KR-20260063728-A - DIGITAL TWIN SYSTEM FOR MANUFACTURING PROCESSES

KR20260063728AKR 20260063728 AKR20260063728 AKR 20260063728AKR-20260063728-A

Abstract

A digital twin system for a manufacturing process according to one aspect of the present invention comprises: a simulation unit including a simulation model that performs a simulation according to a simulation scenario using data from a manufacturing process and generates simulation result data; and an optimization unit that calculates the optimization result data by adjusting the parameters of the simulation model so that the simulation model satisfies an objective function and constraints. The optimization unit includes first and second candidate optimization result data calculation units, and the first and second candidate optimization result data calculation units include different algorithms or different artificial intelligence models for adjusting the parameters of the simulation model.

Inventors

양원모
윤지상
김준영
신국남
박승도
김태석
이성훈
손정현
권태웅
김동언
현진원

Assignees

주식회사 포스코디엑스

Dates

Publication Date: 20260507
Application Date: 20241031

Claims (17)

A simulation unit including a simulation model that performs a simulation according to a simulation scenario using data from a manufacturing process and generates simulation result data; and The optimization unit for producing the optimization result data by adjusting the parameters of the simulation model so that the simulation model satisfies the objective function and constraints; A digital twin system for a manufacturing process, wherein the optimization unit includes first and second candidate optimization result data output units, and the first and second candidate optimization result data output units include different algorithms or different artificial intelligence models for adjusting parameters of the simulation model.
In paragraph 1, The above optimization unit is, It further includes an optimization management unit that manages the process and execution time of the first and second candidate optimization result data output units mentioned above, and The above-mentioned first candidate optimization result data calculation unit calculates first candidate optimization result data using at least one of a mathematical algorithm and a statistical algorithm, and The above-mentioned second candidate optimization result data output unit outputs second candidate optimization result data using at least one of at least one deep learning model and at least one reinforcement learning model, and A digital twin system for a manufacturing process, wherein the optimization management unit executes the first and second candidate optimization result data output units simultaneously only for a predetermined limited execution time.
In paragraph 1, The above optimization unit is, It further includes an optimization management unit that manages the process and execution time of the first and second candidate optimization result data output units mentioned above, and The above-mentioned first and second candidate optimization result data calculation units each calculate the first candidate optimization result data and the second candidate optimization result data, and A digital twin system for a manufacturing process, wherein the optimization management unit determines one of the first and second candidate optimization result data as the optimization result data.
In paragraph 3, A digital twin system for a manufacturing process, wherein the above optimization result data includes at least one of the results of performing iterative simulations and the calculation results of the above objective function, parameter values for the next simulation, whether constraints on parameter values are satisfied, an optimization method, and option values, and is stored in an optimization result database.
In paragraph 1, The above optimization unit is, It further includes an optimization management unit that manages the process and execution time of the first and second candidate optimization result data output units mentioned above, and The above-mentioned first candidate optimization result data calculation unit calculates first candidate optimization result data using at least one of a mathematical algorithm and a statistical algorithm, and The above-mentioned second candidate optimization result data output unit outputs second candidate optimization result data using at least one of at least one deep learning model and at least one reinforcement learning model, and The above optimization management unit is, The above first and second candidate optimization result data output units are executed simultaneously only for a predetermined limited execution time, and If the above first candidate optimization result data includes a global solution, the above first candidate optimization result data is determined as the optimization result data, and A digital twin system for a manufacturing process that determines the second candidate optimization result data as the optimization result data when the first candidate optimization result data does not include a global solution.
In paragraph 1, The above optimization unit is, A digital twin system for a manufacturing process further comprising an optimization parallel processing unit that divides and parallelizes at least one of the first and second candidate optimization result data output units.
In paragraph 1, A digital twin system for a manufacturing process, wherein the second candidate optimization result data output unit calculates at least one simulation scenario using at least one of at least one deep learning model and at least one reinforcement learning model, and calculates the second candidate optimization result data using simulation result data received from the simulation model by repeatedly executing the simulation model corresponding to each simulation scenario.
In paragraph 1, A digital twin system for a manufacturing process that further adjusts the variables, attributes, and parameters of the simulation model before performing the simulation to improve the prediction accuracy of the simulation model.
In paragraph 1, It further includes a real-time connection unit that generates field data using microdata collected from the manufacturing process, and The above optimization unit is, A digital twin system for a manufacturing process, further comprising: a simulation model verification unit that measures the prediction accuracy of the simulation model by comparing the field data and the simulation result data corresponding to the field data, and retrains and remodels the simulation model when a retraining condition is satisfied.
In Paragraph 9, The above relearning condition is when the measured prediction accuracy is lower than a predetermined reference prediction level, and at least one of a certain period, a digital twin system for a manufacturing process.
In paragraph 1, The above simulation unit is, A simulation modeling unit that models objects to be simulated to create unit models, connects the unit models according to a simulation scenario to create a simulation model, maps the created simulation model to the simulation scenario, and stores it in a simulation model database; and A simulator that performs the simulation using the simulation model to produce the simulation result data; A digital twin system for a manufacturing process, comprising a simulation model corresponding to a simulation scenario to be performed, and result data for the simulation provided to the optimization unit.
In Paragraph 11, A digital twin system for a manufacturing process, wherein the parameters of the simulation model include the parameters of the unit models constituting the simulation model.
In paragraph 1, The above optimization unit is, It further includes an optimization modeler that receives the objective function from the user and stores it in an objective function database, and receives the constraints in real time and stores them in a constraint database. A digital twin system for a manufacturing process, wherein the first and second candidate optimization result data output unit calculates the first and second candidate optimization results, respectively, using at least one of the objective functions stored in the objective function database and at least one of the constraints stored in the constraint database.
In paragraph 1, It further includes a real-time connection unit that generates field data using microdata collected from the manufacturing process, and The above optimization unit is, A digital twin system for a manufacturing process comprising: an optimization data pipeline that receives the simulation model and simulation result data generated from the simulation model from the simulation unit and transmits them to the first and second candidate optimization result data output unit, and receives the optimization result data from the first and second candidate optimization result data output unit and transmits it to the simulation unit.
In paragraph 1, A digital twin system for a manufacturing process, wherein the parameters of the above simulation model are hyperparameters of the above simulation model.
In paragraph 1, The above simulation unit is, A digital twin system for a manufacturing process, comprising an automatic control unit that generates control values for the manufacturing process based on the simulation result data and controls equipment placed in the manufacturing process according to the generated control values.
In paragraph 1, The method further includes a virtualization unit that renders a virtual object corresponding to each of the objects in the above manufacturing process and modifies the rendered virtual object based on at least one of the field data and the simulation result data. The simulation unit selects simulation result data that is different from the (n-1)th simulation result data among the nth simulation result data and transmits it to the virtualization unit, and A digital twin parallel processing system for a manufacturing process, characterized in that 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 nth simulation result data.

Description

Digital Twin System for Manufacturing Processes The present invention relates to a digital twin system, and more specifically, to a digital twin system for a manufacturing process for producing a product. 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. Existing factory management systems collect and process large volumes of data in real time and store them in big data repositories. By analyzing this stored data, they can determine the presence of abnormalities in intermediate goods and products or monitor equipment failures in real time. However, while existing systems can easily monitor and analyze the operational status of actual manufacturing plants, they do not easily create virtual factories identical to the real ones to support operations remotely, nor can they easily simulate various operational situations and strategies based on such virtual factories. FIG. 1 is a block diagram schematically showing a digital twin system for a manufacturing process according to one embodiment of the present invention. Figure 2 is a block diagram showing the configuration of the real-time connection part of the digital twin platform. Figure 3 is a block diagram showing the configuration of the simulation section of the digital twin platform. Figure 4 is a block diagram showing the configuration of the optimization section of the digital twin platform. Figure 5 is a block diagram showing the configuration of the optimization engine of the digital twin platform. Figure 6 is a diagram showing the configuration of the virtualization section of the digital twin platform. Figure 7 is a diagram illustrating the processing of simulation result data in a digital twin system. Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings. The meaning of the terms described in this specification should be understood as follows. 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. Terms such as "include" or "have" should be understood as not excluding in advance the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. 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. Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings. Hereinafter, embodiments according to the present invention will be described in detail with reference to the attached drawings. FIG. 1 is a schematic diagram showing a digital twin system for a manufacturing process according to one embodiment of the present invention. 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). 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. 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 process site. 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. For exa