Method and platform for deploying industrial applications on edge computing devices of machine tools

Abstract

Method and platform for deploying industrial applications on an edge computing device of a machine tool. A deployment platform for deploying industrial applications on an edge computing device ECD connected to a controller of a machine tool MT, the deployment platform comprising: a model management component MMC, which is adapted to instantiate a general machine tool model GMTM stored in a memory based on a machine tool data report MTDR received by the model management component MMC from the edge computing device ECD of the corresponding machine tool MT to provide a machine instance model MIM of the corresponding machine tool MT, and is further adapted to convert general data requirements gR of a general industrial application into machine tool specific requirements mtsR using the machine instance model MIM of the corresponding machine tool MT, wherein the general industrial application is instantiated by extending its configuration data with the machine tool specific requirements mtsR or by converting requirements at runtime, providing an instantiated industrial application of the machine tool MT, wherein the instantiated industrial application is deployed by the model management component MMC of the deployment platform on the edge computing device ECD of the corresponding machine tool MT.

Description

Technical Field

[0001] The present invention relates to a method and a platform for deploying industrial applications on an edge computing device connected to a controller of a machine tool in a manufacturing facility. Background Art

[0002] With the advent of cloud computing, industrial assets (such as motors, robots, industrial equipment, automation equipment) within a manufacturing facility can be connected to a cloud platform via an agent such as an IoT gateway. Digitalization and the development of related technologies such as the Internet of Things, cloud computing, and artificial intelligence can be used in a manufacturing environment, giving rise to smart factories where processes can be optimized with the help of data-driven decision-making. In many cases, decisions must be made near the data source to reduce response times, address privacy issues, and / or process large amounts of data. Therefore, edge computing devices are collocated with physical systems in a manufacturing environment. However, the computing devices process data generated by the physical systems to provide value-added functions, particularly diagnostics, process monitoring and process optimization, as well as predictive maintenance, security monitoring, etc. This can be based on a variety of underlying technologies, such as complex event processing, machine learning, or reasoning.

[0003] These value-added functions can generally be deployed as reusable software components, i.e., industrial applications. This forms an effective approach whereby the core functions are programmed only once and then can be deployed to multiple manufacturing environments that can perform most equivalent tasks. A manufacturing facility can include different machine tools controlled by associated controllers. For example, a machine tool in a manufacturing environment can be provided to machine a metal workpiece.

[0004] However, due to the heterogeneity of manufacturing environments, it is difficult to realize this vision because different device configurations can be used in each factory to perform tasks. For example, different series or versions of digital controllers (even from different manufacturers) can be used. In addition, machines can be configured in different ways (e.g., with different numbers of axes).

[0005] This makes it difficult to create a general-purpose application that can work in different manufacturing environments. So far, application developers can provide multiple versions of industrial applications suitable for different manufacturing environments, as Figure 1As shown. A user or developer D can release different versions V1, V2 of the industrial application. In a typical setup, different manufacturing environments can be connected to a common set of remote (e.g., cloud-based) computing services, namely the so-called backend BE. The backend BE can provide different services common to all managed manufacturing environments. In this setup, the application developer or user D needs to provide different versions V1, V2 of the application APP suitable for different machine configurations (e.g., released to the backend). The backend BE can then decide which version of the industrial application to execute on a given manufacturing environment ME. The different versions of the released industrial application can be deployed on the edge computing device ECD of the corresponding manufacturing environment ME, also as Figure 1 As shown. The drawback of this traditional approach is that different machine configurations of the machines or machine tools MT within the manufacturing environment ME are required to produce multiple versions of the industrial application for different environments.

[0006] A more advanced conservative approach is to use standard protocols for industrial automation, such as OPC Unified Architecture (OPC-UA). Such protocols can provide a unified advertisement processing mechanism for data received from different types of devices. Assuming that all controllers of the machine tool MT implement such a common protocol, the app developer D can provide a single application for all manufacturing environments ME. However, programming a general application by resorting to a common protocol has many limitations of its own. These applications are not compatible with devices and machines using other protocols. This may prevent the application from accessing all the data available in the manufacturing environment ME. In addition, the application is developed based on a priori assumptions about the capabilities of the target manufacturing environment. For example, the protocol can use a general static information model. In addition, there is no mechanism for the developer D to learn about the specific characteristics of some potential target manufacturing environments ME, such as data points or additional sensor devices. In addition, industrial applications do not know about each other, so it is difficult for independent developers to gradually build complex functions based on other applications.

[0007] Therefore, there is a need to provide a method and platform that allows the deployment of industrial applications without binding the applications to a unique protocol or information model. Summary of the Invention

[0008] According to a first aspect of the present invention, this object is achieved by a method for deploying industrial applications comprising the features of claim 1.

[0009] According to the first aspect, the present invention provides a method for deploying an industrial application on an edge computing device connected to a controller of a machine tool, wherein the method comprises the following steps:

[0010] Providing a general machine tool model of machine tool components, component configurations, and / or dynamic data items,

[0011] Convert the machine tool data report extracted by the edge computing device from the controller of the machine tool into a machine instance model of the machine tool that instantiates the general machine tool model.

[0012] Use the machine instance model of the machine tool to convert the general data requirements of the general industrial application into machine tool specific requirements.

[0013] Instantiate the general industrial application by extending its configuration data with the machine tool specific requirements or converting subset requirements or all requirements during the operation of the machine tool, providing an instantiated industrial application of the machine tool, and deploy the instantiated industrial application on the edge computing device of the machine tool.

[0014] The method according to the first aspect of the present invention allows customization and extension. It provides and maintains a common general machine tool model GMTM, which is extensible and can be used as a basis for developing industrial applications. The general machine tool model GMTM can form a superset of existing machine tool data standards and can provide a scheme for describing drive parameters and kinematic transformations of the machine tool MT. The method can automatically deploy general industrial applications to a specific manufacturing environment ME using different machines, configurations, and data protocols. The method can automatically generate an instance of the general machine tool model GMTM corresponding to the specific manufacturing environment ME.

[0015] The method allows a developer D to create a general industrial application that can be deployed on multiple heterogeneous manufacturing environments ME, thereby fully utilizing the capabilities of those manufacturing environments and the functions provided by other applications. The possibility of creating general industrial applications for different manufacturing environments ME greatly improves the efficiency and usability of the deployment platform.

[0016] In a possible embodiment of the method according to the first aspect of the present invention, the machine instance model of the machine tool is enriched with OEM specific data of the original equipment manufacturer of the corresponding machine tool.

[0017] In another possible embodiment of the method according to the first aspect of the present invention, the general machine tool model and the machine instance model of the machine tool are stored in the memory or database of a model management component within a backend connected to the edge computing device.

[0018] In another possible embodiment of the method according to the first aspect of the present invention, the general machine tool model provides a scheme suitable for describing the configuration, kinematics, and associated data items of machine tool components.

[0019] In another possible embodiment of the method according to the first aspect of the present invention, the general data requirements of the general industrial application include queries.

[0020] In a possible embodiment of the method according to the first aspect of the present invention, a query for a general industrial application is evaluated by a query engine using a rich machine instance model of a machine tool to generate a query result that forms the machine tool specific requirements.

[0021] In another possible embodiment of the method according to the first aspect of the present invention, the general data requirements of a general industrial application that are converted into machine tool specific requirements using a rich machine instance model of a machine tool include static data requirements that are converted at deployment and dynamic data requirements that are converted during application runtime when the industrial application is executed on the processor of an edge computing device.

[0022] In another possible embodiment of the method according to the first aspect of the present invention, a machine tool data report extracted by an edge computing device from a controller of a machine tool includes a predefined format that includes a controller specific data report format or an industrial standard format, and includes a mapping from the predefined format to a general machine tool model stored in a database of a model management component at the backend.

[0023] In another possible embodiment of the method according to the first aspect of the present invention, a general machine tool model stored in a memory or a database at the backend is extended by a general application model of a general industrial application.

[0024] In another possible embodiment of the method according to the first aspect of the present invention, a machine instance model of a machine tool is enriched by applying rules specified in a general industrial application.

[0025] In another possible embodiment of the method according to the first aspect of the present invention, an instantiated industrial application with its configuration data extended with machine tool specific requirements is deployed on an edge computing device of a machine tool, wherein the edge computing device uses the machine tool specific requirements and / or general data requirements to obtain data from a data source of the machine tool via a machine tool interface of the edge computing device.

[0026] In another possible embodiment of the method according to the first aspect of the present invention, a copy of the machine instance model of a machine tool is stored in a local memory of an edge computing device of the corresponding machine tool, wherein the machine instance model copy is used by an industrial application to convert general data requirements into machine tool specific requirements.

[0027] In another possible embodiment of the method according to the first aspect of the present invention, the machine instance model copy is updated whenever the machine instance model of a machine tool stored in the memory of a model management component is changed.

[0028] In another possible embodiment of the method according to the first aspect of the present invention, data quality predicates are used to qualify the dynamic data requirements of an industrial application with respect to data obtained by an edge computing device from a data source of a machine tool.

[0029] In a possible embodiment of the method according to the first aspect of the present invention, the data quality predicates include

[0030] data acquisition frequency,

[0031] probability of data correctness,

[0032] data precision,

[0033] data integrity,

[0034] data currency,

[0035] historical data availability range, and

[0036] data security constraints.

[0037] In another possible embodiment of the method according to the first aspect of the present invention, the general data requirements of a general industrial application expressed according to a general machine tool model are converted into machine tool specific requirements, including data points and / or data protocols of the machine tool, its controller, and its edge computing device.

[0038] In another possible embodiment of the method according to the first aspect of the present invention, after successfully checking that the corresponding edge computing device meets the machine tool specific requirements, the industrial application instance is deployed on the edge computing device of the machine tool.

[0039] According to a further second aspect, the present invention further provides a deployment platform including the features of claim 15.

[0040] According to the second aspect, the present invention provides a deployment platform for deploying an industrial application on an edge computing device connected to a controller of a machine tool,

[0041] The deployment platform includes:

[0042] a model management component adapted to instantiate a general machine tool model stored in a memory based on a machine tool data report received by the model management component from an edge computing device of a corresponding machine tool to provide a machine instance model of the corresponding machine tool, and further adapted to convert general data requirements of a general industrial application into machine tool specific requirements using the machine instance model of the corresponding machine tool,

[0043] wherein, by extending its configuration data with machine tool specific requirements or converting subset requirements or all requirements during machine tool operation, the general industrial application is instantiated to provide an instantiated industrial application of the machine tool,

[0044] wherein the instantiated industrial application is deployed on the edge computing device of the corresponding machine tool by the model management component of the deployment platform. Description of the Drawings

[0045] In the following, possible embodiments of different aspects of the present invention are described in more detail with reference to the accompanying drawings.

[0046] Figure 1 Schematically shows a conventional method for deploying industrial applications on edge computing devices;

[0047] Figure 2 Shows a schematic diagram for illustrating a possible exemplary embodiment of a deployment platform for deploying industrial applications on edge computing devices according to one aspect of the present invention;

[0048] Figure 3 Shows a flowchart of a possible exemplary embodiment of a method for deploying industrial applications on edge computing devices according to one aspect of the present invention;

[0049] Figure 4 Schematically shows an example of a general machine tool model with a machine instance model;

[0050] Figure 5 Shows an example of automatically generating a machine instance model from a SINUMERIK data archive;

[0051] Figure 6 Shows the enrichment of a machine instance model of a machine tool with OEM-specific data;

[0052] Figure 7 Shows a flowchart for illustrating a possible exemplary embodiment of deploying an instantiated industrial application on an edge computing device;

[0053] Figure 8 Schematically shows a query evaluation in which a query engine uses a machine instance model of a machine tool to generate a query result forming machine tool-specific requirements;

[0054] Figure 9 Schematically shows the deployment of an industrial application instantiated on an edge computing device of a machine tool;

[0055] Figure 10 Shows a flowchart of a possible exemplary embodiment of the deployment process. Detailed Description of the Invention

[0056] Figure 2 Shows a schematic diagram for illustrating possible exemplary embodiments of a method and a platform for deploying industrial applications on one or more edge computing devices ECD of a machine tool MT within a manufacturing facility. In Figure 2In the exemplary embodiment shown, different manufacturing environments ME1, ME2, ME3 are connected to a common backend BE. Each manufacturing environment ME includes at least one machine tool MT, which has a machine tool controller connected to an edge computing device ECD of the corresponding manufacturing environment. The machine tool MT may include mechanical components for processing workpieces during the manufacturing process. The machine tool MT may be connected to at least one machine tool controller, which is adapted to control the mechanical components of the machine tool. The machine tool controller includes an interface to the edge computing device ECD of the manufacturing environment ME to exchange data.

[0057] The backend BE of the deployment platform includes a model management component MMC, which is adapted to instantiate a general machine tool model GMTM stored in the memory or database of the platform. The model management component MMC is adapted to instantiate the stored general machine tool model GMTM based on a machine tool data report MTDR received by the model management component MMC from the edge computing device ECD of the corresponding machine tool MT, to provide a machine instance model MIM of the corresponding machine tool MT. The model management component MMC of the backend BE is further adapted to convert the general data requirements gR of a general industrial application into machine tool specific requirements mtsR using the machine instance model MIM of the corresponding machine tool MT. In Figure 2 In the embodiment shown, each edge computing device ECD is connected to the controller of the machine tool MT. In this particular embodiment, the edge computing device ECD and the machine tool MT form separate physical entities, but are connected via a data interface. The edge computing device ECD itself may also be physically embedded in the machine tool MT, i.e., integrated with other components of the machine tool MT such as the machine tool controller.

[0058] In a possible embodiment, a general industrial application gAPP is instantiated by extending its configuration data with machine tool specific requirements mtsR to provide an instantiated industrial application for the machine tool MT. This can be done if some of the general data requirements gR are selected to be converted into machine tool specific requirements mtsR. In an alternative embodiment, all requirements are converted at runtime of the machine tool MT. The instantiated industrial application of the machine tool MT is deployed by the model management component MMC of the backend BE on the edge computing device ECD of the corresponding machine tool MT. Figure 2 The deployment platform shown is adapted to execute a method for deploying an industrial application on an edge computing device ECD connected to the controller of a machine tool MT within a manufacturing environment, which method in a possible embodiment includes Figure 3 the steps shown.

[0059] In a first step S1, a general machine tool model GMTM of the machine tool components of the corresponding machine tool MT is provided and may be stored in the memory or database of the model management component MMC.

[0060] In another step S2, by instantiating the General Machine Tool Model GMTM, the Machine Tool Data Report MTDR extracted by the Edge Computing Device ECD from the controller of the corresponding machine tool MT is converted into the corresponding Machine Instance Model MIM of the machine tool MT.

[0061] In another step S3, the General Data Requirements gR of the General Industrial Application gAPP can be converted into Machine Tool Specific Requirements mtsR using the Machine Instance Model MIM of the machine tool MT provided in step S2.

[0062] In another step S4, an instantiated industrial application is provided. In a possible embodiment, the General Industrial Application gAPP is instantiated by extending its configuration data with the Machine Tool Specific Requirements mtsR, providing an instantiated industrial application. Additionally, in an alternative embodiment, it is possible to convert subset requirements or all requirements at runtime of the machine tool MT.

[0063] In the final step S5, the instantiated industrial application can be deployed on the Edge Computing Device ECD of the machine tool MT.

[0064] In a possible embodiment, the Machine Instance Model MIM of the machine tool MT generated in step S2 can be enriched with OEM - specific data of the Original Equipment Manufacturer OEM of the corresponding machine tool MT, as also Figure 2 shown. The General Machine Tool Model GMTM stored in the database of the Model Management Component MMC provides a scheme suitable for describing the configuration of machine tool components, their kinematics, and associated data items. Figure 2 The Platform Owner PO of the deployment platform shown in can generate and / or update the General Machine Tool Model GMTM. The General Data Requirements gR of the General Industrial Application can include queries Q. These queries Q can be evaluated by the Query Engine QE of the platform using the Machine Instance Model MIM of the machine tool MT to generate query results QR, thereby forming the Machine Tool Specific Requirements mtsR. The Machine Instance Model MIM can be enriched with OEM - specific data by the OEM of the corresponding machine tool MT, as also Figure 2 shown.

[0065] The General Data Requirements gR of the General Industrial Application can be converted into Machine Tool Specific Requirements mtsR using the enriched Machine Instance Model MIM of the machine tool MT. The Machine Tool Specific Requirements mtsR can include static data requirements converted at deployment and dynamic data requirements converted during application runtime when the industrial application is executed on the processor of the Edge Computing Device ECD in the manufacturing environment.

[0066] In a possible embodiment, the machine tool data report MTDR extracted by the edge computing device ECD from the controller of the machine tool MT in step S2 may include a predefined format. The predefined format may include a controller-specific data report format or an industry standard format. The machine tool data report MTDR may also be accompanied by a mapping from the predefined format to a general machine tool model GMTM stored in the database of the model management component MMC of the backend BE.

[0067] In a possible embodiment, the general machine tool model GMTM stored in the memory or database of the backend BE can be extended by a general application model gAPPM of a general industrial application gAPP, as also Figure 2 shown. In a possible embodiment, the machine instance model MIM of the machine tool MT provided in step S2 can also be enriched by applying the rules specified in the general industrial application.

[0068] An instantiated industrial application with its configuration data extended by machine tool specific requirements mtsR can be deployed on the edge computing device ECD of the machine tool MT using the platform. Data can be obtained by the edge computing device ECD from the data sources of the machine tool MT using the machine tool specific requirements mtsR and / or general data requirements gR via the machine tool interface of the edge computing device ECD.

[0069] In a possible embodiment, a copy of the machine instance model MIM of the machine tool MT can be stored in the local memory of the edge computing device ECD of the machine tool MT. Then, the machine instance model MIM copy is used by the industrial application to convert the general data requirements gR into machine tool specific requirements mtsR. In addition, whenever the machine instance model MIM of the machine tool MT stored in the memory of the model management component MMC is changed, the machine instance model MIM copy stored in the local memory can be updated.

[0070] In a possible embodiment, data quality predicates are used to define the dynamic data requirements of an industrial application with respect to the data obtained by the edge computing device ECD from the data sources of the machine tool MT. These data quality predicates may include data acquisition frequency, probability of data correctness, data accuracy, data integrity, data currency, historical data availability range, and data security constraints.

[0071] The general data requirements gR of a general industrial application expressed according to the general machine tool model GMTM can be converted into machine tool specific requirements mtsR, including data points and / or data protocols of the machine tool MT, its controller, and its edge computing device ECD. In a possible embodiment, the instantiated industrial application is deployed on the edge computing device ECD of the machine tool MT only after it has been successfully verified that the corresponding edge computing device ECD meets the machine tool specific requirements mtsR.

[0072] The General Machine Tool Model (GMTM) stored in the database of the Model Management Component (MMC) within the Back End (BE) provides a schema for describing the machine tool components of a machine tool (MT), their configurations, and dynamic data. The General Machine Tool Model (GMTM) can include a repository of physical and abstract machine tool component classes (types), their properties, and their relationships (such as subclasses, lexical, and connectivity). An example of the General Machine Tool Model (GMTM) is also shown in Figure 4 . The General Machine Tool Model (GMTM) can include a repository of dynamic data item classes (types), both analog and discrete, their properties, and their subclass hierarchies, as well as relationships to machine tool components. The General Machine Tool Model (GMTM) can include concepts and constructs for specifying the kinematic transformations of machine tool components. The General Machine Tool Model (GMTM) can support multiple hierarchies and is also extensible. The General Machine Tool Model (GMTM) stored in the database of the Model Management Component (MMC) within the Platform Back End (BE) forms a superset of machine tool data standards, such as MTConnect and the OPC-UA Companions of the VDW. The Machine Instance Model (MIM) can define the configuration data and dynamic data of a specific machine tool (MT) by means of the schema provided by the General Machine Tool Model (GMTM). The Machine Instance Model (MIM) describes the machine tool (MT) and its machine tool components, as well as the associated data items, and can align them with the General Machine Tool Model (GMTM). The dynamic data items are in a format defined by the native controller / OEM (Original Equipment Manufacturer) or are mapped to machine-specific addresses, for example, by means of the OPC-UA address space.

[0073] Exemplary fragments of the General Machine Tool Model (GMTM) and the Machine Instance Model (MIM) are shown in Figure 4 . The General Machine Tool Model (GMTM) provides subclasses of the top-level asset (ASS) to describe the classes of the machine tool (MT) and its machine tool components. In the example shown in Figure 4 , the machine tool (MT) is a subcategory of the asset (ASS) that forms the top level. Additionally, in the example shown, the milling machine (MM) forms a subclass of the machine tool (MT). Additionally, the axis (AX) is a subclass of the asset (ASS) and is defined as part of (po) the machine tool (MT). Additionally, the linear axis 1AX is a subclass of the axis, as shown in Figure 4 . The asset (ASS) has a characteristic (hP) that includes a data item (DI). The data item (DI) can be an analog data item (aDI) or a discrete data item (dDI), as shown in Figure 4 . The actual position (aPOS) is a subclass of the position (POS), which in turn is a subclass of the analog data item (aDI), which forms a subclass of the data item (DI), as shown in Figure 4 . The asset (ASS) can be connected to (ct) another asset (ASS). The data item (DI) has an asset as its physical source (hps). The axis (AX) is connected to (ct) the motor (MOT).

[0074] InFigure 4 In the example of, a model instance model MIM fragment is shown. Figure 4 An example of a milling machine, millmach - 123, is shown in. The instance millmach - 123 has a part axis 1, which is an instance of the linear axis 1AX and has an attribute named "X". Axis 1 has the following characteristics: axisl.Enc2ActPos, which forms an instance of the actual position aPOS; and the characteristic axis1.ComTorque, which is an instance of the CommandedTorque cTOR. axis1.ComTorque has additional attributes such as unit = Nm and address = " / nck / servodata / nckServoDataCmdTorque64[1]", which is specific to SINUMERIK. The model management component MMC of the backend BE is a component suitable for managing the life cycles of both the general machine tool model GMTM and the machine instance model MIM. In addition to the CommandedTorque cTOR, the actualTorque aTOR forms a subclass of the Torque TOR. Furthermore, the commandPosition cPOS and the actualPosition aPOS form subclasses of the PositionPOS, as Figure 4 shown in.

[0075] The model management component MMC of the backend BE can provide the following function set. The model management component MMC can be used to generate and / or update the general machine tool model GMTM. In addition, it can be used to generate and / or update the mapping from the controller - specific address to the general machine tool model GMTM. In addition, the model management component MMC can be used to generate and / or update the mapping from the industrial machine tool data standard to the general machine tool model GMTM and / or the machine instance model MIM. The generation and / or update of the model is usually performed by the backend platform owner PO in the backend BE.

[0076] The model management component MMC can also be used to parse the input and convert it into a part of the machine instance model MIM and store it in the memory of the backend BE. The required input can include the machine tool data report MTDR provided by the edge computing device ECD. The machine tool data report MTDR can be provided in a controller - specific data report format (e.g., the data archive and dynamic data specification of SINUMERIK) or an industrial standard format (such as MTConnect and the OPC - UA Companions of VDW).

[0077] The required input can further include the mapping from the received report format to the general machine tool model GMTM. These can be stored in the backend or together with the MTDR.

[0078] Figure 5 An example of automatically generating a machine configuration instance model MIM from a SINUMERIK data archive is shown. The SINUMERIK data archive shown on the left can be part of the machine tool data report MTDR of a machine tool MT, which is converted into a machine instance model MIM of the machine tool MT of an instantiated general machine tool model GMTM, as Figure 5 shown. Channel1 is an instance of the channel CH, the channel CH is a subclass of the asset ASS, the linear axis 1AX and the rotary axis rAX are two subclasses of the axis AX, both forming subclasses of the asset ASS. Axis3 is an instance of the spindle SP, and the spindle SP is a subclass of the rotary axis rAX.

[0079] In a possible embodiment, Figure 5 the transformation shown can be executed by a processor of the model management component MMC.

[0080] The model management component MMC may include an interface tool that allows enriching the machine instance models MIM of different machine tools MT, or allows the OEM to enrich a series of machine instance models MIM with OEM-specific data. For example, the machine tool data stored in the PLC can be addressed in an OEM-specific manner, where only the corresponding OEM can understand the semantics of these data. For example, as Figure 6 shown, the OEM can insert a data item addressed by DB13.DBDBB148 into the machine instance model MIM. This is an instance of the workpiece state WPS and can be defined as a characteristic of the corresponding instance of the pallet changer PCH, as Figure 6 shown. If this instance does not exist in the machine instance model MIM, it also needs to be added. The model management component MMC of the backend BE may also include an interface tool for enriching the general machine tool model GMTM via the OEM if necessary. For example, if a machine component class (such as "PalletChanger") or a data item class (such as "WorkpieceState") required for instance generation does not exist in the general machine tool model GMTM, the OEM can generate its own extension of the general machine tool model GMTM. The OEM can, for example, introduce the missing classes and align them with the general machine tool model GMTM, as Figure 6 shown in the example shown.

[0081] oem1: PalletChanger (PCH) subclassOf mt: Asset ASS

[0082] oeml: PalletChanger (PCH) partOf mt: Machine Tool MT

[0083] oeml: WorkpieceState subclassOf mt: DiscreteDataitem dDI

[0084] The OEM can also decide to keep the model extension private. In this embodiment, only applications having access to the OEM-specific model can also use the OEM-specific data. Alternatively, the OEM can also send a request to the owner of the General Machine Tool Model GMTM, who can adopt the proposed class extension to the General Machine Tool Model GMTM.

[0085] Also as Figure 2 shown, the General Machine Tool Model GMTM stored in the database of the Model Management Component MMC can also be extended via the General Application Model gAPPM provided by the Application Management Component APPMC.

[0086] For example, if the application "XYZ" (Application XYZ-AxisLoadIndex) calculates the load index of all machine axes, the General Application Model gAPPM can contain the following extension of the General Machine Tool Model GMTM:

[0087] mt: Axis hasProperty application: XYZ-AxisLoadlndex

[0088] application: XYZ-AxisLoadlndex subclassOf mt: DiscreteDataitem

[0089] In this way, the application announces its model extension so that other applications can subscribe to its output in a general way. This extension can be initiated by the Application Management Component APPMC and the Application Deployment Component APPDC.

[0090] As Figure 2 shown, the Machine Instance Model MIM of the Machine Tool MT on which the Industrial Application APP is to be deployed can also be enriched by the Application Deployment Component APPDC. This can be done by applying rules specified by the application, for example, each instance a of the axis in the Machine Tool MT generates a new feature a.XYZ-AxisLoadIndex of type XYZ-AxisLoadIndex. This extension can be initiated by the Application Management (APPMC) and Deployment Component (APPDC).

[0091] In the case of triggering a lifecycle change, different functions can be applied. For example, if the Edge Computing Device ECD extracts a new Machine Tool Data Report MTDR, a new Machine Instance Model MIM can be generated. In addition, statically deployed applications can apply the enriched Machine Instance Model MIM (for example, the number of axes may change).

[0092] If a new application is released, an application general model extension for extending the General Machine Tool Model (GMTM) can be released.

[0093] In addition, if a new application is deployed on the Edge Computing Device (ECD) of the Machine Tool (MT), an Application Instance Model (MIM) can be generated.

[0094] If the General Machine Tool Model (GMTM) is insufficient to describe the new configuration, the interface tools of the Model Management Component (MMC) can be used to enrich and / or extend the Machine Instance Model (MIM).

[0095] The Application Management Component (APPMC) and the Application Development Component (APPDC) allow the development and release of a General Application (gAPP) via the Back-end System (BE). The gAPP can provide reusable functions for different Manufacturing Environments (ME). In addition to software, the application can contain the following information or data:

[0096] The application can include a general model extension. This can indicate how the released application extends the upper-layer model. For example, if Application XYZ is created, which measures the load on a specific machine axis, it can be indicated that it extends the general model by adding the parameter XYZ - AxisLoadIndex to the axis class.

[0097] The application can further include rules for extending the Machine Instance Model (MIM). This indicates how to extend an instance of the General Machine Tool Model (GMTM). In the previous example, it can indicate to which specific objects of the axis class in the model instance the parameter XYZ - AxisLoadIndex must be assigned.

[0098] The application can further contain General Requirements (gR). These include descriptions of data requirements, such as a list of data points that can be utilized by the application, which can be referenced with the help of the General Machine Tool Model (GMTM), for example, by using a query Q (such as a SPARQL query).

[0099] The application can further contain other dependencies. These other dependencies can include dependencies in aspects such as system services, other applications, resources, machine details, etc., required by the corresponding application.

[0100] For example, the application can describe data requirements in a general way without knowing exactly how many axes the Machine Tool (MT) has. In addition, the application can describe how to process dynamic data on a specific machine. For example, the application can subscribe to the torque data of all axes.

[0101] First, query AXES_CONFIG with the help of SPARQL

[0102] SELECT

[0103] Among them

[0104] 。

[0105] The application queries the static axis configuration data of all axes: their names (e.g., X, Y) and types (e.g., linear / rotary / spindle).

[0106] In addition, by querying AXES_COMMANDED_TORQUE

[0107] Select

[0108] Among them

[0109] 。

[0110] The application subscribes to axis torque data and the context of this data, i.e., which axis it is related to and in which unit it is measured.

[0111] After submission, the backend BE can manage the lifecycle of these artifacts. Both can be used during the application deployment process and the model lifecycle process.

[0112] A user, including the manufacturer, operator, or owner of the machine tool MT, can deploy an industrial application to a specific machine tool MT in a specific manufacturing environment ME via the backend system BE. When this occurs, the steps of the flowchart shown in Figure 7 are executed at the backend BE.

[0113] In step S70, application deployment is triggered. In step S71, relevance is checked, i.e., it is checked whether the application is ready to be deployed on a specific edge computing device ECD. It is checked whether all necessary requirements are met. These include physical requirements, software requirements, and / or resource requirements. Physical requirements are, for example, the model and type of the associated machine model required for the corresponding application. For example, software requirements can include the relevance to other applications, or the relevance to specific services or service versions provided by the edge computing device ECD. Resource requirements can include the available physical memory requirements or CPU requirements on the device. If some relevance is not met, the backend BE can attempt to deploy them, although not all requirements are met. Otherwise, the deployment step may fail.

[0114] As Figure 7As shown in [figure], in step S72, it is determined whether the relevance has been satisfied. If the relevance is satisfied, the parsing of the static data requirements is performed in step S74. If the relevance is not satisfied, the deployment of the relevance is triggered in step S73. The application developer D can provide a general description of the data points, for example, in the form of a query Q to the general model. For any running instance of the application, all general data requirements can be converted into data requirements specific to the machine tool MT and the edge computing device ECD for the deployed application. These specific requirements can include, for example, the physical addresses where data must be read on the specific machine of interest, and the specific protocols that must be used to communicate with the machine tool MT. This conversion step can be carried out by evaluating the query Q on the query engine QE of the backend BE, which can use the machine instance model MIM corresponding to the edge computing device ECD of the machine tool MT of interest, as also shown in [figure]. The general data requirements gR of the general application can include a list of queries Q evaluated by the query engine QE using the machine instance model MIM of the machine tool MT to generate a query result QR that forms the machine tool specific requirements mtsR, as schematically shown in [figure]. The general industrial application can be instantiated by extending its configuration data with the machine tool specific requirements mtsR to provide the instantiated industrial application for the corresponding machine tool MT. Figure 8 As shown in [figure], Figure 8 As schematically shown in [figure].

[0115] During deployment, it can be decided which general data requirements gR in the form of queries Q are to be parsed into specific requirements, and which general data requirements gR are not parsed before deployment and are left for the edge computing device ECD to parse during the runtime of the application. Parsing the general data requirements gR into specific data requirements during deployment is more efficient because it does not consume the resources of the edge device. However, the solution during deployment is less flexible because it does not allow the running application to react to changes in the model. In a possible embodiment, the backend BE can select which general data requirements gR to statically parse (e.g., before deployment) based on the knowledge of the likelihood of the requested parameters changing during the runtime of the application.

[0116] The machine tool specific requirements mtsR (possibly together with the general data requirements gR to be parsed at runtime) can be deployed with the application in Figure 7 step S75, and can be used by the edge computing device ECD to obtain specific data values from the machine tool MT and provide the obtained data to the application via the machine interface component on the edge computing device ECD, which uses a specific protocol to implement communication with the actual machine tool MT.

[0117] In some cases, the query Q is not statically resolved at deployment but rather at runtime when the application APP is executed on the processor of the edge computing device ECD. In a possible embodiment, the edge computing device ECD may include a local copy of the machine instance model MIM and a local query engine QE, as also shown in Figure 9 as shown. The edge computing device ECD may include a local memory that contains a local copy of the machine model instance cMIM. Additionally, the edge computing device ECD may include a local query engine QE as shown in Figure 9 as shown. The application APP may use the local query engine QE to resolve the general query Q into data points for a specific machine tool MT and / or resolve it from other applications using an appropriate addressing mechanism into the general query Q. These can then be forwarded to the machine tool interface component MT-INT of the edge computing device ECD.

[0118] In a possible embodiment, data quality predicates can be used to define dynamic data requirements. If the application APP provides such predicates at runtime, the query engine QE will not only be used to resolve the data points required by the application but also to match the required data quality with the available data sources such as machine sensors, cameras, audio input devices, etc.

[0119] The attributes of data quality can be defined by the application as predicates for time series data sets over time, such as the acquisition frequency of data from machine sensors or provided by other applications. Additionally, the predicates can include the probability of uncertainty regarding data correctness. The predicates can also be related to data accuracy or information about missing data points in time series data. Furthermore, the predicates can refer to the currency of the received data and / or the amount or range of historical data availability. The predicates can also contain data security constraints that can be defined in a static manner (e.g., role-based access control of the application to the data) or in a dynamic manner (e.g., based on the need-to-know context of the application that only allows access to data in a predefined context, which is continuously evaluated during the data transmission of the application).

[0120] To coordinate the entire data point resolution process, the semantics of data quality attributes and predicates can be added to the general model and managed in the instance model. This allows the merging of the data quality requirements of the application when resolving data points from them based on the local model instance. For example, when requesting the query engine QE to resolve data points based on the model, each application can provide one or more attributes as optional predicates for resolving the required data points based on their principle availability and the provided quality attributes. By applying the predicates to the model instance that manages the data quality attributes of the incoming data over time, the query engine QE can match the defined quality requirements.

[0121] This means that the running application may potentially depend on the local machine instance model MIM stored in the local memory of the edge computing device ECD, as Figure 9 shown. The local machine instance model MIM can be managed at the backend BE and can be changed by multiple parties. Therefore, whenever the corresponding machine instance model MIM on the backend BE is updated, a mechanism is implemented to update the local machine instance model MIM on the edge computing device ECD. The backend BE can execute this process.

[0122] When changing the machine instance model MIM, the backend BE can (automatically or manually by the machine operator or machine owner) deploy the updated machine model instance on the edge computing device ECD. Subsequently, the backend BE checks for each individual application currently deployed on the edge computing device ECD whether it must be redeployed, including whether the relevance or data requirements are still met. In case of a change in the statically resolved data requirements, application redeployment can be performed.

[0123] As Figure 10 shown, if the machine instance model MIM is updated in step S10-0, the updated machine instance model MIM can be deployed in step S10-1, where for each installed application, steps similar to the process Figure 7 shown are performed. In step S10-2, the relevance is checked, similar to step S71 Figure 7 shown. In step S10-3, it is determined whether the check has been successfully executed. If the relevance is met, the static requirements are parsed in step S10-5. Otherwise, the deployment of the task is triggered in step S10-4. In step S10-6, it is determined whether a change has been made. If a change has occurred, then in step S10-7, the application binary file and configuration data are deployed on the edge computing device ECD. The process ends in step S10-8.

Claims

1. A method for deploying industrial applications on an edge computing device ECD connected to a controller of a machine tool MT, The method includes the following steps: (a) Provide (S1) a general machine tool model GMTM of a machine tool component, wherein the general machine tool model GMTM provides a scheme suitable for describing the configuration, kinematics, and associated data items of the machine tool component; (b) Instantiate (S2) the general machine tool model GMTM of the machine tool component by the model management component MMC of the deployment platform based on the machine tool data report MTDR extracted by the edge computing device ECD from the controller of the machine tool MT, to provide a machine instance model MIM of the machine tool MT, wherein the machine instance model MIM defines the configuration data and dynamic data of the machine tool by means of the scheme provided by the general machine tool model GMTM; (c) Use the machine instance model MIM of the machine tool MT by the model management component MMC to convert (S3) the general data requirements gR of a general industrial application into requirements specific to the machine tool mtsR; (d) Instantiate the general industrial application by extending its configuration data with the requirements specific to the machine tool mtsR, to provide (S4) an instantiated industrial application for the machine tool MT, and (e) Deploy (S5) the instantiated industrial application on the edge computing device ECD of the machine tool MT by the model management component MMC.

2. The method according to claim 1, wherein The machine instance model MIM of the corresponding machine tool MT is enriched with OEM-specific data of the original equipment manufacturer OEM of the machine tool MT.

3. The method according to claim 1 or 2, wherein The general machine tool model GMTM and the machine instance model MIM of the machine tool MT are stored in the memory of the model management component MMC in the backend connected to the edge computing device ECD.

4. The method according to claim 1 or 2 above, wherein, The general data requirements gR of the general industrial application include queries Q evaluated by a query engine QE using the enriched machine instance model MIM of the machine tool MT to generate query results QR, forming requirements specific to the machine tool mtsR.

5. The method according to claim 1 or 2 above, wherein, The general data requirements gR of the general industrial application that are converted into requirements specific to the machine tool mtsR using the enriched machine instance model MIM of the machine tool MT include static data requirements converted at deployment and dynamic data requirements converted during application runtime when the industrial application is executed on the processor of the edge computing device ECD.

6. The method according to claim 1 or 2 above, wherein The machine tool data report MTDR extracted by the edge computing device ECD from the controller of the machine tool MT includes a predefined format containing a controller-specific data report format or an industrial standard format, and includes a mapping from the predefined format to the general machine tool model GMTM stored in the database of the model management component MMC in the backend.

7. The method according to claim 1 or 2 above, wherein, The general machine tool model GMTM stored in the memory of the backend is extended by the general application model of the general industrial application.

8. The method according to claim 1 or 2 above, wherein, The machine instance model MIM of the machine tool MT is enriched by applying rules specified in the general industrial application.

9. The method according to claim 1 or 2 above, wherein, The instantiated industrial application having its configuration data extended with the machine tool-specific requirements mtsR is deployed on the edge computing device ECD of the machine tool MT, where the edge computing device ECD uses the machine tool-specific requirements mtsR and / or the general data requirements gR to obtain data from the data sources of the machine tool MT via the machine tool interface of the edge computing device ECD.

10. The method according to claim 1 or 2 above, wherein, A copy of the machine instance model MIM of the machine tool MT is stored in the local memory of the corresponding edge computing device ECD of the machine tool MT, where the copy of the machine instance model MIM is used by the industrial application to convert the general data requirements into machine tool-specific requirements mtsR. Wherein, whenever the machine instance model MIM of the machine tool MT stored in the memory of the model management component MMC is changed, the copy of the machine instance model MIM is updated.

11. The method according to claim 1 or 2 above, wherein, Data quality predicates are used to define the dynamic data requirements of the industrial application with respect to the data obtained by the edge computing device ECD from the data sources of the machine tool MT, where the data quality predicates include Data acquisition frequency, Probability of data correctness, Data accuracy, Data integrity, Data currency, Historical data availability range, and Data security constraints.

12. The method according to claim 1 or 2 above, wherein, Converting the general data requirements gR of a general industrial application expressed according to the general machine tool model GMTM into machine tool-specific requirements mtsR, including data points and / or data protocols of the machine tool MT, its controller, and its edge computing device ECD.

13. The method according to claim 1 or 2 above, wherein After successfully checking that the corresponding edge computing device ECD meets the machine tool-specific requirements mtsR, the instantiated industrial application is deployed on the edge computing device ECD of the machine tool MT.

14. A deployment platform for deploying an industrial application on an edge computing device ECD connected to the controller of a machine tool MT. The deployment platform executes the method according to any one of claims 1-13.

Images