipe.

Description

ENABLING AND OPTIMIZING PRODUCTION IN THE ONE OR MORE PLANTS

TECHNICAL FIELD

[0001] The present disclosure relates to apparatuses and methods for automating production processes. More particularly, the present disclosure relates to equipment control software that allows controlling a process.

BACKGROUND

[0002] A process flow is a sequence of chemical, physical, and/or biological activities for the conversion, transport, or storage of material or energy. Process controllers manipulate hardware to ensure that a process flow is completed in a satisfactory manner. Prior art process controllers present to an operator information directly related to the hardware. Present-day process control systems use instruments, control devices, and/or communication systems to monitor and/or manipulate controlled elements, such as valves and switches, and to control the values of one or more process variables, including temperature, pressure, flow, etc. The process variables are selected and controlled to achieve a desired process objective, such as attaining the safe and/or efficient operation of machines and equipment utilized in the process. Process control systems have widespread application in the automation of industrial processes such as the processes used in chemical, petroleum, and manufacturing industries. [0003] For example, US patent US7369913B2 describes apparatuses and methods for automating a sequence of basic chemistry operations through a recipe. Therein, a workstation software uses a recipe to control the process. The recipe is expressed in terms of the process. The controller runs software routines for process control and for hardware control. SUMMARY AND DESCRIPTION

[0004] Today’s customers, for example, in the consumer packaged goods industries are struggling with transformation of an industrial process that was built based on a research formula, to the actual factories. Using consumer packaged goods terminology, the problem is to transfer a general recipe to a master recipe that represents a plant specific implementation. This plant specific implementation contains the equipment and/or volumes of the ingredients to be processed per specific batch size.

[0005] It is very tedious work when process engineers from the laboratory are trying to understand the existing equipment in plants, how they are connected, and what is their characteristics, in order to express their generic process in terms of the plants. This task is to be repeated for each plant that is intended to produce the general recipe due to differences in the factories, batch sizes, regulations, target market, supplies, etc.

[0006] Thus, a (semi-)automatic creation and/or validation when transferring a (general) recipe to factories without manual intervention is desired. Further, a specific language that provides that a recipe may be implemented in some or all production facilities (e.g., of the same company) and, for example, without developing a specific software for the company may be needed.

[0007] According to a first aspect, a method of generating a master recipe including resources and describing a process flow for a specific production in a specific plant is provided. The method includes the steps of obtaining a general recipe including information related to the process flow without identifying the resources to be used to perform the process, and determining required capabilities of the general recipe in the form of a first semantic graph. The first semantic graph is labeled with relations and concept names of and between ingredients, volumes, processes in an industry specific language/ontology. The method includes determining provided capabilities of at least one plant in the form of a second semantic graph. The second semantic graph is labeled with the capabilities of resources of the one or more plants and/or based on an equipment topology of the respective plant also in the industry specific language/ontology. The method includes semantically matching the required capabilities with the provided capabilities by identifying one or more sub-graphs in the second semantic graph including the resources for carrying out the process flow, and outputting the one or more sub-graphs as a master recipe.

[0008] According to a second aspect, a method of generating an equipment control code in an equipment control language based on a process description/master recipe is provided. The method includes the steps of determining a process graph of the master recipe in the form of a first semantic graph. The first graph is labeled with relations and concept names of and between ingredients, volumes, processes in an industry specific language/ontology and describing a skill sequence. The method includes determining a capability model of at least one plant in the form of a second semantic graph. The second semantic graph is labeled with the capabilities of resources of the one or more plants and/or based on an equipment topology of the respective plant also in the industry specific language/ontology and is describing skills of the one or more equipment in the plant. The method includes determining an execution sequence of the skills of the one or more equipment by semantically matching the process model with capability model, and identifying one or more application programming interfaces (APIs) representing the skills of the one or more equipment. Each of the APIs includes one or more parameters. The method includes searching the parameter space of the APIs in order to find an interoperable API sequence and outputting the interoperable API sequence as an equipment control code.

[0009] According to a third aspect, a production in the one or more plants is initiated and/or running based on the master recipe and/or the equipment control code obtained according to the first aspect and/or the second aspect.

[0010] According to a fourth aspect, an apparatus operative to perform the method steps of the first aspect and/or the second aspect is proposed.

BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 illustrates required capabilities in the form of a first semantic graph that is created based on a domain ontology and a general recipe.

[0012] Figure 2 illustrates provided capabilities in the form of a second semantic graph that is created based on a capability ontology and an equipment topology of one or more plants.

[0013] Figure 3 illustrates an identification of sub-graphs in the second semantic graph based on the required capabilities.

[0014] Figure 4 illustrates the identification of one or more sub-graphs within the second graph, the mapping of a labelling of the sub-graphs onto the first semantic tree representing the required capabilities, and the resulting description of the generic recipe in a terminology of the capability ontology.

[0015] Figure 5 illustrates the generation of a process graph of a master recipe and a capability model of at least one plant based on an industry specific language/ontology. [0016] Figure 6 illustrates the matching of the process model with capability model.

[0017] Figure 7 illustrates the identification of an execution sequence of the skills of the one or more equipment.

[0018] Figure 8 illustrates the identification of one or more application programming interfaces representing the skills. DETAILED DESCRIPTION

[0019] Customers in consumer packaged goods (CPG) industries are struggling with transformation of an industrial process that was built based on a research formula, to the actual factories. Consumer packaged goods are items that require routine replacement or replenishment, such as food, beverages, clothes, tobacco, makeup, and household products. [0020] The research formula may be given in the form of a general recipe. Using CPG terminology, the problem is to transfer a general recipe to a master recipe that represents a plant specific implementation. The plant specific implementation may contain the equipment and/or volumes of the one or more ingredients to be processed per specific batch size.

[0021] Today, it is very tedious work when process engineers (e.g., from the laboratory) are trying to understand the existing equipment in one or more plants, how they are connected, what their characteristics are, to express their generic process in terms of the plants. As mentioned, the work may need to be repeated for each plant that shall produce the general recipe due to differences in the factories, batch sizes, regulations, target market, and/or supplies, etc.

[0022] The proposed solutions allow an automatic match and/or validation of transferring a general recipe to one or more factories, especially without or with minimal manual intervention. Hence, methods and/or systems that help create a company specific language that provides that a recipe defined in the terms of that language may be implemented in one or more production facilities (e.g., plant and/or factories) of the company are provided.

[0023] Turning to Figure 1, a general or generic recipe 1 is provided. Based on the general recipe 1, a semantic tree 2 (e.g., including ingrediencies, volumes, processes (transformation) in industry specific terms, with industry specific constraints) that represents the general recipe 1 is generated. The semantic tree 2 is also referred to as a process graph in the figures or semantic graph in general. This representation allows to express the process steps in terms of required capabilities (e.g., the needed equipment types, categories, characteristics, and/or usage, such as without designating a specific equipment). This representation may be given textually (e.g., via a Resource Description Framework (RDF) description, such as via a graphical application or any other way that allows to describe such representation). For example, a data model for metadata including/describing the generic recipe may be given, for example, by the Resource Description Framework (RDF) standard. RDF is a directed graph composed of triple statements. An RDF graph statement is represented by: 1) a node for the subject; 2) an arc that goes from a subject to an object for the predicate; and 3) a node for the object. Each of the three parts of the statement may be identified by a Uniform Resource Identifier (URI).

[0024] Further, an ontology, such as a domain ontology 3 that describes all of the relevant concepts in a single domain of interest may be given. For example, such an ontology 3 may be given or derived from a standard such as S88, shorthand for ANSI/ISA-88, which addresses batch process control. Therein, ANSI/ISA-88.01-2010 Batch Control Part 1 provides Models and terminology, ANSI/ISA-88.00.02-2001 Batch Control Part 2 provides Data structures and guidelines for languages, ANSI/ISA-88.00.03-2003 Batch Control Part 3 provides general and/or site recipe models and representation, ANSVISA-88.00.04-2006 Batch Control Part 4 provides Batch Production Records, and ISA- TR88.00.02-2015 Machine and Unit States provides an implementation example of ISA-88. Thus, such an ontology 3 (e.g., domain ontology) provides a consistent set of standards and/or terminology for control (e.g., batch control) and/or defines the physical model, procedures, and recipes. Alternatively and/or additionally, ANSI/ISA-95, or ISA-95 may be used; alternatively, another ontology that provides consistent terminology that may serve as a foundation for supplier and manufacturer communications, provide consistent information models, and provide consistent operations models that are a foundation for clarifying application functionality and/or how information is to be used.

[0025] A general recipe 1 may include information related to a specific process flow for the production of a product. A general recipe may also include definitions of resources such as equipment that is deployed to perform the process flow, as well as materials input to perform the process flow and/or output materials resulting from the execution of the process flow. [0026] In general, different types of recipes are known: a general recipe 1 includes information related to the process flow without necessarily identifying the resources to be used to perform the process (flow); a site recipe includes site-specific information with local constraints; and a master recipe includes resource capabilities and describes the recipe for a specific production (e.g., on a specific production line or other production unit).

[0027] Returning to Figure 1, the resulting process graph 2 provides a sequence of steps that are to be carried out in order to implement the general recipe 1. The generic recipe 1 may be semantically matched to and/or enriched with the terms included in the ontology 3. As a result of this, the process graph 2 is obtained. This process graph 2, as well as the (e.g., terms of) ontology and/or the generic recipe may be stored in a database (e.g., in a non-volatile memory).

[0028] Now turning to Figure 2, for each plant 4 or factory, an equipment topology and/or characteristics (e.g., in the form of a plant model 4 in terms of skills and/or constraints) may be obtained. The skills refer to the capabilities (or skills) of the equipment in the plant or factory. Based thereon, a graph (e.g., second semantic graph) or tree 5 including provided capabilities may be created. To that end, the same terminology that was used when creating the required capabilities may be used. As before, the terms or terminology may be provided in the form of an ontology 3 (e.g., capability ontology). In Figure 2, this ontology is referred to as capability ontology 3. Again, the information may be imported based on various industrial standards such as ISA88/95, as described above, or authored via textual or graphic user interfaces.

[0029] As a result, a second semantic graph 5 representing or including the provided capabilities of the one or more plants or factories may be obtained. In one embodiment, a plurality of plants or factories are considered. In other words, the equipment and/or equipment topology of the plurality of plants 4 or factories are obtained, and these equipment topologies are included in the second semantic graph 5.

[0030] As shown in Figure 3, the plant(s) or factory(ies) that are capable of the required capabilities and/or possess the provided capabilities to fulfill the required capabilities (e.g., that may do the job) are identified. The provided capabilities (and the corresponding equipment) may be expressed in terms of virtual units 6, 7. Hence, one or more equipment in the one or more plants may be grouped in the form of virtual unit 6, 7. The virtual unit 6, 7 is able to execute at least a part of the process steps as required by the general recipe. Further, a designation or name is provided to each virtual unit 6, 7, or at least to a number of the virtual units 6, 7. A suggestions for such name or designation may be provided based on a classification system for products and services (e.g., an industrial classification such as eCl@ss) that may be based on industrial standards. Such a classification system may support the digital exchange of product descriptions and service descriptions, based on standardized data formats based on IEC 61360. The classification system may be used in engineering tools as base for transfer of planning data, in ERP systems as base for product master data, and/or as base for exchange of product data. Hitherto, the process graph (e.g., first semantic graph/tree) may be matched to the second semantic graph. In other words, the provided capabilities of the second semantic graph are matched with the required capabilities of the first semantic graph. To that end, a semantic matching may be performed. Semantic matching is a technique to determine whether two or more elements have similar meaning. Hence, the one or more nodes and/or edges (e.g., carrying the semantic information) of the first semantic graph may be compared (e.g., semantically matched) to the one or more nodes and/or edges (e.g., carrying the semantic information) of the second semantic graph.

[0031] As illustrated in Figure 4, the provided capabilities (e.g., in form of the second semantic graph 5) may thus be mapped to one or more sub-trees (e.g., minimal sub-trees), and/or a name may be assigned for each subtree. For explanatory purposes, a sub-tree is referred to but a sub-graph in general may be used. As shown in Figure 4 the (minimal) subtree of a virtual unit 6 is identified. The (minimal) sub-tree of that virtual unit 6 is extracted from the second semantic graph 5 and thus may be considered separately and/or in isolation from the other process steps within the second semantic graph 5 including the provided capabilities. The name and/or denomination of the virtual unit may then become part of a language to be used for executing the master recipe (e.g., in the one or more plants, such as of the enterprise). The method allows to express the general recipe with terms that are defined for the one or more plants, factories, and/or enterprises that are capable of carrying out and thus implementing the general recipe. In case a new recipe that, for example, has different ingredients or requires new processes is obtained, the steps as described in herein may be repeated. Thereby, an enterprise language may be created and/or extended.

[0032] It is an advantage of the aspects disclosed herein that the automation of the customization process and/or the definition of customer specific processes is enabled, in particular without developing a dedicated software, by capturing knowledge of a company in a semantic way. The process is repeatable and extendable, still, without any software changes. Hence, definition and/or extension of a company specific language is enabled, which ensures the executability of a general recipe on various plants.

[0033] The steps as described herein may be implemented and thus carried out by an integrated development environment (IDE). An IDE is a software application that enables computer programming for software development. An IDE may include a source code editor, build automation tools, and a debugger. Some IDEs, such as NetBeans and Eclipse, contain the necessary compiler, interpreter, or both. The IDE may also include a version control system and/or various tools or routines to carry out the steps described herein.

[0034] Turning to Figure 5, further aspects are described. As mentioned herein, today’s customers in CPG industries are struggling with the transformation of an industrial process that was built for a specific plant into an automation control software to execute this process. The challenge lies in a translation of various process requirements to an actual equipment control language that keeps the various devices in a plant coordinated. The control language of many devices is different, and even when it may be controlled via OPC-UA, specific parameters and values may be required. Manually performing this task is tedious and requires mapping the process specifications into control sequences of many control devices.

Currently, it is manual work that involves control engineers and process engineers to describe the process and comprehend it in terms of the control language(s). Even further, after this mapping, there is another manual task for the control engineer to translate it to the control programs (e.g., using an IDE). Thereafter, the control program is to be validated before the control program can be deployed.

[0035] It is thus proposed to generate the one or more control programs directly from the process description of the plant specific process (e.g., the master recipe; obtained as described herein, such as in connection with any one of the Figures 1 to 4). The plant specific process manifests the capabilities the plant specific process requires that are using the “language” (i.e., terminology) of the process. It can be generated from a more virtual process or from the master recipe 8. The master recipe 8, as, e.g., shown in Figure 5, may include resource capabilities and/or may describe the recipe for a specific production, e.g., in a specific plant or factory and for example a specific production line therein. The master recipe 8 may also describe the process for a specific production on specific hardware. A batch description may be received based on which a master recipe 8 is identified. Furthermore, as shown in Figure 5, a plant description or plant model 9 may be obtained. The plant description or plant model 9 can be based on standards such as ISA88/95, as described herein, and/or a skill modeling description.

[0036] As shown in Figure 5, based on required capabilities, and for example additional process constraints, a first semantic graph, also referred to as process graph 12 in Figure 5, comprising actions and/or properties of the master recipe is created. To that end, skill semantics 10, or an ontology in general, may be used. The same ontology may be used to describe a plant, factory and/or the equipment therein. The process graph 12 or semantic graph or semantic tree may thus comprise a skill sequence or may be used to determine a skill sequence. To that end, the process graph 12 is created based on the master recipe and may be enriched and/or semantically matched with the ontology or skill semantics 10.

[0037] Furthermore, a capability model 11 of the plant or plant model 9 may be obtained. The plant model may also be enriched based on the skill semantics 10 or an ontology in general. Hence, the capability model may comprise a description of a plant and associated skills (e.g., from the skill description). Hitherto, semantical mapping of the skill semantics or the ontology to the plant model may be performed.

[0038] The plant model 9 and/or the representation in the form of the capability model may comprise one or more skill models of the one or more equipment in the plant. A skill model may be a model of an equipment in the plant that comprises the skills of the equipment. These skill models of the one or more equipment may comprise definitions of one or more APIs. These skill models and/or APIs may correspond to the provided capabilities, i.e., provided by the plant. These skill models can be mapped (or to put it differently, semantically matched) to the required capabilities of the master recipe. Thus, a mapping between required capabilities and provided capabilities is obtained.

[0039] As shown in Figure 6, from provided capabilities, e.g., in the form of said capability model 11, a matched graph may be identified which comprises properties, e.g., from the plant description and/or the skill semantics, of the capability model 11 associated with the matched graph. Hitherto, the process graph 12 is matched or mapped, e.g., using semantical matching, to the capability model 11, or vice versa. As a result, a matched graph 13 may be obtained. [0040] Turning to Figure 7, based on associated skills the one or more APIs, Application Programming Interface, associated with the defined properties may be identified. Thereupon, one or more perturbations may be generated to find a complete coverage with the correct sequence of properties. The perturbations may use one or more stochastic techniques such as genetic algorithms or simulated annealing or another stochastic solution search approach. As a result, a sequence 14 of APIs and/or API calls and corresponding API parameters are obtained. Hence, the sequence 14 and/or parameters of API (calls) are obtained (more than one possible solution can be ranked and presented to user too select). This sequence 14 reflects the process with the specific equipment may be generated as shown in Figure 8. As shown in Figure 9, based on the API sequence 14 an OPC-UA sequence, e.g., in the form of an OPC-UA code 15, may be created. The OPC-UA sequence or OPC code 15 may then be validated via simulation in a simulation tool. It should be understood that the OPC UA code may serve as a control program or that a control program may be created based on the OPC UA code. Thus, a control program which corresponds to the Master recipe, and for example comprises a specific order of process steps and/or skills to be applied, may be obtained.

[0041] It is thus proposed to use semantic mapping with skill based model and skill based interface definitions for equipment in order to generate a control program.

[0042] The elements and features recited in the appended claims may be combined in different ways to produce new claims that likewise fall within the scope of the present invention. Thus, whereas the dependent claims appended below depend from only a single independent or dependent claim, it is to be understood that these dependent claims may, alternatively, be made to depend in the alternative from any preceding or following claim, whether independent or dependent. Such new combinations are to be understood as forming a part of the present specification.

[0043] While the present invention has been described above by reference to various embodiments, it should be understood that many changes and modifications can be made to the described embodiments. It is therefore intended that the foregoing description be regarded as illustrative rather than limiting, and that it be understood that all equivalents and/or combinations of embodiments are intended to be included in this description.

Claims

PATENT CLAIMS

1. A method of generating a master recipe (8) including resources and describing a process flow for a specific production in a specific plant, the method comprising: obtaining a general recipe (1) including information related to the process flow without identifying the resources to be used to perform a process; determining required capabilities of the general recipe (1) in a form of a first semantic graph (2), wherein the first semantic graph (2) is labeled with relations and concept names of and between ingredients, volumes, processes in an industry specific language or ontology (3); determining provided capabilities of at least one plant in a form of a second semantic graph (5), wherein the second semantic graph (5) is labeled with capabilities of resources of the at least one plant and/or based on an equipment topology of the respective plant also in the industry specific language or ontology (3); semantically matching the required capabilities with the provided capabilities, the semantically matching comprising identifying one or more sub-graphs (6, 7) in the second semantic graph (5) comprising the resources for carrying out the process flow; and outputting the one or more sub-graphs (6, 7) as a master recipe (8).

2. The method of claim 1, further comprising: determining separate graphs for each plant of a plurality of plants (4); and assembling the separate graphs into the second semantic graph (5).

3. The method of claim 2, further comprising: assigning a labeling to the one or more sub-graphs (6, 7), the one or more sub-graph (6, 7) representing a virtual unit comprising the resources for carrying out the process flow.

4. The method of claim 3, wherein the virtual unit possesses capabilities to fulfill the requirements of one or more process steps of the general recipe (1).

5. The method of claim 3, further comprising: determining a sub-graph (6, 7) for a number of plants (4) of the plurality of plants (4) in the second semantic graph (5).

6. The method of claim 5, further comprising: employing the same labeling for each of the sub-graphs (6, 7).

7. The method of claim 6, further comprising: determining the one or more sub-graphs (6, 7) based on an optimization goal for carrying out the process flow of the generic recipe (1), wherein the optimization goal relates to a number of resources, a time to produce, and/or a usage of transport equipment.

8. The method of claim 7, wherein the optimization goal relates to the usage of the transport equipment, and wherein the transport equipment includes containers.

9. The method of claim 8, further comprising: ranking the sub-graphs (6, 7) in an order according to an extent the optimization goal is achieved.

10. A method of generating an equipment control code (15) in an equipment control language based on a process description and/or master recipe (8), the method comprising: determining a process graph of the master recipe (8) in a form of a first semantic graph (12), wherein the first semantic graph (12) is labeled with relations and concept names of and between ingredients, volumes, processes in an industry specific language/ontology, and describing a skill sequence; determining a capability model of at least one plant in a form of a second semantic graph (11), wherein the second semantic graph (11) is labeled with the capabilities of resources of the at least one plant and/or based on an equipment topology of the respective plant also in the industry specific language/ontology, and describes skills of one or more equipment in the plant; determining an execution sequence (13) of skills of the one or more equipment, the determining of the execution sequence (13) of the skills of the one or more equipment comprising semantically matching a process model with the capability model; identifying one or more application programming interfaces (APIs) representing the skills of the one or more equipment, each of the APIs comprising one or more parameters; searching a parameter space of the APIs in order to find an interoperable API sequence (14); and outputting the interoperable API sequence (14) as an equipment control code (15).

11. The method of claim 10, further comprising validating the equipment control code (15), the validating of the equipment control code comprising running a simulation of the plant based on the equipment control code.

12. The method of claim 11, further comprising searching the parameter space of the APIs based on a stochastic, preferably genetic, algorithm.

13. Initiating and/or running production in the one or more plants (4) based on the master recipe (8) according to any one of the preceding claims.

14. An apparatus, preferably comprising a processor and a memory, operative to perform the method steps of any one of the preceding claims.

15. A software application, such as an integrated development environment, preferably stored on a non-transitory medium, operative to perform the method steps of any one of the claims 1 to 13.