Equipment tuning system using digital twin and equipment tuning method using same

Abstract

Disclosed are an equipment tuning system using a digital twin and an equipment tuning method using same, wherein the equipment tuning system comprises a digital twin unit and an equipment control unit. The digital twin unit sets test conditions for optimizing control parameters of equipment. The equipment control unit drives the equipment on the basis of the set test conditions to obtain test results for each test condition. In this case, the digital twin unit resets the test conditions until optimal control parameters of the equipment are derived on the basis of the obtained test results.

Description

Equipment tuning system using a digital twin and equipment tuning method using the same

[0001] The present invention relates to an equipment tuning system and an equipment tuning method using the same, and more specifically, to an equipment tuning system using a digital twin and an equipment tuning method using the same that can achieve automation by setting control parameters of equipment, such as a motor of roll-to-roll equipment, to optimal parameters and continuously and sequentially improving the performance of the equipment without process interruption.

[0002] Digital twin technology is a system that creates a virtual model of physical equipment to perform experiments and simulations on that equipment, and its application is expanding across various industrial fields due to recent developments in digital and artificial intelligence technologies.

[0003] For example, Japanese Published Patent No. 2024-500323 discloses a technology for estimating one or more variable values ​​of a hydrocarbon system in real time using digital twin technology. In addition, Japanese Registered Patent No. 7401934 discloses a technology for performing a dynamic evaluation of an industrial sieve effect by applying digital twin technology.

[0004] As described above, while technologies for performing simulations using digital twin technology are being developed in various fields, technology for automatically optimizing control parameters of equipment, such as motors—the most representative physical devices—has not yet been developed. In particular, in the case of roll-to-roll equipment where multiple motors are used simultaneously, the optimization of individual motor control parameters is performed by skilled operators relying on their experience and know-how. Consequently, this leads to reduced process efficiency, the impossibility of real-time optimization, and a decline in consistency across multiple process devices.

[0005] Accordingly, the technical problem of the present invention is conceived from this point, and the objective of the present invention is to provide an equipment tuning system using a digital twin that enables automation by setting control parameters of equipment, such as the motor of roll-to-roll equipment, to optimal parameters and continuously and sequentially improving equipment performance without process interruption, thereby improving process efficiency, enabling real-time optimization, and allowing consistent optimization for multiple process equipment.

[0006] In addition, another objective of the present invention is to provide a method for tuning equipment using the equipment tuning system.

[0007] An equipment tuning system according to one embodiment for realizing the purpose of the present invention described above includes a digital twin unit and an equipment control unit. The digital twin unit sets test conditions for optimizing control parameters of the equipment. Based on the set test conditions, the equipment control unit drives the equipment to obtain test results for each test condition. In this case, the digital twin unit resets the test conditions based on the obtained test results until optimal control parameters of the equipment are derived.

[0008] In one embodiment, the system may further include a server unit that receives the set test conditions from the digital twin unit and transmits them to the equipment control unit, and receives the acquired test results from the equipment control unit and transmits them to the digital twin unit.

[0009] In one embodiment, the server unit can perform security processing to prevent the test conditions and test results from being leaked to the outside.

[0010] In one embodiment, the equipment may be equipment driven by the motor, comprising at least one motor.

[0011] In one embodiment, the equipment is a roll-to-roll equipment driven by the plurality of motors, and the control parameter may be an indicator representing the operating characteristics of the motors as the tension changes to a target tension after the roll-to-roll equipment is driven with an initial tension.

[0012] In one embodiment, the control parameters may be the time constant (T), overshoot (O), and settling time (S) of each of the motors.

[0013] In one embodiment, the optimal control parameter is,

[0014] Equation (1)

[0015] (Here, α, β, and γ represent the respective weights of the control parameters, each being an arbitrary value between 0 and 1.)

[0016] The time constant, overshoot, and settling time may be the case where the cost function defined by the above equation (1) is minimized.

[0017] In one embodiment, the equipment control unit may include a driving unit that drives the equipment based on the set test conditions, a data acquisition unit that acquires test data according to the test conditions from the driving unit, and a preprocessing unit that preprocesses the acquired test data to acquire the test result.

[0018] In one embodiment, the system may further include a system evaluation unit that evaluates the operating status of the equipment based on the derived optimal parameters and operates a maintenance algorithm when a problem occurs in the equipment.

[0019] In one embodiment, the equipment may be located remotely separated from the digital twin.

[0020] In one embodiment, the digital twin unit can set the test conditions using the results of previous experiments.

[0021] In one embodiment, the digital twin unit can derive optimal control parameters of the equipment through self-supervised learning and transfer learning using prior experimental results.

[0022] In a method for tuning equipment according to another embodiment for realizing the purpose of the present invention described above, test conditions are set to optimize the control parameters of the equipment. Based on the set test conditions, the equipment is driven to obtain test results for each test condition. Based on the obtained test results, the test conditions are reset until the optimal control parameters of the equipment are derived.

[0023] In one embodiment, the method may further include the step of receiving the set test conditions from the digital twin unit and transmitting them to the equipment control unit, and the step of receiving the acquired test results from the equipment control unit and transmitting them to the digital twin unit.

[0024] In one embodiment, the method may further include the step of driving the equipment with the optimal control parameter when the derived optimal control parameter is normal, and the step of driving the equipment with a maintenance algorithm when the derived optimal control parameter is abnormal.

[0025] In one embodiment, the step of obtaining the test result may include the step of operating the equipment based on the set test conditions, the step of obtaining test data according to the test conditions, and the step of obtaining the test result by preprocessing the obtained test data.

[0026] In one embodiment, in the step of setting the test conditions, the P gain and I gain of the motor included in the equipment can be selected.

[0027] In one embodiment, in the step of setting the test conditions, the test conditions can be set by transfer learning using previous experimental results.

[0028] In one embodiment, when deriving the optimal control parameters of the equipment, learning can be performed by selecting control parameters in an even pattern from all control parameters, and then additionally learning can be performed on control parameters other than the candidate group, and then learning can be performed to derive the optimal control parameters.

[0029] According to embodiments of the present invention, the digital twin unit autonomously sets test conditions for optimizing the control parameters of the equipment, derives optimal control parameters based on the test results for each test condition, and controls the equipment to perform optimal control. Thus, by moving away from the conventional method of relying on manual work to derive optimal control parameters for specific equipment and implementing automation through the digital twin, process efficiency is improved, real-time optimization is possible, and consistent optimization for multiple process equipment can be achieved.

[0030] Since the derivation of these optimal control parameters can be performed by analyzing and updating process data in real time while the equipment is in operation, the performance of the equipment can be continuously improved without interrupting the process.

[0031] In addition, in deriving the optimal control parameters, transfer learning based on previously learned results can be applied, thereby minimizing the time and computation required to derive the optimal control parameters.

[0032] In particular, when multiple motors are used simultaneously, such as in roll-to-roll equipment, the derivation of optimal control parameters as described above can be automated for each of the multiple motors, thereby enabling rapid real-time automatic tuning of the motors while eliminating complex manual tuning procedures.

[0033] Furthermore, since the digital twin unit can derive the optimal control parameters remotely, the same optimal control parameters can be applied to multiple different factories located remotely, thereby enabling consistent control of the same equipment as well as remotely deriving optimal control parameters that take into account the field conditions of each factory.

[0034] In addition, regarding the transmission and reception of test conditions or test results between the digital twin unit and the equipment control unit, by performing transmission and reception through a security-enhanced server unit, security issues such as the leakage of data to the outside can be resolved, thereby enabling safe data transmission and reception.

[0035] At this time, the control parameters for optimal control in driving the motor are derived based on three variables: time constant, overshoot, and settling time. The digital twin unit can automatically and quickly derive the optimal control parameters by setting test conditions through feedback that considers the test results for the three variables.

[0036] Furthermore, the system evaluation unit can evaluate the operating status of the equipment based on the derived optimal control parameters to determine in real time whether a problem has occurred with the equipment, and if a problem occurs, it can quickly resolve the problem through a maintenance algorithm to continuously maintain the stable operation of the equipment.

[0037] FIG. 1 is a block diagram illustrating an equipment tuning system using a digital twin according to an embodiment of the present invention.

[0038] Figure 2 is a flowchart illustrating an equipment tuning method using the equipment tuning system of Figure 1.

[0039] Figure 3 is a schematic diagram illustrating the step of applying the previously learned data of Figure 2.

[0040] Figure 4 is a flowchart illustrating the steps of operating the equipment of Figure 2 to obtain test results.

[0041] Figure 5 is a graph illustrating control parameters when the equipment tuning system of Figure 1 performs motor tuning.

[0042] Figure 6 is a graph illustrating how the digital twin unit of Figure 1 sets test conditions.

[0043] FIGS. 7a to 7c are images illustrating an example of deriving optimal control parameters while varying the control parameters of a motor in the equipment tuning system of FIG. 1, and FIGS. 8a to 8c are graphs illustrating the control state of a motor when deriving the control parameters of FIGS. 7a to 7c.

[0044] FIGS. 9a to 9d are graphs illustrating the state in which a cost function is derived during the process of optimizing the control parameters of FIG. 5 in the equipment tuning system of FIG. 1.

[0045] <Explanation of Symbols>

[0046] 1 : Equipment Tuning System 100 : Digital Twin Unit

[0047] 110: Condition setting section 120: Result analysis section

[0048] 130 : Command section 140 : Database

[0049] 200: Server Unit 300: Equipment Control Unit

[0050] 310: Driving unit 320: Data acquisition unit

[0051] 330 : Preprocessing Unit 400 : Equipment

[0052] 410, 420, 430 : Motor 500 : System Evaluation Department

[0053] The present invention is susceptible to various modifications and may take various forms, and embodiments are to be described in detail in the text. However, this is not intended to limit the invention to the specific disclosed forms, and it should be understood that the invention includes all modifications, equivalents, and substitutions that fall within the spirit and scope of the invention. Similar reference numerals have been used for similar components in the description of each figure. Terms such as "first," "second," etc., may be used to describe various components, but said components should not be limited by said terms.

[0054] The above terms are used solely for the purpose of distinguishing one component from another. The terms used in this application are used merely to describe specific embodiments and are not intended to limit the invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this application, terms such as "comprising" or "consisting of" are intended to specify the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0055] Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the attached drawings.

[0056] FIG. 1 is a block diagram illustrating an equipment tuning system using a digital twin according to an embodiment of the present invention.

[0057] Referring to FIG. 1, the equipment tuning system using the digital twin according to the present embodiment (hereinafter referred to as the equipment tuning system) (1) includes a digital twin section (100), a server section (200), an equipment control section (300), and a system evaluation section (500).

[0058] At this time, a digital twin refers to a system capable of achieving optimization of the actual equipment while minimizing the number of tests on the actual equipment by creating a virtual model of the actual equipment (400) to be tuned and performing tests and simulations on the equipment (400). Accordingly, the equipment tuning system (1) in this embodiment also applies such digital twin technology, thereby achieving optimization of the equipment (400) through the digital twin as described above, and by directly applying the optimized control parameters thus achieved to the actual equipment, the performance of the actual equipment is improved without interrupting the operation of the actual equipment.

[0059] Accordingly, the digital twin unit (100) sets test conditions for optimizing control parameters of the equipment (400) and includes a condition setting unit (110), a result analysis unit (120), a command unit (130), and a database (140).

[0060] At this time, the equipment (400) is a target for deriving optimal control parameters using the equipment tuning system (1) according to the present embodiment, and may be located physically separated from the digital twin unit (100). That is, the location where the equipment (400) is located is not limited, and thus remote control of the equipment (400) is possible through the digital twin unit (100).

[0061] The above equipment (400) may be, for example, a roll-to-roll equipment. The above roll-to-roll equipment is a piece of equipment capable of rapidly manufacturing various products in large quantities through a continuous process and is utilized in various industrial fields.

[0062] In particular, such roll-to-roll equipment includes multiple motors that control the operation of substrates to continuously supply and retrieve substrates or electrodes, and each motor is controlled through a separate control means. At this time, since the operation of each of the motors directly affects performance and product quality in manufacturing using the roll-to-roll equipment, optimally controlling the operation of the motors is very important.

[0063] Accordingly, the equipment tuning system (1) performs optimization for each of the control parameters of at least one motor (410, 420, ..., 430) included in the equipment (400), such as P gain, I gain, etc.

[0064] However, the equipment (400) subject to such optimization tuning is not limited to roll-to-roll equipment and may include various equipment, but may be equipment including at least one motor for which optimization of control parameters is required. Furthermore, it is obvious that this can be extended and applied to various mechanical parts requiring optimization of control parameters in addition to motors.

[0065] The equipment control unit (300) drives the equipment (400) based on the test conditions set through the digital twin unit (100) to obtain test results for each test condition. That is, the equipment control unit (300) may be located physically separated from the digital twin unit (100) and receives the remotely set test conditions to drive the equipment (400) directly. Accordingly, the equipment control unit (300) includes a driving unit (310), a data acquisition unit (320), and a preprocessing unit (330) to directly control the driving of the equipment (400) and obtain test results.

[0066] The server unit (200) performs data transmission and reception between the digital twin unit (100) and the equipment control unit (300) and maintains the security of the transmitted and received data. That is, the test conditions generated through the digital twin unit (100) are primarily transmitted to the server unit (200), and the server unit (200) strengthens the security of the test conditions and transmits the test conditions to the equipment control unit (300).

[0067] At this time, the server unit (200) may operate a security enhancement program, such as a firewall to block hacking or data intrusion from the outside, or transmit the test conditions in an encrypted manner in order to strengthen security.

[0068] In addition, the server unit (200) receives the test results obtained from the equipment control unit (300) regarding the corresponding test conditions and transmits the test results to the digital twin unit (100). At this time, the transmitted test results may also be transmitted through a security enhancement program or transmitted in an encrypted form. Thus, the security of all data generated through the equipment tuning system (1) can be enhanced through the server unit (200). That is, the server unit (200) must transmit data such as the test conditions and test results between the digital twin unit (100) and the equipment control unit (300), which may be physically separated from each other, while performing security.

[0069] Furthermore, the server unit (200) not only strengthens security for data such as the test conditions or test results, but can also apply a security system to prevent hacking or data intrusion from the outside, as previously exemplified, when performing operations of the digital twin unit (100), the equipment control unit (300), and the system evaluation unit (500).

[0070] Of course, the server unit (200) may be omitted, and the test conditions may be directly processed and transmitted from the digital twin unit (100) to the equipment control unit (300), and likewise, the test results may also be directly processed and transmitted from the equipment control unit (300) to the digital twin unit (100). For such direct transmission and reception of data, it is obvious that a predetermined remote network for transmitting and receiving remote data must be provided in the digital twin unit (100) and the equipment control unit (300), even if a separate server unit is omitted.

[0071] The system evaluation unit (500) determines whether the parameters derived as optimal control parameters for the equipment (400) from the digital twin unit (100) fall outside the range of a predetermined reference value or expected value. Thus, if the derived parameters do not fall outside the range of the reference value or expected value, the equipment (400) is in a normal state, and operation of the equipment (400) is performed based on the derived parameters.

[0072] However, if the parameters derived above fall outside the range of the reference values ​​or expected values, the equipment (400) is determined to be in an abnormal state, and it is determined that the equipment (400) is in a state where unexpected problems have occurred, such as parts wearing out or sensors malfunctioning. Accordingly, when the equipment (400) is in an abnormal state, the system evaluation unit (500) applies a so-called maintenance algorithm to restore the equipment (400) to a normal state. At this time, the meaning of restoring the equipment (400) to a normal state, that is, applying the maintenance algorithm, refers to a series of operations or algorithms to restore the equipment (400) to a normal state, such as replacing the problematic parts or sensors in the equipment (400) or rebooting the equipment (400).

[0073] In the case of the above maintenance algorithm, it may include analyzing the cause of the equipment (400) being determined to be in an abnormal state and performing appropriate measures, and the above maintenance algorithm may be performed in various ways depending on the type of equipment (400) or the type of abnormal state.

[0074] As described above, through the equipment tuning system (1), optimal control parameters for each of the motors (410, 420, ..., 430) constituting the equipment (400) are derived using a so-called digital twin, and each of the motors is controlled based on this to perform optimized operation of the equipment (400). The derivation of these optimal control parameters is ultimately performed using artificial intelligence learning, and, for example, can be performed through self-supervised learning based on the test conditions and the test results in the digital twin unit (100).

[0075] Meanwhile, a more detailed explanation of the above-mentioned equipment tuning system (1) is explained simultaneously with the following equipment tuning method using the above-mentioned equipment tuning system (1) for convenience of explanation.

[0076] FIG. 2 is a flowchart illustrating an equipment tuning method using the equipment tuning system of FIG. 1. FIG. 3 is a schematic diagram explaining the step of applying the prior learning data of FIG. 2. FIG. 4 is a flowchart illustrating the step of driving the equipment of FIG. 2 to obtain test results. FIG. 5 is a graph explaining the control parameters when the equipment tuning system of FIG. 1 performs motor tuning. FIG. 6 is a graph explaining the method of setting test conditions for the digital twin unit of FIG. 1. FIG. 7a to 7c are images illustrating examples of deriving optimal control parameters while varying the control parameters of the motor in the equipment tuning system of FIG. 1, and FIG. 8a to 8c are graphs illustrating the control state of the motor when deriving the control parameters of FIG. 7a to 7c. FIG. 9a to 9d are graphs illustrating the state in which a cost function is derived during the process of optimizing the control parameters of FIG. 5 in the equipment tuning system of FIG. 1.

[0077] Referring to FIGS. 1 and 2, in the above equipment tuning method, first, in the condition setting unit (110) of the digital twin unit (100), a test condition to perform a test on the equipment (400) is set (step S10).

[0078] At this time, the above test conditions are conditions for performing a test to derive optimal control parameters for the equipment (400), and since this corresponds to the first test, the initial test conditions are arbitrarily set.

[0079] Meanwhile, as previously explained, the equipment (400) may be a roll-to-roll process equipment comprising a plurality of motors (410, 420, ..., 430), and such roll-to-roll process equipment is described below as an example.

[0080] In setting the initial test conditions above, since it is difficult to know the optimal test conditions from the beginning, an approximate range is set as the test conditions based on existing test results or the characteristics of the equipment (400). For example, when optimizing the control parameters of each of the motors, the range of the P gain of each of the motors can be set between 1 and 10, and the range of the I gain can be set between 0.01 and 1.

[0081] The test conditions are divided equally within the above-mentioned range, for example, if divided equally into 4*5, a total of 20 basic test conditions can be set.

[0082] Meanwhile, in setting the initial test conditions, data of previously tested results stored in the database (140) can be utilized (step S15). Referring simultaneously to FIG. 3, the digital twin unit (100) may have already received test conditions set for a specific device (400) and test results optimized according to those test conditions, and these previously tested results are stored in the database (140).

[0083] Accordingly, for example, when data regarding test conditions and optimal test results is stored in the equipment (400) when a process is performed on a PET film substrate as a substrate, the digital twin unit (100) can set the optimal test conditions for performing a process on a PI film substrate as a new substrate in the equipment (400) through so-called transfer learning. Thus, when performing learning to obtain test results according to the conditions by newly setting initial test conditions, hundreds of data points may need to be learned, but by performing transfer learning based on the previously stored data, optimized control parameters can be derived by learning about 30 to 50 data points.

[0084] As described above, the initial test conditions are set based on previously stored data, or if there is no previously stored data, the conditions are roughly set and transmitted to the equipment control unit (300) with enhanced security through the server unit (200) (step S20). Of course, as previously explained, the server unit (200) may be omitted, and the test conditions may be transmitted directly from the digital twin unit (100) to the equipment control unit (300) with security applied.

[0085] Afterwards, referring to FIGS. 1 and FIGS. 2, the equipment control unit (300) performs operation of the equipment (400) based on the transmitted test conditions and obtains test results for each test condition therefrom (step S30).

[0086] More specifically, referring to FIG. 4, in the driving and test result acquisition step (step S30), first, the driving unit (310) of the equipment control unit (300) drives the motors (410, 420, ..., 430) constituting the equipment (400) based on the test conditions transmitted through the server unit (200) or the digital twin unit (100) (step S31).

[0087] In this case, since the above test conditions may include different test conditions for each motor, the driving unit (310) is driven individually for each motor (410, 420, ..., 430) under each test condition.

[0088] For example, to drive the motor, if the equipment (400) is a roll-to-roll process equipment, the drive unit (310) drives the substrate or electrode provided by the roll-to-roll process equipment with an initial tension value, and then drives it by changing the tension to a target tension. At this time, the initial tension value or the target tension may be a value included in the pre-set test conditions.

[0089] Afterward, the data acquisition unit (320) of the equipment control unit (300) acquires the operating characteristics of the system, i.e., the motors, of the equipment (400) through time-series data according to the change in tension during the process in which the equipment (400) is driven through the driving unit (310) (step S32). At this time, the acquired motor operating characteristic data may include, for example, a time constant (T), an overshoot (O), a settling time (S), etc.

[0090] That is, referring to FIG. 5, when a predetermined tension value is applied to the substrate or electrode of the equipment (400) through the driving unit (310) and the motor is driven, the motor is driven with characteristics such as a time constant (T), overshoot (O), and settling time (S). That is, the important dynamic performance indicators of the motor are the time constant (T), overshoot (O), and settling time (S), and the operation of the motor can be controlled through these indicators.

[0091] Of course, the data acquisition unit (320) can acquire various data regarding related driving characteristics in addition to the time constant (T), overshoot (O), and settling time (S) data exemplified above as data regarding the driving characteristics of the equipment (400). To this end, the data acquisition unit (320) may include various sensors.

[0092] Afterwards, the preprocessing unit (330) of the equipment control unit (300) performs preprocessing on the data regarding the driving characteristics of the equipment (400) obtained through the data acquisition unit (320) (step S33), and derives only the key indicators exemplified above, namely the time constant (T), overshoot (O), and settling time (S).

[0093] That is, among the data related to driving characteristics obtained from the data acquisition unit (320), only the time constant (T), overshoot (O), and settling time (S) are derived, and other data are removed, thereby selecting only the most important indicators among the indicators representing the driving characteristics of each motor included in the equipment (400), thereby improving the efficiency of the analysis of test results.

[0094] At this time, the data processing method performed by the preprocessing unit (330) is not limited to a specific method.

[0095] As described above, the equipment control unit (300) performs a test on the equipment (400) based on the test conditions generated through the digital twin unit (100), and derives the result data of the test as the test result.

[0096] Afterward, the test results derived above are transmitted to the digital twin unit (100) with enhanced security through the server unit (200) (step S40). Of course, as previously explained, the server unit (200) may be omitted, and the test results derived above may be transmitted directly to the digital twin unit (100).

[0097] Afterward, the digital twin unit (100) analyzes the transmitted test results (step S50). That is, the result analysis unit (120) analyzes the transmitted test results to derive control parameters.

[0098] Specifically, the previously transmitted test results are three indicators: time constant (T), overshoot (O), and settling time (S). Accordingly, the result analysis unit (120) defines a so-called cost function that comprehensively considers the three performance indicators as shown in the following equation (1) and derives it.

[0099] Equation (1)

[0100] (Here, α, β, and γ represent the respective weights of the control parameters, each being an arbitrary value between 0 and 1.)

[0101] At this time, in the above equation (1), if the weights of the three performance indicators are equal to each other, α, β, and γ can all be defined as 0.3, and these α, β, and γ can be arbitrarily set considering the importance of the performance indicators.

[0102] Thus, the result analysis unit (120) derives the case where the value of the cost function defined by the above equation (1) becomes the lowest, and derives the P gain and I gain under the corresponding test conditions. That is, when the value of the above cost function is the lowest, accurate control of the motor is possible, and this corresponds to the P gain and I gain that allow for optimal control while minimizing disturbances.

[0103] Ultimately, the case where the value of the cost function calculated by the result analysis unit (120) becomes the lowest can be said to be the case where the optimal control parameter for the equipment (400) is derived (step S60).

[0104] Accordingly, until the value of the cost function calculated by the result analysis unit (120) becomes the lowest (step S60), the condition setting unit (110) must vary the previously set test conditions (step S10) and obtain test results for each test condition.

[0105] As described above, the digital twin unit (100) obtains test results for each variable test condition while varying the test conditions, and determines whether the value of the cost function derived from the test results is at its lowest. Thus, if an optimal control parameter for the equipment (400) is derived, the operation of the equipment (400) is performed under the corresponding condition. Such variation of test conditions in the digital twin unit (100) and acquisition of test results are performed through self-supervised learning and transfer learning, as previously explained.

[0106] Additionally, referring to FIG. 6, when performing learning to derive optimal control parameters in the digital twin unit (100), it can be performed in multiple stages. For example, in the initial section (A), learning is performed by selecting control parameters that can be selected by methods such as so-called Grid search and OPTUNA in an overall even and sparse pattern.

[0107] Afterwards, when specific control parameters are considered to be candidates for optimal parameters, in the intermediate period (B), learning is performed on a range outside the candidate group using methods such as Bayesian EI to solve the so-called local minimum problem that occurs during the process of adjusting the control parameters.

[0108] Thus, after learning is performed using the above-mentioned Baysain EI method, learning is performed in the final section (C) to derive optimal control parameters using the Baysian PI method. Through this, the above-mentioned digital twin section (100) can derive optimal control parameters more accurately and quickly.

[0109] Meanwhile, the execution of the algorithm in the aforementioned result analysis unit (120) is exemplified through FIGS. 7a to 8c. That is, as shown in FIGS. 7a to 7c, as the number of times the test conditions are newly set (trial) is increased and the value (score) of the cost function is derived, it can be confirmed that the derived cost function is derived as a minimum value.

[0110] In addition, as shown in FIGS. 9a to 9d, information regarding the state in which the value of the cost function varies as each of the important indicators constituting the cost function, such as the time constant (T), overshoot (O), and settling time (S), changes can be obtained in real time.

[0111] Furthermore, in performing this algorithm, the result analysis unit (120) can perform modeling by deriving a three-dimensional map by assigning the X-axis and Y-axis to P-gain and I-gain, respectively, and assigning the Z-axis to the value (score) of the cost function, as shown in FIGS. 8a to 8c. In addition, during the process of performing this modeling, data obtained from testing the equipment (400) is treated as a true value, and the change of the cost function can be predicted by applying a probabilistic standard deviation to other areas.

[0112] Thus, the condition setting unit (110) can set the next test condition by selecting a new P gain and I gain that are expected to minimize the value of the cost function based on the predicted change of the cost function (step S10).

[0113] Accordingly, the setting of these new test conditions and the analysis of results based on the test results are performed repeatedly until optimal control parameters for the equipment (400) are derived (step S60), thereby finally determining the optimal control parameters for the equipment (400).

[0114] Meanwhile, in performing such tests, as previously explained, the so-called local minimum is a repetitive test performed only on a specific area to derive a minimum value in that area, but to prevent a situation where the minimum value of the cost function exists in a practically different area, the result analysis unit (120) may apply a supplementary algorithm, such as appropriately adjusting the probability of the standard deviation, when performing the algorithm.

[0115] In the case of this embodiment, when optimal control parameters for each motor (410, 420, ..., 430) included in the equipment (400) are derived through the algorithm described above (step S60), the system evaluation unit (500) further determines whether there is an abnormality in the derived optimal control parameters (step S70).

[0116] That is, as explained above, the system evaluation unit (500) determines whether the parameters derived as optimal control parameters deviate from a predetermined standard value or expected value range.

[0117] Thus, if the derived parameters do not exceed the range of the reference value or expected value, the equipment (400) is in a normal state, and the operation of the equipment (400) is performed based on the derived parameters (step S80).

[0118] However, if the parameters derived above fall outside the range of the reference value or expected value (step S70), the equipment (400) is determined to be in an abnormal state, and the equipment (400) is determined to be in a state where an unexpected problem has occurred, such as parts wearing out or sensors malfunctioning.

[0119] Accordingly, when the equipment (400) is in an abnormal state, the system evaluation unit (500) applies a so-called maintenance algorithm to restore the equipment (400) to a normal state (step S90). At this time, the meaning of restoring the equipment (400) to a normal state, that is, applying the maintenance algorithm, refers to a series of operations or algorithms to restore the equipment (400) to a normal state, such as replacing a problematic part or sensor in the equipment (400) or rebooting the equipment (400).

[0120] In the case of the above maintenance algorithm, it may include analyzing the cause of the equipment (400) being determined to be in an abnormal state and performing appropriate measures, and the above maintenance algorithm may be performed in various ways depending on the type of equipment (400) or the type of abnormal state.

[0121] The execution of the above-mentioned maintenance algorithm may be repeated until the equipment (400) is in a normal state, and when it is finally determined that the equipment (400) is in a normal state (step S70), the operation of the equipment (400) is performed based on the derived parameters (step S80).

[0122] That is, when the optimal control parameter derived for each motor (410, 420, ..., 430) included in the equipment (400) is within the normal range, the command unit (130) selects the P gain and I gain in the corresponding control parameter and commands the equipment (400) to be driven (step S80). This driving command for the equipment (400) is a driving command for the actual equipment and can be provided to the actual equipment as a control signal for the equipment.

[0123] According to the embodiments of the present invention described above, the digital twin unit automatically sets test conditions for optimizing the control parameters of the equipment, derives optimal control parameters based on the test results for each test condition, and controls the equipment to perform optimal control. Thus, by moving away from the conventional method of relying on manual work to derive optimal control parameters for specific equipment and implementing automation through the digital twin, process efficiency is improved, real-time optimization is possible, and consistent optimization for multiple process equipment can be achieved.

[0124] Since the derivation of these optimal control parameters can be performed by analyzing and updating process data in real time while the equipment is in operation, the performance of the equipment can be continuously improved without interrupting the process.

[0125] In addition, in deriving the optimal control parameters, transfer learning based on previously learned results can be applied, thereby minimizing the time and computation required to derive the optimal control parameters.

[0126] In particular, when multiple motors are used simultaneously, such as in roll-to-roll equipment, the derivation of optimal control parameters as described above can be automated for each of the multiple motors, thereby enabling rapid real-time automatic tuning of the motors while eliminating complex manual tuning procedures.

[0127] Furthermore, since the digital twin unit can derive the optimal control parameters remotely, the same optimal control parameters can be applied to multiple different factories located remotely, thereby enabling consistent control of the same equipment as well as remotely deriving optimal control parameters that take into account the field conditions of each factory.

[0128] In addition, regarding the transmission and reception of test conditions or test results between the digital twin unit and the equipment control unit, by performing transmission and reception through a security-enhanced server unit, security issues such as the leakage of data to the outside can be resolved, thereby enabling safe data transmission and reception.

[0129] At this time, the control parameters for optimal control in driving the motor are derived based on three variables: time constant, overshoot, and settling time. The digital twin unit can automatically and quickly derive the optimal control parameters by setting test conditions through feedback that considers the test results for the three variables.

[0130] Furthermore, the system evaluation unit can evaluate the operating status of the equipment based on the derived optimal control parameters to determine in real time whether a problem has occurred with the equipment, and if a problem occurs, it can quickly resolve the problem through a maintenance algorithm to continuously maintain the stable operation of the equipment.

[0131] Although the present invention has been described above with reference to preferred embodiments, those skilled in the art will understand that various modifications and changes can be made to the invention without departing from the spirit and scope of the invention as set forth in the following claims.

Claims

1. A digital twin unit for setting test conditions to optimize the control parameters of the equipment; and Based on the above-mentioned test conditions, it includes an equipment control unit that drives the equipment to obtain test results for each test condition, and The above digital twin unit is characterized by resetting the test conditions until the optimal control parameters of the equipment are derived based on the acquired test results, in an equipment tuning system.

2. In Paragraph 1, An equipment tuning system further comprising a server unit that receives the above-set test conditions from the digital twin unit and transmits them to the equipment control unit, and receives the above-mentioned test results from the equipment control unit and transmits them to the digital twin unit.

3. In paragraph 2, the above server unit, An equipment tuning system characterized by security processing to prevent the above test conditions and above test results from being leaked to the outside.

4. In paragraph 1, the above equipment is, An equipment tuning system characterized by comprising at least one motor and being equipment driven by said motor.

5. In Paragraph 4, The above equipment is a roll-to-roll equipment driven by the above plurality of motors, and An equipment tuning system characterized in that the above control parameter is an indicator representing the operating characteristics of the motors as the tension changes to a target tension after the roll-to-roll equipment is driven with an initial tension.

6. In paragraph 5, the above control parameter is, An equipment tuning system characterized by the time constant (T), overshoot (O), and settling time (S) of each of the above motors.

7. In paragraph 6, the optimal control parameter is, Equation (1) (Here, α, β, and γ represent the respective weights of the control parameters, each being an arbitrary value between 0 and 1.) A device tuning system characterized by the time constant, overshoot, and settling time when the cost function defined by the above equation (1) is at its minimum.

8. In paragraph 1, the equipment control unit is, A driving unit that drives the equipment based on the above-mentioned set test conditions; A data acquisition unit that acquires test data according to the test conditions from the above driving unit; and An equipment tuning system characterized by including a preprocessing unit that preprocesses the acquired test data to obtain the test result.

9. In Paragraph 1, An equipment tuning system further comprising a system evaluation unit that evaluates the operating status of the equipment based on the optimal parameters derived above and operates a maintenance algorithm when a problem occurs in the equipment.

10. In paragraph 1, the above equipment is, An equipment tuning system characterized by being located remotely separated from the above-mentioned digital twin.

11. In paragraph 1, the digital twin unit is, An equipment tuning system characterized by setting the above test conditions using the results of a previous experiment.

12. In Clause 11, the digital twin unit is, An equipment tuning system characterized by deriving optimal control parameters of the equipment through self-supervised learning and transfer learning using prior experimental results.

13. A step of setting test conditions to optimize the control parameters of the equipment; A step of operating the equipment based on the above-described test conditions to obtain test results for each test condition; and A method for tuning equipment that includes the step of resetting test conditions until optimal control parameters of the equipment are derived based on the above-mentioned test results.

14. In Paragraph 13, A step of receiving the above-mentioned set test conditions from the digital twin unit and transmitting them to the equipment control unit; and A method for tuning equipment that further includes the step of receiving the above-mentioned test results from the equipment control unit and transmitting them to the digital twin unit.

15. In Paragraph 13, A step in which, when the optimal control parameter derived above is normal, the equipment is driven with the optimal control parameter; and A method for tuning equipment that further includes the step of operating the equipment with a maintenance algorithm when the optimal control parameter derived above is abnormal.

16. In Paragraph 13, the step of obtaining the above test results is, A step of operating the equipment based on the above-determined test conditions; A step of obtaining test data according to the above test conditions; and A method for tuning equipment characterized by including a step of obtaining test results by preprocessing the test data obtained above.

17. In Paragraph 13, at the step of setting the above test conditions, A method for tuning equipment characterized by selecting the P gain and I gain of the motor included in the above equipment.

18. In Paragraph 13, at the step of setting the above test conditions, A method for tuning equipment characterized by setting the above test conditions through transfer learning using the results of previous experiments.

19. In claim 11, when deriving the optimal control parameters of the above equipment, A method for tuning equipment characterized by selecting control parameters in a uniform pattern from all control parameters to perform learning, performing additional learning on control parameters other than the candidate group, and then performing learning to derive optimal control parameters.

Images