Digital twin system for manufacturing processes includign equipment health simulation model

Abstract

A digital twin system for a manufacturing process including an equipment integrity simulation model according to one aspect of the present invention comprises: a simulation data pipeline that receives field torque data for equipment to be simulated and virtual torque data predicted by an AI model; an equipment integrity simulation model that determines the integrity of the equipment by predicting stress values ​​of each component included in the equipment based on simulation torque data including field torque data and virtual torque data; and a simulator that performs a simulation of the equipment based on the equipment integrity simulation model to generate simulation result data.

Description

Technology Field

[0001] The present invention relates to a digital twin system, and more specifically, to a digital twin system for a manufacturing process including an equipment health simulation model. Background Technology

[0002] Factory management systems are being utilized to improve production efficiency by automatically managing various events occurring in manufacturing plants. In particular, thanks to recent advancements in sensor technology, technologies are being developed to attach sensors to various pieces of equipment and analyze the status of facilities or the factory in real time using the sensing data collected by these sensors. In the case of such technologies, not only is a large number of sensors installed, but the sensing cycles of each sensor are also short, which can generate a massive amount of collected data.

[0003] As described in the 'Digital Twin-based Smart Factory Operation System and Method (Korean Registered Patent 10-2523738)', existing factory management systems collect and process large amounts of data in real time and store them in a big data repository. By analyzing the large amounts of data stored in the big data repository, they can determine whether there are abnormalities in intermediate goods and products or monitor equipment failures in real time. However, while existing factory management systems can easily monitor and analyze the operational status of a real manufacturing plant, they cannot easily create a virtual factory identical to the real manufacturing plant to support operations remotely or simulate various operational situations and strategies based on the virtual factory. The problem to be solved

[0004] The technical objective of the present invention is to provide a digital twin system for a manufacturing process that includes an equipment health simulation model capable of accurately predicting the remaining lifespan and health of the equipment. means of solving the problem

[0005] A digital twin system for a manufacturing process including an equipment integrity simulation model according to one aspect of the present invention comprises: a simulation data pipeline that receives field torque data for equipment to be simulated and virtual torque data predicted by an AI model; an equipment integrity simulation model that determines the integrity of the equipment by predicting stress values ​​of each component included in the equipment based on simulation torque data including field torque data and virtual torque data; and a simulator that performs a simulation of the equipment based on the equipment integrity simulation model to generate simulation result data. Effects of the invention

[0006] According to the present invention, by performing a simulation using not only torque data generated in actual equipment but also torque data generated virtually, and by selecting torque data that affects the remaining lifespan to accurately calculate stress values, the accuracy of the remaining lifespan can be improved.

[0007] Furthermore, the present invention can continuously analyze equipment health and predict and analyze long-term equipment health to estimate the repair and replacement cycles of parts. The present invention can extend the service life of equipment or parts up to a maximum value considering a safety factor from the remaining lifespan.

[0008] In addition, since the health of the equipment is predicted through an equipment health simulation model rather than manual work, the time required for equipment health analysis can be reduced. Brief explanation of the drawing

[0009] FIG. 1 is a schematic diagram showing a digital twin system for a manufacturing process according to one embodiment of the present invention. Figure 2 is a diagram showing the configuration of the real-time connection part of the digital twin platform. Figure 3 is a diagram showing the configuration of the simulation section of the digital twin platform. Figure 4 is a diagram showing the configuration of the virtualization section of the digital twin platform. Figure 5 is a diagram illustrating the simulation of the health of equipment in a digital twin system. Figure 6 is a diagram showing the configuration of the equipment health simulation model of Figure 5. Figure 7a is a graph showing an example of simulated torque data over time. Figure 7b is a table showing the torque change and number of cycles for the simulation torque data shown in Figure 7a. Figure 8a is a table showing an example of the minimum and maximum number of iterations for each part. Figure 8b is a graph showing an example of an SN diagram for part 1 of Figure 8a. Figure 9 is a diagram illustrating the processing of simulation result data in a digital twin system. Specific details for implementing the invention

[0010] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.

[0011] The meaning of the terms described in this specification should be understood as follows.

[0012] A singular expression should be understood to include a plural expression unless the context clearly defines otherwise, and terms such as "first," "second," etc., are intended to distinguish one component from another, and the scope of rights shall not be limited by these terms.

[0013] Terms such as "include" or "have" should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0014] The term “at least one” should be understood to include all combinations that can be presented from one or more related items. For example, the meaning of “at least one of the first item, the second item, and the third item” is not only the first item, the second item, or the third item individually, but also all combinations of items that can be presented from two or more of the first item, the second item, and the third item.

[0015] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.

[0016] Hereinafter, embodiments according to the present invention will be described in detail with reference to the attached drawings.

[0017] FIG. 1 is a schematic diagram showing a digital twin system for a manufacturing process according to one embodiment of the present invention.

[0018] Referring to FIG. 1, a digital twin system (1) according to one embodiment of the present invention is composed of layers such as a device (10), a network (20), a digital twin platform (30), and an application (40).

[0019] The device (10) includes a plurality of devices placed in a manufacturing plant. A manufacturing process for producing a predetermined product may be carried out in the manufacturing plant. The manufacturing process includes a plurality of processes, and at least one facility and a data collection device may be placed in each of the plurality of processes.

[0020] A data collection device collects data generated during the process. In one embodiment, the data collection device can collect microdata generated during the process. Here, microdata refers to raw data, which is the data itself collected through various sensors at the manufacturing plant site.

[0021] The data acquisition device includes various measuring instruments, sensors, actuators, etc. for collecting microdata. The data acquisition device may further include a P / C, PLC (Programmable Logic Controller), DCS (Distributed Control System), etc., for integrating or controlling data collected by the measuring instruments, sensors, actuators, etc.

[0022] For example, a data collection device can collect data generated in a continuous process. A continuous process refers to a method in which multiple processes are performed consecutively to produce a finished product using raw materials, and the outputs of each process are mixed together or the state of the output of a specific process changes and is supplied to a subsequent process. The steelmaking process is a representative example of such a continuous process. Since the steelmaking process consists of various processes such as ironmaking, steelmaking, continuous casting, and rolling, when a data collection device is applied to the steelmaking process, it can collect microdata generated during the progress of these various processes, such as ironmaking, steelmaking, continuous casting, and rolling.

[0023] Such microdata is operational data collected at the manufacturing plant site and may include equipment status information, operation events, material information such as temperature, thickness, and humidity, and input material and energy data related to product production. Such microdata may include real-time data in the form of time series generated by sensors and event data from equipment or control devices.

[0024] The network (20) transmits large volumes of microdata collected from the device (10) to the digital twin platform (30). The network (20) may include a network cable, gateway, router, wireless AP, etc., for transmitting microdata generated from the device (10) to the digital twin platform (30).

[0025] In one embodiment, the network (20) according to the present invention may support various communication methods depending on the type of sensor, etc. included in the device (10). For example, the network (2) may support various communication methods such as iBA, OPC (OLE for Process control), TCP / IP, etc. depending on the type of sensor, etc.

[0026] The digital twin platform (30) receives large volume microdata collected from the device (10) through the network (20). The digital twin platform (30) processes the received large volume microdata in real time to generate field data. Additionally, the digital twin platform (30) can generate simulation result data by performing simulations according to various simulation scenarios based on the generated field data. Furthermore, the digital twin platform (30) can monitor the operation and equipment status of the manufacturing plant in a virtual space based on the field data, or verify the simulation results in a virtual space based on the simulation result data. Such a digital twin platform (30) may be an IT platform that processes and stores a series of data and executes simulations and virtualization.

[0027] The digital twin platform (30) may include a real-time connection unit (100), a simulation unit (200), a virtualization unit (300), a management unit (400), and a standardization unit (500) as illustrated in FIG. 1.

[0028] The real-time connection unit (100) receives large volume microdata collected from the device (10) through the network (20). In addition, the real-time connection unit (100) preprocesses the microdata collected from the device (10) to generate field data. A detailed description of the real-time connection unit (100) will be provided later with reference to FIG. 2.

[0029] The simulation unit (200) performs a simulation according to a simulation scenario based on data collected at the manufacturing plant site. Specifically, the simulation unit (200) receives field data from the real-time connection unit (100). The simulation unit (200) performs a simulation according to a simulation scenario using the received field data and generates simulation result data.

[0030] In addition, the simulation unit (200) can provide the user with the simulation progress and simulation results. The user can make various decisions at the operation site by considering the simulation results provided by the simulation unit (200). A detailed description of the simulation unit (200) will be provided later with reference to FIG. 3.

[0031] The virtualization unit (300) virtualizes the manufacturing plant to create a virtual plant and makes the virtual plant identical to the manufacturing plant site or the running simulation. Specifically, the virtualization unit (300) receives site data from the real-time connection unit (100) or receives simulation result data from the simulation unit (200). In order to implement the state represented by the site data or simulation result data in the virtual plant, the virtualization unit (300) may change at least one of the shape, operation, and state of at least one of the virtual objects included in the virtual plant.

[0032] In addition, the virtualization unit (300) can provide a virtual factory to the user. The user can monitor the manufacturing plant through the virtual factory and visually check the simulation results through the virtual space. A detailed description of the virtualization unit (300) will be provided later with reference to FIG. 4.

[0033] The management unit (400) manages individual configurations belonging to the digital twin platform (30). The management unit (400) manages UI / UX, user rights, simulation models and virtualization models, configuration files for configurations for data collection, individual monitoring of each configuration, management of linkage information between preset configuration values, and processing performance and integrated monitoring information of the entire system.

[0034] The standardization department (500) defines standardized data specifications for transmitting data of virtual objects in virtual space.

[0035] The application (40) processes and provides screens and data necessary for business based on the digital twin platform (30).

[0036] Figure 2 is a diagram showing the configuration of the real-time connection part of the digital twin platform.

[0037] Referring to FIG. 2, a real-time connection unit (100) included in a digital twin platform (30) according to one embodiment of the present invention may include an interface middleware (110), a real-time processing unit (120), a big data storage (130), and a service unit (140).

[0038] The interface middleware (110) receives large volumes of microdata collected from the device (10) through the network (20). The device (10) may include a plurality of sensors (11), a plurality of actuators (12), and a control device (13). The control device (13) may integrate the data collected by the plurality of sensors (11) and the plurality of actuators (12) and transmit it to the interface middleware (110) through the network (20).

[0039] In one embodiment, the interface middleware (110) may support various communication protocols. For example, the interface middleware (110) may support iBA, OPC (OLE for Process Control), TCP / IP, FTP (File Transfer Protocol), etc. In this way, the interface middleware (110) may support all communication methods capable of communicating with various networks (20).

[0040] Interface middleware (110) parses microdata into meaningful units based on a preset layout and standardizes the parsed data. In one embodiment, for microdata consisting of one item ID, a collection time, and one measurement value, the interface middleware (110) can standardize the parsed data by converting the item ID into a standard item ID according to a preset standard conversion criterion and unifying the unit and number of digits of the measurement value.

[0041] In this way, the interface middleware (110) can preprocess data so that the same standard item ID is used for measurement values ​​that have measured the same attributes, thereby enabling the collected data from the continuous process to be processed in conjunction based on each standard item ID.

[0042] In addition, the format of the measurement values ​​included in the collected data may vary depending on the type of device (10), such as a sensor or actuator. As a result, in order to process the collected data of different formats collected from the continuous process, it is necessary to additionally perform a process of converting the unit and length of the data according to the type of device (10), making it impossible to process the large volume of data collected from the continuous process in real time. Therefore, the interface middleware (110) according to the embodiment of the present invention can standardize microdata so that the real-time processing unit (120) can process the large volume of data in real time. For convenience of explanation, the standardized microdata will be referred to as field data below.

[0043] The real-time processing unit (120) maps process identifiers to standardized field data through the interface middleware (110) and sorts the field data so that data between areas such as operation, equipment, and quality can be analyzed in conjunction.

[0044] The process identifier may include at least one of the equipment identifier of the equipment performing each process or the material identifier of the material processed by the equipment. At this time, the real-time processing unit (120) can extract the equipment identifier of the equipment where the microdata was generated based on the microdata collection time and the attribute information of the sensor that collected the microdata, and can extract the material identifier of the material processed by the equipment corresponding to the equipment identifier mapped to the mapping data based on the work instruction information performed in each process.

[0045] In one embodiment, the equipment identifier may be an equipment number assigned to each piece of equipment, and the material identifier may be a material number assigned to each material.

[0046] Meanwhile, the real-time processing unit (120) may sequentially sort field data mapped with the same material identifier according to the collection time, and sort the field data sorted according to the time order based on the location where the data was collected on the material corresponding to the same material identifier.

[0047] The real-time processing unit (120) can generate field data for missing values ​​at points where no sensor is present and micro data is not generated, or in the middle of the collection cycle, based on the aligned field data.

[0048] The big data storage unit (130) stores field data sorted by the real-time processing unit (120) in a big data storage space. In addition, the big data storage unit (130) manages data to prevent data loss and provides functions for querying and analyzing historical data.

[0049] The service unit (140) is structured to reuse standardized processing processes and business standards as services, and facilitates the linkage between planning, execution, and control through connections between services defined by functional units by repositories of business know-how, and provides analysis results by calling and executing an analysis model that includes a quality judgment model or an anomaly prediction model for materials or products.

[0050] Since the analysis model is stored in advance in the model storage unit (not shown), when an execution call event for the analysis model occurs, the service unit (140) can extract data required for the analysis model from the real-time processing unit (120) or the big data storage unit (130) and execute the analysis model based on the extracted data.

[0051] The service unit (140) can directly receive and analyze data processed by the real-time processing unit (120), or analyze the data by referring to it when the data is stored in the big data storage unit (130).

[0052] Figure 3 is a diagram showing the configuration of the simulation section of the digital twin platform.

[0053] Referring to FIG. 3, a simulation unit (200) included in a digital twin platform (30) according to one embodiment of the present invention includes a simulation data pipeline (210), a simulation database (215), a simulation modeling unit (220), a simulation model database (225), a simulator (230), a simulation result database (255), and a data interface unit (280). In one embodiment, the simulation unit (200) may further include at least one of an optimization unit (260), a simulation monitoring unit (270), and an automatic control unit (275).

[0054] The simulation data pipeline (210) receives field data in real time from the real-time connection unit (100). The simulation data pipeline (210) extracts data required as initial values ​​in the simulation from the received field data and processes the extracted data into simulation field data for performing the simulation. The simulation data pipeline (210) maps the processed simulation field data to the simulation so that it can be used when performing the simulation and stores it in the simulation database (215). For example, the simulation data pipeline (210) can extract operational information of the equipment from the field data and process it into simulation field data in order to perform an equipment aging prediction model. The simulation data pipeline (210) can map the processed simulation field data to the aging prediction model and store it in the simulation database (215).

[0055] The simulation database (215) stores simulation field data processed in the simulation data pipeline (210). The simulation field data is operational data collected at the manufacturing plant site and may include equipment status information, operation events, material status information such as temperature, thickness, and humidity of the material, and input material and energy data according to product production. The simulation field data may have a time-series form by sensors or an event form of equipment or control device (13).

[0056] The simulation modeling unit (220) models the objects to be simulated among the objects included in the manufacturing plant. Here, the objects included in the manufacturing plant include not only tangible objects but also intangible objects. Tangible objects may include equipment placed at the manufacturing plant site, parts of the equipment, materials, products, workers, etc. Intangible objects may include work processes, work standards, AI models, physical analysis models, control algorithms, anomaly detection models, torque models, stress models, formula models, data models, etc.

[0057] The simulation modeling unit (220) creates unit models by modeling objects to be simulated and creates a simulation model by connecting the unit models according to a simulation scenario. In this case, if the object is an intangible object stored externally in the form of a program, the unit model may be implemented as connection information that can access the object. The simulation model may include input parameters, output parameters, operations or processes, and connection relationships with other unit models for each of the unit models. The simulation modeling unit (220) stores the created simulation model in the simulation model database (225).

[0058] The simulation model database (225) stores simulation models that have been completed by the simulation modeling unit (220). The simulation model database (225) can store simulation models mapped for each of a plurality of simulation scenarios.

[0059] The simulator (230) reads a simulation model stored in the simulation model database (225) and performs a simulation. The simulator (230) includes a simulation management unit (240), a simulation execution unit (245), and a simulation parallel processing unit (250).

[0060] The simulation management unit (240) manages the simulation execution unit (245) and the simulation parallel processing unit (250). The simulation management unit (240) can also manage the optimization unit (260). The simulation management unit (240) can manage the process and time of the simulation.

[0061] Additionally, the simulation management unit (240) performs a simulation by calling at least one of the simulation execution unit (245) and the simulation parallel processing unit (250). The simulation management unit (240) stores the simulation execution results in the simulation result database (255).

[0062] Specifically, the simulation management unit (240) can control the execution of a simulation using a simulation model and simulation site data. The simulation management unit (240) reads a simulation model corresponding to the simulation scenario to be executed from the simulation model database (225). That is, the simulation management unit (240) can read all unit models required to perform the simulation, the input and output parameters of each of all unit models, and the connection relationships between unit models from the simulation model database (225).

[0063] In addition, the simulation management unit (240) reads simulation field data required for the simulation execution process from the simulation database (215).

[0064] The simulation management unit (240) can call the simulation execution unit (245) and transmit the simulation model and simulation site data read into the called simulation execution unit (245).

[0065] The simulation management unit (240) may call the simulation execution unit (245) when performing a simulation of discrete events and continuous events, either individually or in combination. A discrete event may be an event consisting of a single event. A continuous event is an event consisting of events that occur continuously over time, for example, an event in which the speed of a specific facility decelerates or accelerates. The simulation management unit (240) may transmit a simulation model and simulation site data to the simulation execution unit (245). The simulation management unit (240) may transmit at least one of the unit models included in the simulation model to the simulation execution unit (245) along with the simulation site data. The simulation management unit (240) may transmit unit models to the simulation execution unit (245) sequentially according to the connection relationships of the unit models included in the simulation model, thereby performing a simulation for the unit models.

[0066] Meanwhile, if the simulation management unit (240) wishes to process the simulation in parallel, it may call the simulation parallel processing unit (250). The simulation management unit (240) can execute multiple simulation scenarios in parallel by passing node identification information for identifying each of the multiple simulation scenarios to the simulation parallel processing unit (250).

[0067] The simulation management unit (240) stores simulation result data, which is executed by the simulation execution unit (245) or the simulation parallel processing unit (250), in the simulation result database (255).

[0068] The simulation execution unit (245) is a simulation engine device that executes a simulation model in response to a call from the simulation management unit (240). Specifically, the simulation execution unit (245) creates simulation objects corresponding to each of the unit models included in the simulation model and creates a simulation space by connecting the simulation objects. For example, if there are 7 identical part FMs in a specific facility, the simulation management unit (240) can read model data for the part FMs from the simulation model database (225) and transmit it to the simulation execution unit (245). The simulation execution unit (240) can create 7 simulation part FMs using the model data for the part FMs. The simulation execution unit (245) can create a simulation space by connecting the 7 simulation part FMs according to their connection relationships.

[0069] The simulation execution unit (245) executes a simulation in a simulation space created based on simulation field data. The simulation execution unit (245) can perform simulations by linking various heterogeneous models, such as discrete event models or continuous time models.

[0070] The simulation parallel processing unit (250) is a simulation engine device and processes multiple simulation scenarios in parallel upon a call from the simulation management unit (240). Additionally, the simulation parallel processing unit (250) divides one of the simulation data, simulation logic, and simulation range to process the simulation in parallel.

[0071] For example, the simulation parallel processing unit (250) can divide the simulation data and process the simulation in parallel. For example, when simulating the operation of a specific facility for 10 years, the simulation parallel processing unit (250) can divide the data into 1-year units and process the operation simulation for each of the 10 years in parallel.

[0072] As another example, the simulation parallel processing unit (250) can process the simulation in parallel by dividing the simulation logic. For example, when calculating the stress of a specific facility including 10 parts, the simulation parallel processing unit (250) can divide it into units of one part and process the stress calculation simulation for each of the 10 parts in parallel. In addition, the simulation parallel processing unit (250) can process the simulation in parallel by dividing the simulation range. A detailed description of the configuration of the simulation parallel processing unit (250) will be provided later with reference to FIGS. 5 to 9.

[0073] The simulator (230) can perform simulations regardless of the type of model through the simulation execution unit (245), and can perform simulations without a decrease in execution speed even when linking models using heterogeneous languages. In addition, the simulator (230) can improve the simulation execution speed by processing the simulation in parallel through the simulation parallel processing unit (250).

[0074] The optimization unit (270) finds the optimal simulation parameter in which the objective function has a maximum or minimum value. Here, the objective function may refer to a function used for a specific purpose. For example, the optimization unit (270) may include an objective function representing work time for the purpose of completing a task in the minimum time. The optimization unit (270) can find the value of the optimal simulation parameter in which the objective function representing work time has a minimum value. If there are multiple simulation parameters, the optimization unit (270) can find the optimal value for each of the multiple simulation parameters to have the objective function have a minimum value.

[0075] The optimization unit (270) can adjust the simulation parameters and repeat the simulation until the objective function has a maximum or minimum value. Here, adjusting the simulation parameters may include adjusting the value of the simulation parameters, adjusting the type of the simulation parameters, or adjusting the input order of the simulation parameters.

[0076] Specifically, the optimization unit (270) may adjust the simulation parameters and request the simulation management unit (240) to perform a simulation based on the adjusted simulation parameters. The optimization unit (270) receives simulation result data from the simulation management unit (240) and can calculate an objective function based on the simulation result data. If the objective function is not a maximum or minimum value, the optimization unit (270) may adjust the simulation parameters again and request the simulation management unit (240) to perform a simulation based on the adjusted simulation parameters.

[0077] In one embodiment, the optimization unit (270) can find the optimal simulation parameter in which the objective function has a maximum or minimum value through an optimization algorithm.

[0078] The optimization unit (270) stores the optimized simulation result data in the simulation result database (255) through the simulation management unit (240).

[0079] The simulation result database (255) stores simulation result data that is performed and produced by the simulator (230). The simulation result data may include changed result values ​​for simulation objects and parameters.

[0080] The simulation monitoring unit (270) provides the simulation progress and simulation results to the user through a user screen. The simulation monitoring unit (270) can provide the user with not only data that can be expressed in a virtual space through the virtualization unit (300) but also data that cannot be expressed in a virtual space.

[0081] The automatic control unit (275) controls the manufacturing plant based on simulation result data. The automatic control unit (275) can generate control values ​​for the manufacturing plant based on simulation result data and transmit the generated control values ​​to the device (10). The device (100) can control the equipment placed in the manufacturing plant according to the control values ​​received from the automatic control unit (275).

[0082] The automatic control unit (275) can control the manufacturing plant based on optimal simulation result data to achieve a specific purpose. The automatic control unit (275) can generate an optimal control value for the manufacturing plant based on the optimal simulation result data and transmit the generated optimal control value to the device (10).

[0083] The data interface unit (280) transmits simulation result data to the virtualization unit (300). The simulation result data may be data stored in the simulation result database (255) or data directly transmitted from a simulation model or unit model. The data interface unit (280) selects simulation result data in which a data change has occurred among the simulation result data, and transmits only the selected simulation result data to the virtualization unit (300).

[0084] A digital twin system (1) according to one embodiment of the present invention can reduce the number of commissioning tests or experiments performed directly at a manufacturing plant by performing a simulation through a simulation unit (200), thereby reducing time and costs and eliminating safety risks.

[0085] In addition, the digital twin system (1) according to one embodiment of the present invention is equipped with a simulation data pipeline (210) that connects data between the manufacturing plant site and the simulation space in real time, thereby making it possible to immediately simulate the current state of the manufacturing plant site.

[0086] In addition, the digital twin system (1) according to one embodiment of the present invention can store the simulation results in a separate simulation result database (255) so that the simulation results can be monitored in real time or utilized for alternative analysis.

[0087] In addition, the digital twin system (1) according to one embodiment of the present invention can provide optimal values ​​for various scenarios and parameters with different operating conditions by linking with an optimization function for simulation. The digital twin system (1) according to one embodiment of the present invention can control the manufacturing plant site based on the optimal values. By applying the optimal operating method to the site, the digital twin system (1) according to one embodiment of the present invention can realize an automated and unmanned manufacturing site.

[0088] In addition, the digital twin system (1) according to one embodiment of the present invention can link various models and simulate them in a single simulation space. Through this, the digital twin system (1) according to one embodiment of the present invention does not require a development step for porting a separate model, and can simulate existing models or systems currently in use in the field by connecting them directly to the simulation space.

[0089] Figure 4 is a diagram showing the configuration of the virtualization section of the digital twin platform.

[0090] Referring to FIG. 4, a virtualization unit (300) included in a digital twin platform (30) according to one embodiment of the present invention may include a virtual data pipeline (310), a virtual database (315), a shape modeling unit (320), a motion modeling unit (330), a state modeling unit (340), a virtualization model database (345), a virtual factory creation unit (350), a first virtual service unit (360), and a second virtual service unit (365).

[0091] The virtual data pipeline (310) receives field data in real time from the real-time connection unit (100). The virtual data pipeline (310) processes the received field data into virtual field data to be displayed in the virtual factory.

[0092] Additionally, the virtual data pipeline (310) receives simulation result data in real time from the simulation unit (200). The virtual data pipeline (310) processes the received simulation result data into virtual simulation data to be displayed in the virtual factory.

[0093] The virtual data pipeline (310) can process field data and simulation result data into data to be represented in a three-dimensional virtual factory. The field data and simulation result data may not contain three-dimensional information. Accordingly, the virtual data pipeline (310) can process the data so that the state or location represented by the field data or simulation result data can be represented in the virtual factory.

[0094] The virtual data pipeline (310) can process field data or simulation result data into virtual data for changing at least one of the shape, operation, and state of at least one virtual object in a virtual factory. The virtual data may include virtual field data processed based on field data and virtual simulation data processed based on simulation result data.

[0095] The virtual data pipeline (310) maps field data or simulation result data to a 3D virtual object stored in a virtualization model database (345) and can generate information related to each other in a shape model, motion model, and state model for the 3D virtual object. The virtual data pipeline (310) can generate virtual field data or virtual simulation data by additionally generating shape information of the 3D virtual object, location information of the 3D virtual object, and motion information of the 3D virtual object that can be processed in a 3D virtual factory based on the field data or simulation result data.

[0096] As described above, the virtual data pipeline (310) can process not only data from the manufacturing plant site but also additional data that is not from the manufacturing plant site into virtual site data. The virtual data pipeline (310) can additionally generate additional information required in the 3D virtual factory, such as shape information of the 3D virtual object, location information of the 3D virtual object, operation information of the 3D virtual object, and state information of the 3D virtual object.

[0097] For example, field data may only include temperature information of a first point and a second point of the material. In this case, since the temperature between the first point and the second point of the material must be represented in the virtual factory, the virtual data pipeline (310) may additionally generate temperature information between the first point and the second point, and furthermore, generate information for representing the material in three dimensions. The virtual data pipeline (310) may additionally generate shape information of a virtual object corresponding to the material and changes according to the state of the virtual object for representing the temperature.

[0098] As another example, field data may include roll speed information for equipment. In this case, since the virtual factory must represent a roll having a three-dimensional shape rotating at a speed corresponding to the roll speed information, the virtual data pipeline (310) may additionally generate shape information of a virtual object corresponding to the roll and motion information of a virtual object.

[0099] The virtual data pipeline (310) transmits virtual field data or virtual simulation data to the virtual factory created by the virtual factory creation unit (350) through the virtual database (315).

[0100] The virtual database (315) stores virtual field data and virtual simulation data processed in the virtual data pipeline (310). The virtual field data may be operational data collected at the manufacturing plant site. The virtual field data may be state data such as equipment state information, operation events, material state information such as temperature, thickness, and humidity of the material, and input materials and energy data according to product production, and may be data mainly generated by the control values ​​of the sensor (11) or the equipment. The virtual field data may have a time-series form by the sensor or an event form of the equipment or control device (13).

[0101] In order to convert a real-world manufacturing plant into a virtual plant, it is necessary to model the shape, operation, and state of each object included in the real-world manufacturing plant. At this time, the objects to be modeled may include equipment, parts, materials, products, workers, and the environment included in the real-world manufacturing plant. To model the virtual plant, the virtualization unit (300) includes a shape modeling unit (320), an operation modeling unit (330), and a state modeling unit (340).

[0102] The shape modeling unit (320) creates a shape model by modeling the three-dimensional shape of each object included in the manufacturing plant. The shape modeling unit (320) models the shape consisting of coordinates, lines, and surfaces for each object included in the manufacturing plant. If a three-dimensional drawing is available, the shape modeling unit (320) can model the shape of the object using the three-dimensional drawing. If a three-dimensional drawing is not available, the shape modeling unit (320) can model the shape of the object using scanning information, visual estimation information, or a two-dimensional drawing.

[0103] The motion modeling unit (330) creates a motion model by modeling the motion of each object included in the manufacturing plant. The motion modeling unit (330) can model the motion of equipment, materials, and processes. The motion modeling unit (330) can implement the motion model for equipment in the form of a function. For example, when modeling a roll, the shape modeling unit (320) can create a shape model by modeling the three-dimensional shape of the roll. The motion modeling unit (330) can create a motion model by implementing the rotational motion and vertical movement motion of the roll in the form of a function.

[0104] The motion modeling unit (330) can generate a motion model for a material in conjunction with a physics engine that simulates various physical phenomena. For example, when a roll presses the top and bottom of a slab during a rolling process, the slab may become flattened and widened due to physical phenomena. The motion modeling unit (330) can define the motion of the slab becoming flat and the motion of the slab becoming widened, and generate these as motion models.

[0105] The motion modeling unit (330) can generate a motion model for a process. The motion modeling unit (330) can generate a motion model by implementing each of the motions included in the process as a function and defining the order of the motions.

[0106] The state modeling unit (340) creates a state model by modeling changes according to the state of each object included in the manufacturing plant. The state modeling unit (340) can create a state model such that the color or shape changes or additional visualization effects appear depending on the state of the object.

[0107] The state modeling unit (340) can visualize invisible information such as temperature. For example, when modeling the temperature of a material surface, the state modeling unit (340) can generate a state model so that high temperature areas are represented in red and low temperature areas are represented in blue, so that the model can be visually represented in the shape of contour lines.

[0108] The virtualization model database (350) stores a virtualization model including a shape model, a motion model, and a state model for each of the objects included in the manufacturing plant. The virtualization model database (350) can store each of the shape model, motion model, and state model by mapping them to a virtual object. The virtual object can correspond to an object included in the actual manufacturing plant.

[0109] The virtual factory creation unit (350) creates a three-dimensional virtual object corresponding to an object using a virtualization model stored in the virtualization model database (350), and creates a virtual factory composed of virtual objects. The virtual factory creation unit (350) can reflect changes in the operation or state of the virtual objects included in the virtual factory based on virtual site data or virtual simulation data stored in the virtual database (315). Here, the virtual site data is data processed by the virtual data pipeline (310) based on the site data. The virtual simulation data is data processed by the virtual data pipeline (310) based on the simulation result data.

[0110] The virtual factory creation unit (350) can change at least one of the shape, operation, and state of at least one of the virtual objects in the virtual factory based on virtual field data or virtual simulation data.

[0111] A virtual factory created by the virtual factory creation unit (350) can be provided to a user. At this time, the virtual factory can be provided to the user through a virtual service unit. The virtual service unit may be divided into a first virtual service unit (360) and a second virtual service unit (365) depending on the method of access by the user.

[0112] The first virtual service unit (360) can provide a virtual space to the first terminal (410) equipped with a general-purpose operating system, such as a PC or a mobile device. The PC supports both client-server and web methods, and the mobile device can support both Android and iOS systems. The first virtual service unit (360) can support bidirectional interaction.

[0113] The second virtual service unit (365) can provide a virtual space to a second terminal (420), such as a Head Mounted Display (HMD) device capable of implementing Augmented Reality (AR), Virtual Reality (VR), and Extended Reality (XR). The second virtual service unit (365) can support bidirectional interaction.

[0114] A digital twin system (1) according to one embodiment of the present invention can implement objects such as equipment, materials, parts, products, and workers of a real manufacturing plant in a virtual plant through a virtualization unit (300). A digital twin system (1) according to one embodiment of the present invention can model not only the shapes of objects included in the manufacturing plant but also changes according to their operation and state. A digital twin system (1) according to one embodiment of the present invention can improve the visibility of the manufacturing plant site and visualize site data to enable users to make quick decisions. In addition, a digital twin system (1) according to one embodiment of the present invention can support remote operation through a virtual plant.

[0115] In addition, the digital twin system (1) according to one embodiment of the present invention is equipped with a virtual data pipeline (310) that connects data between the manufacturing plant site and a virtual space in real time, so that the equipment, materials, and movements or status of workers of the actual manufacturing plant can be displayed in the virtual plant in real time. The digital twin system (1) according to one embodiment of the present invention makes the virtual objects of the virtual plant identical to the actual manufacturing plant, so real-time monitoring is possible.

[0116] In addition, the digital twin system (1) according to one embodiment of the present invention can connect the simulation space to a virtual factory and reflect the simulation results in the virtual factory. That is, the digital twin system (1) according to one embodiment of the present invention can implement the simulation in the virtual factory. The user can intuitively check the simulation results through the virtual factory.

[0117] In addition, the digital twin system (1) according to one embodiment of the present invention provides a real-time connection unit (100), a simulation unit (200), and a virtualization unit (300) in the form of a single platform, thereby facilitating the construction of a simulation environment and making it easy to apply various models currently operating in a manufacturing plant back to the manufacturing plant after performing a simulation. The digital twin system (1) according to one embodiment of the present invention enhances connectivity between digital twins and the reuse of models, and enables the implementation of an automation function that automatically reflects changes in objects of the real manufacturing plant to the virtual plant.

[0118] In this digital twin system (1) according to one embodiment of the present invention, various tasks can be easily and quickly implemented on a virtual factory, and it is easy to expand the application of the digital twin.

[0119] Figure 5 is a diagram illustrating the simulation of equipment health in a digital twin system, and Figure 6 is a diagram showing the configuration of the equipment health simulation model of Figure 5. Figure 7a is a graph showing an example of simulation torque data over time, and Figure 7b is a table showing torque change and the number of cycles for the simulation torque data shown in Figure 7a. Figure 8a is a table showing an example of the minimum and maximum number of iterations per part, and Figure 8b is a graph showing an example of an SN curve for Part 1 of Figure 8a.

[0120] Referring to FIGS. 5 and 6, a digital twin system (1) according to one embodiment of the present invention can simulate the soundness of the equipment in a simulation unit (200).

[0121] Specifically, the simulation data pipeline (210) receives torque data for equipment health simulation. The simulation data pipeline (210) may include a field torque data receiving unit (212), a virtual torque data receiving unit (213), and a preprocessing unit (214).

[0122] The field torque data receiving unit (212) can receive field torque data. The field torque data may be generated from actual equipment and sensed by a sensor at the site where the manufacturing process is performed. The field torque data receiving unit (212) can receive field torque data in real time or periodically from the real-time connection unit (100).

[0123] The virtual torque data receiving unit (213) can receive virtual torque data. The virtual torque data may be created based on past production scenarios and may be predicted by an AI model (205). The AI ​​model (205) can predict torque data by equipment type based on past torque data by equipment type that occurred in the past. The torque data may vary depending on the type of material or product manufactured through the equipment. The AI ​​model (205) can also predict torque data for types of materials or products that are not manufactured in the actual equipment. The AI ​​model (205) may be a model stored in the simulation model database (225) of the simulation unit (200), but is not necessarily limited thereto. The AI ​​model (205) may also be a model stored externally. In this case, the virtual torque data receiving unit (213) can periodically receive virtual torque data by equipment type from the external AI model (205).

[0124] The preprocessing unit (214) preprocesses field torque data and virtual torque data. The preprocessing unit (214) can generate simulation torque data by preprocessing the field torque data and virtual torque data into a data format for performing equipment health simulation. The preprocessing unit (214) can store the simulation torque data in the simulation database (215).

[0125] The simulator (230, FIG. 3) executes the equipment health simulation model (500) to perform equipment health simulation. The simulator (230, FIG. 3) can perform equipment health simulation by reading the equipment health simulation model (500) from the simulation model database (225, FIG. 3).

[0126] The equipment health simulation model (500) can be generated by the simulation modeling unit (220, FIG. 3). Specifically, the simulation modeling unit (220, FIG. 3) can generate unit models by modeling objects to be simulated, and generate the equipment health simulation model (500) by connecting the unit models according to the equipment health simulation scenario. In this case, if the object is an intangible object stored externally in the form of a program, the unit model can be implemented as connection information that can access the object. The equipment health simulation model (500) may include input parameters, output parameters, operations or processes, and connection relationships with other unit models for each of the unit models.

[0127] The equipment health simulation model (500) may include unit models for each of the tangible objects, such as equipment placed at the manufacturing plant site, parts of the equipment, materials, and products. Additionally, the equipment health simulation model (500) may include intangible objects such as a stress prediction model (510), a remaining life prediction model (520), and an equipment health prediction model (530), as shown in FIG. 5. Furthermore, each of the stress prediction model (510), the remaining life prediction model (520), and the equipment health prediction model (530) may be connected to a stress model (610) and a plurality of formula models (620, 630, 64), as shown in FIG. 6.

[0128] The simulation modeling unit (220, FIG. 3) stores the generated equipment health simulation model (500) in the simulation model database (225, FIG. 3).

[0129] As described above, the simulator (230, FIG. 3) can perform equipment health simulation by reading and executing an equipment health simulation model (500) from the simulation model database (225, FIG. 3). This simulator (230, FIG. 3) includes a simulation management unit (240) and a simulation execution unit (245).

[0130] The simulation management unit (240) can call the simulation execution unit (245) to perform an equipment health simulation. Specifically, the simulation management unit (240) can read an equipment health simulation model (500) to be performed from the simulation model database (225). The simulation management unit (240) can read all unit models required to perform the equipment health simulation, the input and output parameters of each of all unit models, and the connection relationships between the unit models from the simulation model database (225).

[0131] Additionally, the simulation management unit (240) can read simulation torque data from the simulation database (215).

[0132] The simulation management unit (240) can call the simulation execution unit (245) and transmit the read equipment health simulation model (500) and simulation torque data to the called simulation execution unit (245). The simulation management unit (240) can sequentially transmit unit models to the simulation execution unit (245) according to the connection relationships of the unit models included in the equipment health simulation model (500), thereby performing simulations on the unit models.

[0133] The simulation management unit (240) can store the simulation result data, which is executed by the simulation execution unit (245), in the simulation result database (255, FIG. 3).

[0134] The simulation execution unit (245) is a simulation engine device that executes an equipment health simulation model (500) in response to a call from the simulation management unit (240). Specifically, the simulation execution unit (245) can create simulation objects corresponding to each of the equipment to be simulated and the parts included in the equipment, and can create a simulation space by connecting the simulation objects. The simulation execution unit (245) can perform an equipment health simulation in the simulation space using the equipment health simulation model (500) based on simulation torque data for the equipment.

[0135] The equipment health simulation model (500) may include a unit model, a stress prediction model (510), a remaining life prediction model (520), and an equipment health prediction model (530).

[0136] The stress prediction model (510) can predict stress values ​​for each part based on simulation torque data. To this end, the stress prediction model (510) may include a torque filtering unit (512) and a stress calculation unit (514) as shown in FIG. 6.

[0137] The torque filtering unit (512) filters simulation torque data that affects the remaining lifespan of the equipment among the simulation torque data. Specifically, if the torque value is large, the equipment may wear out as the stress value increases, and consequently, the remaining lifespan may be affected. On the other hand, if the torque value is small, the stress value is small, so it may not affect the remaining lifespan of the equipment. Accordingly, the torque filtering unit (512) can select only the simulation torque data that affects the remaining lifespan of the equipment due to the large torque value among the simulation torque data and use it for equipment health simulation. The torque filtering unit (512) measures the torque change over time for the simulation torque data and can select the simulation torque data in which the torque change exceeds a threshold value.

[0138] For example, simulation torque data may change over time as shown in FIG. 7a and FIG. 7b. The torque change may appear as 6 Nm in the interval from 0 ms to 50 ms, as 5 Nm in the intervals from 50 ms to 100 ms and from 450 ms to 500 ms, as 4 Nm in the interval from 100 ms to 150 ms, and as 2 Nm in the interval from 150 ms to 450 ms. When the threshold is set to 4.7, the torque filtering unit (512) can select simulation torque data in the interval where the torque change exceeds the threshold. The torque filtering unit (512) can select simulation torque data in the intervals from 0 ms to 50 ms, from 50 ms to 100 ms, and from 450 ms to 500 ms. In one embodiment, the torque filtering unit (512) can filter the simulation torque data using a Rainflow algorithm.

[0139] The torque filtering unit (512) can transmit the filtered simulation torque data and the number of cycles indicating the number of times the filtered simulation torque data is repeated to the stress calculation unit (514).

[0140] The stress calculation unit (514) can calculate stress values ​​for each part based on simulation torque data filtered by the torque filtering unit (512).

[0141] In one embodiment, the stress calculation unit (514) can calculate a stress value for each part using a stress model (610). The stress calculation unit (514) can transmit simulation torque data filtered by the torque filtering unit (512) to the stress model (610). The stress model (610) may be a unit model included in the equipment health simulation model (500), but is not necessarily limited thereto. The stress model (610) may also be a model stored externally. In this case, the stress calculation unit (514) can access the external stress model (610) using connection information that allows access to the stress model (610). The stress calculation unit (514) can transmit simulation torque data filtered by the torque filtering unit (512) to the external stress model (610).

[0142] The stress model (610) may include a physical analysis engine that performs physical analysis for each of the equipment and the parts included in the equipment. The stress model (610) may calculate stress values ​​for each part based on information regarding the parts included in the equipment, historical performance data regarding the equipment and parts, and simulation torque data regarding the equipment and parts. At this time, the stress values ​​calculated for each part may be one or more. One torque data may correspond to one torque data, but is not necessarily limited thereto. If there are multiple filtered simulation torque data, multiple stress values ​​may be calculated. Meanwhile, the calculated stress values ​​may correspond to the number of cycles of the corresponding simulation torque data.

[0143] In one embodiment, the stress model (610) may include a plurality of physical analysis engines divided into parts. In this case, each of the plurality of physical analysis engines can calculate a stress value of a part based on simulation torque data for that part.

[0144] The stress model (610) can transmit stress data including the calculated stress values ​​for each part to the stress calculation unit (514).

[0145] The stress prediction model (510) can transmit stress data, including the number of cycles obtained by the torque filtering unit (512) and the stress values ​​per part obtained by the stress calculation unit (514), to the remaining life prediction model (520).

[0146] The remaining life prediction model (520) predicts the remaining life of each part based on the stress value and the number of cycles for each part. To this end, the remaining life prediction model (520) may include a wear calculation unit (522) and a remaining life calculation unit (524) as shown in FIG. 6.

[0147] The wear calculation unit (522) calculates the degree of wear indicating the degree of wear for each part based on the stress value and number of cycles for each part.

[0148] In one embodiment, the wear calculation unit (522) can calculate the wear of each part using a wear formula model (620). The wear calculation unit (522) can transmit the stress value and cycle count for each part received from the stress prediction model (510) to the wear formula model (620). The wear formula model (620) may be a unit model included in the equipment health simulation model (500), but is not necessarily limited thereto. The wear formula model (620) may also be a model stored externally. In this case, the wear calculation unit (522) can access the external wear formula model (620) using connection information that allows access to the wear formula model (620). The wear calculation unit (522) can transmit the stress value and cycle count for each part to the external wear formula model (620).

[0149] The wear rate formula model (620) can calculate the wear rate using the S (Stress)-N (Number of Cycle to Failure) curve function. The wear rate formula model (620) can determine the SN curve for each part. The SN curve can represent the number of cycles until damage occurs to the part for each stress. For example, as shown in FIG. 8a, the number of cycles until damage occurs to the part for each of the multiple parts can be set for each stress. At this time, the number of cycles until damage occurs to the part for each stress can be determined through repeated experiments. The SN curve may correspond to a curve created to represent stress on the vertical axis and the number of cycles until damage occurs to the part (Number of Cycle to Failure) on the horizontal axis. Such an SN curve can be determined for each part.

[0150] The wear equation model (620) can derive a straight line equation connecting the first point, which is the lowest number of repetitions (Cycle_Low), and the second point, which is the highest number of repetitions (Cycle_High) in the SN diagram. The wear equation model (620) can use the derived straight line equation to obtain the number of repetitions corresponding to the stress value predicted by the stress prediction model (510), and calculate the wear as the ratio of the number of cycles to the number of repetitions obtained.

[0151] For example, Part 1 has a stress of 40 Kg / mm 2 When this happens, the minimum number of repetitions (Cycle_Low) becomes 1.0E+4, and the stress is 30 Kg / mm 2 When this occurs, the maximum number of repetitions (Cycle_High) can be 3.0E+6. If the first point, which is the minimum number of repetitions (Cycle_Low), and the second point, which is the maximum number of repetitions (Cycle_High), are connected, an SN curve as shown in FIG. 8b can be obtained. The wear equation model (620) can derive a linear equation using the SN curve shown in FIG. 8b.

[0152] The wear formula model (620) can obtain the number of iterations corresponding to the stress value predicted by the stress prediction model (510) using the derived linear equation. The stress value calculated by the stress prediction model (510) is 40 Kg / mm 2 The cycle count for is 0.5, and the stress value is 35 Kg / mm 2 The number of cycles for can be 1.0. The wear formula model (620) uses a linear equation to obtain a stress value of 40 Kg / mm 2 Obtain a repetition count of 1.0E+4 corresponding to, and a stress value of 35 Kg / mm 2 A repetition count of 15E+5 corresponding to can be obtained. The wear formula model (620) has a stress value of 40 Kg / mm 2For the first wear degree, the ratio of a cycle count of 0.5 to a repetition count of 1.0E+4 can be calculated. The first wear degree is a stress value of 40 Kg / mm 2 To destroy the part, 1.0E+4 repetitions must be performed, but this is a numerical representation of the current 0.5 repetitions. Additionally, the wear formula model (620) has a stress value of 35 Kg / mm 2 For the second wear degree, the ratio of a cycle count of 1.0 to a repetition count of 15E+5 can be calculated. The second wear degree is a stress value of 35 Kg / mm 2 To destroy the part, 15E+5 repetitions must be performed, but this is a numerical representation of the current 1.0 repetitions.

[0153] The wear formula model (620) can calculate the wear of a corresponding part by summing the first wear and the second wear. That is, the wear formula model (620) can calculate the wear by summing the ratio of the number of cycles to the number of repetitions for each of the multiple stress values. The wear formula model (620) can calculate the wear for each part.

[0154] The wear formula model (620) can transmit information regarding the wear of each calculated part to the wear calculation unit (522). The wear calculation unit (522) can transmit the wear of each part to the remaining life calculation unit (524).

[0155] The remaining lifespan calculation unit (524) predicts the remaining lifespan of each part based on the degree of wear of each part. In one embodiment, the remaining lifespan calculation unit (524) can calculate the remaining lifespan of each part using a remaining lifespan formula model (630). The remaining lifespan calculation unit (524) can transmit the degree of wear of each part received from the degree of wear calculation unit (522) to the remaining lifespan formula model (630). The remaining lifespan formula model (630) may be a unit model included in the equipment health simulation model (500), but is not necessarily limited thereto. The remaining lifespan formula model (630) may also be a model stored externally. In this case, the remaining lifespan calculation unit (524) can access the external remaining lifespan formula model (630) using connection information that allows access to the remaining lifespan formula model (630). The remaining life calculation unit (524) can transmit the degree of wear for each part to an external remaining life formula model (630).

[0156] The remaining lifespan formula model (630) can calculate the expected lifespan using the accumulated usage time and the degree of wear, and calculate the remaining lifespan by subtracting the accumulated usage time from the calculated expected lifespan. The remaining lifespan formula model (630) can calculate the expected lifespan by dividing the accumulated usage time by the degree of wear.

[0157] In one embodiment, the remaining life formula model (630) can calculate the current life of each part based on past torque data. The equipment health simulation model (500) can store and manage past torque data from a predetermined period prior to the current life in a separate database. For example, the equipment health simulation model (500) can store and manage past torque data from two years prior to the current life in a separate database. The remaining life formula model (630) can also calculate the remaining life by subtracting the current life from the expected life, using the current life of each part as the initial value.

[0158] In one embodiment, the remaining life formula model (630) can calculate the remaining life by reflecting past maintenance history data.

[0159] The remaining life formula model (630) can transmit the remaining life for each part to the remaining life calculation unit (524). The remaining life calculation unit (524) can transmit the remaining life for each part to the equipment health prediction model (530).

[0160] The equipment health prediction model (530) predicts equipment health based on the remaining lifespan of each component. To this end, the equipment health prediction model (530) may include an equipment health calculation unit (532).

[0161] The equipment health calculation unit (532) calculates the equipment health based on the remaining lifespan of each component. In one embodiment, the equipment health calculation unit (532) can calculate the health of the equipment using an equipment health formula model (640). The equipment health calculation unit (532) can transmit the remaining lifespan of each component received from the remaining lifespan prediction model (520) to the equipment health formula model (640). The equipment health formula model (640) may be a unit model included in the equipment health simulation model (500), but is not necessarily limited thereto. The equipment health formula model (640) may also be a model stored externally. In this case, the equipment health calculation unit (532) can access the external equipment health formula model (640) using connection information that allows access to the equipment health formula model (640). The equipment health calculation unit (532) can transmit the remaining lifespan of each part to an external equipment health formula model (640).

[0162] The equipment health formula model (640) can calculate the health score of the equipment by statistically processing the remaining lifespan of each part. The equipment health formula model (640) can calculate the health score for each piece of equipment.

[0163] The equipment health formula model (640) can transmit the health score for each piece of equipment to the equipment health calculation unit (532). The equipment health calculation unit (532) can generate simulation result data including the health score for each piece of equipment and transmit the simulation result data to the simulation management unit (240).

[0164] The equipment health simulation model (500) may generate simulation result data for each unit model, which is a unit of simulation execution. The equipment health simulation model (500) may generate simulation result data from the results of simulations performed in each of the stress prediction model (510), the remaining life prediction model (520), and the equipment health prediction model (530). The stress prediction model (510) may generate information regarding stress values ​​and cycle counts for each component, which is the result of performing a stress prediction simulation, as the first simulation result data. The remaining life prediction model (520) may generate information regarding the remaining lifespan for each component, which is the result of performing a remaining life prediction simulation, as the second simulation result data. The equipment health prediction model (530) may generate information regarding the health score for each piece of equipment, which is the result of performing an equipment health simulation, as the third simulation result data.

[0165] The simulation management unit (240) can perform a simulation according to the equipment health simulation model (500) and store the generated first to third simulation result data in the simulation result database (255).

[0166] A digital twin system (1) according to one embodiment of the present invention can perform simulations using not only torque data generated in actual equipment but also torque data generated in a virtual environment. In addition, a digital twin system (1) according to one embodiment of the present invention can accurately calculate stress values ​​by selecting torque data that affects the remaining lifespan. Through this, the accuracy of the remaining lifespan is increased in the digital twin system (1) according to one embodiment of the present invention.

[0167] In addition, a digital twin system (1) according to one embodiment of the present invention can analyze equipment health at all times and predict and analyze equipment health for up to 10 years to estimate the repair and replacement cycle of parts. A digital twin system (1) according to one embodiment of the present invention can extend the usage period of equipment or parts to a maximum value considering a safety factor from the remaining lifespan.

[0168] In addition, the digital twin system (1) according to one embodiment of the present invention can shorten the analysis time because the health of the equipment is predicted through an equipment health simulation model (500).

[0169] Figure 9 is a diagram illustrating the processing of simulation result data in a digital twin system.

[0170] A digital twin system (1) according to one embodiment of the present invention performs a simulation of an object in a simulation unit (200) and can reflect the results of the simulation performed by the simulation unit (200) in a virtual space virtualized by a virtualization unit (300). To this end, the simulation unit (200) provides simulation result data to the virtualization unit (300).

[0171] The simulation unit (200) generates a large amount of simulation result data whenever a simulation is performed. In particular, a physical analysis model (290) that performs physical analysis on an object can generate hundreds of thousands of simulation result data in a single simulation. In such cases, the simulation unit (200) can transmit hundreds of thousands of simulation result data to the virtualization unit (300). The virtualization unit (300) receives hundreds of thousands of simulation result data and performs rendering processing on the received hundreds of thousands of simulation result data to display them in a virtual space. The virtualization unit (300) may experience a heavy load in receiving a large amount of simulation result data and performing rendering processing on them, and consequently, the simulation result data may not be reflected in the virtual space in real time. Furthermore, since the program performing the rendering work requires a heavy load, the virtualization unit (300) may consume a large amount of resources and time due to a significant increase in memory usage and load.

[0172] In order to effectively process simulation result data, the digital twin system (1) according to one embodiment of the present invention may transmit only the simulation result data in which data changes have occurred to the virtualization unit (300), rather than transmitting all simulation result data to the virtualization unit (300).

[0173] The simulation unit (200) can select simulation result data that is different from the (n-1)th simulation result data among the nth simulation result data and transmit only the selected nth simulation result data to the virtualization unit (300). To this end, the simulation unit (200) includes a data interface unit (280) composed of a queue manager (282), a queue storage (284), a data selection unit (286), and a data transmission unit (288), as shown in FIG. 5.

[0174] The queue manager (282) groups the nth simulation result data into multiple groups. The nth simulation result data may be data stored in the simulation result database (255), or data directly transmitted from a simulation model or unit model stored inside or outside the system. For example, the queue manager (282) may receive the physical analysis results for an object as simulation result data from a physical analysis model (290) stored externally in the form of a program, as shown in FIG. 5.

[0175] Data changes can occur frequently in parts of an object where forces are generated by physical pressure or where material properties change due to heat, humidity, or chemical components. In other words, data changes may occur frequently primarily in the drive parts of equipment or in areas where friction with the material occurs. Conversely, data changes may not occur frequently in parts that are not directly subjected to force and where physical analysis is not visually applied. For example, in rolling equipment included in a rolling process, physical analysis results may change more frequently in areas closer to the motor, while the variation may decrease as the distance from the motor increases.

[0176] The queue manager (282) can divide the nth simulation result data into predetermined group units and group them into multiple groups. The multiple groups can be determined based on the past data change rate. The past data change rate can be determined based on the change value of the simulation result values ​​of past simulation result data that occurred before the nth simulation result data.

[0177] Each of the multiple simulation result values ​​generated at a specific point in time may correspond to a location within the object being simulated. For example, a first simulation result value generated at a specific point in time may be a simulation result value for a first location within the object, and a second simulation result value may be a simulation result value for a second location within the object. Accordingly, each of the multiple simulation result values ​​generated at a specific point in time may be a simulation result value for each of the locations within the object being simulated. Accordingly, the simulation result data may include a simulation result value and location information corresponding to the simulation result value.

[0178] The degree of change in past data may correspond to the difference between past simulation results at multiple time points corresponding to the same location within an object. For example, the degree of change in past data may correspond to the difference between the simulation result at a first time point and the simulation result at a second time point corresponding to the same location within an object.

[0179] The queue manager (282) can group the nth simulation result data into a single group if the past data change rate is within a similar range. For example, the queue manager (282) can group simulation result data with no past data change rate into a single group. For another example, the queue manager (282) can group simulation result data with a low past data change rate into a single group. For yet another example, the queue manager (282) can group simulation result data with a high past data change rate into a single group. In this case, the range of the past data change rate may vary according to user settings.

[0180] The queue manager (282) can group simulation result data into multiple groups by utilizing location information within the object in addition to the past data change map. The queue manager (282) can group simulation result data into one group where the past data change map is within a similar range and the location information within the object is adjacent. The object is divided into multiple regions based on the past data change map and the location within the object, and the multiple groups can correspond to each of the multiple regions. The queue manager (282) can divide the simulation result data by region of the object using the location information within the object of the simulation result data, and group the simulation result data divided into multiple regions into groups.

[0181] In one embodiment, the queue manager (282) can group simulation result data into multiple groups based on multiple pre-separated regions within an object. The multiple regions within an object can be pre-set by the user. The user can divide the object into multiple regions based on past simulation result data.

[0182] In another embodiment, the queue manager (282) can automatically set multiple regions within an object. The queue manager (282) determines the degree of change in past data by location within the object using the change values ​​of the simulation result values ​​of past simulation result data that occurred prior to the nth simulation result data, and can divide the object into multiple regions based on the degree of change in past data.

[0183] The queue manager (282) can calculate a change value by comparing the simulation result value of the simulation result data at the first time point and the simulation result value of the simulation result data at the second time point according to the location, and can determine the change in past data according to the calculated change value.

[0184] The queue manager (282) can divide an object into multiple regions based on the historical data change rate by location. For example, the queue manager (282) can divide the object into a region containing locations with no historical data change rate and a region containing locations with a historical data change rate, but is not necessarily limited thereto. As another example, the queue manager (282) can divide the object into multiple regions based on the historical data change rate. The queue manager (282) can divide the object into a region containing locations where the historical data change rate is within a first range, a region containing locations where the historical data change rate is within a second range greater than the first range, and a region containing locations where the historical data change rate is within a third range greater than the second range.

[0185] The queue manager (282) can divide each of the areas containing locations with no past data change and the areas containing locations with past data change into multiple parts. The queue manager (282) can divide the areas containing locations with no past data change into multiple parts based on the location within the object.

[0186] The queue manager (282) can divide the area containing the locations with past data changes into multiple areas based on the location within the object and the past data change rate. The queue manager (282) can divide the area into the area containing the locations where the past data change rate is within a first range, the area containing the locations where the past data change rate is within a second range greater than the first range, and the area containing the locations where the past data change rate is within a third range greater than the second range.

[0187] The queue manager (282) can divide the area into smaller sizes as the degree of change in past data increases. That is, the queue manager (282) can divide the area with a small degree of change in past data into relatively large sizes and divide the area with a large degree of change in data into relatively small sizes.

[0188] Additionally, the queue manager (282) can divide the area such that the number of simulation result data within the area decreases as the degree of change in past data increases. That is, the queue manager (282) can have a relatively large number of simulation result data in the area where the degree of change in past data is small, and a relatively small number of simulation result data in the area where the degree of change in data is large.

[0189] The queue manager (282) stores simulation result data grouped into multiple groups along with group information in the queue storage (284). The group information may indicate the relative position of an area within an object corresponding to the group among the multiple groups. The nth grouped simulation result data may be stored in the queue storage (284) along with the group information.

[0190] In one embodiment, the queue manager (282) can determine the priority for each of the plurality of groups. The queue manager (282) can have the group with a large degree of historical data change have a high priority, and the group with a low or no degree of historical data change have a low priority. That is, the queue manager (282) can determine that the group in which historical data changes frequently occur has a higher priority than the group in which historical data changes rarely occur. The queue manager (282) can store the grouped nth simulation result data in the queue storage (284) along with group information and priority information.

[0191] The data selection unit (286) selects a group in which a data change has occurred among a plurality of groups stored in the queue storage (284). The data selection unit (286) may create a thread to determine whether there is a data change for the nth simulation result data included in the group. The data selection unit (286) creates threads within a range permissible by the system and may create multiple threads to enable parallel processing.

[0192] The queue manager (282) can assign multiple groups stored in the queue storage (284) to multiple threads generated by the data selection unit (286). The queue manager (282) assigns threads in order of highest priority and assigns threads to groups with lower priority later.

[0193] The data selection unit (286) can determine whether a data change has occurred by comparing the n-th simulation result data and the (n-1)-th simulation result data of the group assigned in each of the multiple threads.

[0194] In one embodiment, the data selection unit (286) can convert the nth simulation result data of a group assigned to each of the plurality of threads into a hash and determine whether there is a change in the data by comparing it with the hash of the (n-1)th simulation result data. Specifically, each thread can convert the nth simulation result data of the assigned group into a first hash value according to a predetermined rule. Each thread can compare the first hash value with a second hash value (Pre Hash) converted based on the (n-1)th simulation result data of the assigned group.

[0195] The data selection unit (286) can determine that no data change has occurred in the nth simulation result data of the group if the first hash value (Hash) and the second hash value (Pre Hash) are the same. The data selection unit (286) can determine that no data change has occurred in the nth simulation result data of the group if the first hash value (Hash) and the second hash value (Pre Hash) are different.

[0196] The data selection unit (286) can select a group in which a data change has occurred and transmit the nth simulation result data included in at least one selected group, along with group information, to the virtualization unit (300) through the data transmission unit (288).

[0197] The data transmission unit (288) transmits the nth simulation result data of the group selected by the data selection unit (286) as the group in which data change occurred, along with group information, to the virtualization unit (300). The data transmission unit (288) does not transmit all nth simulation result data generated from a single simulation execution, but only transmits the nth simulation result data included in the group selected by the data selection unit (286).

[0198] The virtualization unit (300) renders a virtual object corresponding to an object and outputs it on the screen of the terminal, and modifies the rendered virtual object based on field data or simulation result data. The virtualization unit (300) can render a first area of ​​the virtual object where no state change has occurred using the n-1th simulation result data, and render a second area where a state change has occurred using the nth simulation result data.

[0199] The virtualization unit (300) receives the nth simulation result data and group information of a selected group and can update a virtual object based on the received nth simulation result data of the selected group. To this end, the virtualization unit (300) includes a virtual factory creation unit (350) composed of a rendering unit (352) and a parallel controller (354).

[0200] The virtual data pipeline (310) can receive the nth simulation result data and group information of a selected group from the data transmission unit (288) of the simulation unit (200) and store them in the virtual database (315). The virtual data pipeline (310) can process the nth simulation result data and group information of the selected group into virtual simulation data to be displayed in the virtual factory and store them in the virtual database (315).

[0201] The virtual factory creation unit (350) creates virtual objects corresponding to objects included in the manufacturing process and creates a virtual factory composed of multiple virtual objects. This virtual factory creation unit (350) includes a rendering unit (352) and a parallel controller (354).

[0202] The rendering unit (352) renders the object to visualize it as a three-dimensional virtual object. The rendering unit (352) can render the virtual object using a virtualization model that includes a shape model that models the three-dimensional shape of the object, a motion model that models the motion, and a state model that models the change according to the state.

[0203] The rendering unit (352) can output the rendered virtual object to the screen of the terminal through the virtual service unit. In addition to the virtual object output to the screen of the terminal, the rendering unit (352) can additionally render a buffer object identical to the virtual object. The buffer object is identical to the virtual object, but unlike the virtual object, it may not be output to the screen of the terminal.

[0204] The parallel controller (354) manages virtual objects and buffer objects. Specifically, when the parallel controller (354) receives new simulation result data and group information for the nth group, it can control the update of virtual objects and buffer objects based on the newly received simulation result data and group information for the nth group. The parallel controller (354) can update buffer objects based on the received simulation result data and group information for the nth group. The parallel controller (354) can update the (n-1)th simulation result data of the area corresponding to the group information within the buffer object with the received simulation result data for the nth group. At this time, the received simulation result data for the nth group may include only the simulation result data of the group in which a data change occurred among the simulation result data generated during the simulation process. Accordingly, the parallel controller (354) can update only the area in which a data change occurred, without updating all areas of the buffer object. Accordingly, while the virtual object according to the n-1th simulation result is being output, the parallel controller (354) can update the buffer object by rendering the first area where no state change of the buffer object has occurred using the n-1th simulation result data and rendering the second area where a state change of the buffer object has occurred using the nth simulation result data.

[0205] When the update of the buffer object is completed, the parallel controller (354) can control the updated buffer object to be converted into a virtual object and displayed on the terminal screen. Additionally, the parallel controller (354) can convert a virtual object that was being displayed on the terminal screen into a buffer object. The converted buffer object can serve as a buffer object to reflect new simulation result data when new simulation result data is received.

[0206] In a digital twin system (1) according to one embodiment of the present invention, when a simulation unit (200) processes a large amount of simulation result data in real time, the simulation result data is divided into groups, and only the groups in which data changes have occurred are selected and transmitted to a virtualization unit (300). In a digital twin system (1) according to one embodiment of the present invention, since the virtualization unit (300) only needs to perform rendering processing on the simulation result data of the selected groups, rendering resources can be reduced. In a digital twin system (1) according to one embodiment of the present invention, memory usage and load of the virtualization unit (300) are reduced, and the rendering processing speed for simulation result data is improved, thereby ensuring fast response performance for the user.

[0207] Those skilled in the art to which the present invention pertains will understand that the above-described invention may be implemented in other specific forms without altering its technical concept or essential features.

[0208] Those skilled in the art to which this name belongs will understand that the above-described invention may be implemented in other specific forms without altering its technical concept or essential features.

[0209] Therefore, the embodiments described above should be understood as illustrative in all respects and not limiting. The scope of the invention is defined by the claims set forth below rather than by the detailed description above, and all modifications or variations derived from the meaning and scope of the claims and equivalent concepts thereof should be interpreted as being included within the scope of the invention. Explanation of the symbols

[0210] 1: Digital Twin System 10: Device 20: Network 30: Digital Twin Platform 40: Application 100: Real-time connection section 200: Simulation Department 300: Virtualization Department 400: Virtualization Division 500: Standardization Division

Claims

Claim 1 A digital twin system for a manufacturing process including an equipment health simulation model, comprising: a simulation data pipeline receiving field torque data for equipment subject to simulation and virtual torque data predicted by an AI model; an equipment health simulation model that determines the health of the equipment by predicting the stress values ​​of each component included in the equipment based on simulation torque data including field torque data and virtual torque data; and a simulator that performs a simulation on the equipment based on the equipment health simulation model to generate simulation result data, wherein the equipment health simulation model includes a remaining life prediction model that calculates the expected life based on the cumulative usage time of the component and the wear degree of each component, and calculates the remaining life of the component by subtracting the cumulative usage time from the calculated expected life. Claim 2 A digital twin system for a manufacturing process including an equipment health simulation model according to claim 1, wherein the equipment health simulation model includes a stress prediction model that filters simulation torque data affecting the remaining lifespan of the equipment among the simulation torque data and predicts stress values ​​for each part based on the filtered simulation torque data. Claim 3 A digital twin system for a manufacturing process including an equipment health simulation model, wherein, in paragraph 2, the stress prediction model measures the change in torque over time with respect to the simulation torque data and selects simulation torque data among the simulation torque data in which the change in torque exceeds a threshold value. Claim 4 A digital twin system for a manufacturing process including an equipment health simulation model, wherein, in paragraph 2, the stress prediction model transmits the filtered simulation torque data to an externally stored stress model and obtains the stress values ​​for each part through the stress model. Claim 5 A digital twin system for a manufacturing process including an equipment health simulation model, wherein, in claim 1, the equipment health simulation model determines the health of the equipment by predicting the remaining lifespan of each part based on the predicted stress value of each of the parts. Claim 6 A digital twin system for a manufacturing process including an equipment health simulation model, wherein, in claim 5, the remaining life prediction model predicts the degree of wear for each part based on the number of cycles representing the number of times the simulation torque data is repeated and the predicted stress value of each of the parts, and predicts the remaining life for each part based on the degree of wear for each part. Claim 7 A digital twin system for a manufacturing process including an equipment health simulation model, wherein, in claim 6, the remaining life prediction model obtains a number of repetitions corresponding to a predicted stress value using an SN curve function representing the number of repetitions until damage occurs to a part according to stress, and calculates the ratio of the number of cycles to the obtained number of repetitions as the wear degree for each part. Claim 8 delete Claim 9 A digital twin system for a manufacturing process including an equipment health simulation model, wherein, in claim 1, the equipment health simulation model includes an equipment health prediction model that calculates the health of the equipment as a score by statistically processing the remaining lifespan of each component included in the equipment. Claim 10 A digital twin system for a manufacturing process including an equipment health simulation model, characterized in that, in claim 1, it includes a virtualization unit that renders a virtual object corresponding to each of the equipment and the parts included in the equipment, and modifies the rendered virtual object based on at least one of the field torque data and the simulation result data. Claim 11 A digital twin system for a manufacturing process including an equipment health simulation model, characterized in that, in claim 10, it further includes a data interface unit that selects simulation result data different from the (n-1)th simulation result data among the n-th simulation result data and transmits it to the virtualization unit, wherein the virtualization unit includes a virtual factory generation unit that renders a first area among the virtual objects where no state change has occurred using the (n-1)th simulation result data, and renders a second area where a state change has occurred using the n-th simulation result data.

Images