An intelligent manufacturing production line optimization control method based on digital twinning

Abstract

This invention discloses an optimized control method for intelligent manufacturing production lines based on digital twins, belonging to the field of intelligent manufacturing control technology. The method includes: acquiring production site operation status data; constructing an initial control topology composed of control nodes and signal connection edges as the current execution basis; based on state deduction information, identifying physical conservation residuals exceeding the safety envelope boundary value as abnormal deviation data, and identifying risk nodes and connection edges to obtain reconstruction marker information; and dynamically reconstructing the initial control topology into the control topology based on execution by deleting signal connection edges and isolating control nodes. This invention forms a reliable safety envelope by comparing the physical conservation residuals of the deduced state and the measured state, providing a basis for dynamic topology reconstruction, improving the control adaptability of switching between different board types in subsequent PCB processing, reducing the risk of decreased board separation accuracy, increased burrs, and missorting, and improving yield and production line stability.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing control technology, and in particular to an optimized control method for intelligent manufacturing production lines based on digital twins. Background Technology

[0002] Digital twins have been widely applied in the fields of condition monitoring and predictive maintenance of intelligent manufacturing production lines. Through the virtual mapping of physical entities, digital twins can realize real-time simulation and optimization decision-making of the production process, support the pre-verification of control commands and closed-loop management of execution feedback, and provide a technological foundation for the intelligent upgrading of manufacturing.

[0003] For example, the subsequent processing of PCB boards typically involves the collaborative control of multiple processes, including board loading, transport, visual positioning, routing and separation, appearance inspection, and unloading and sorting. Because different batches of PCB boards vary in thickness, shape, panelization method, and processing path, visual positioning deviations, abnormal spindle vibrations, tool wear, and transport cycle fluctuations can easily lead to decreased separation accuracy, increased burrs on board edges, station waiting times, and missorting. Existing general-purpose digital twin control methods struggle to reliably verify and dynamically adjust for real-time process constraints in subsequent PCB board processing scenarios. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an intelligent manufacturing production line optimization control method based on digital twins to solve the problems of unreliable verification of control command credibility and decreased production line operation stability during subsequent PCB board processing.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a method for optimizing and controlling an intelligent manufacturing production line based on digital twins. The method includes: acquiring production site operation status data; constructing an initial control topology composed of control nodes and signal connection edges as the current execution basis; based on the current execution basis, parsing the control node status and performing signal encapsulation processing to generate control instructions marked with pending transaction states, forming a virtual verification object; using digital twins to virtually execute the virtual verification object, calculating the physical conservation residual between the current projected state and the current measured state, obtaining state projection information and a safety envelope; and based on the state projection information, applying physical conservation values ​​exceeding the safety envelope boundary values. The residual is identified as abnormal deviation data, and risk nodes and connecting edges are identified to obtain reconstruction marker information. By deleting signal connecting edges and isolating control nodes, the initial control topology is dynamically reconstructed into the control topology based on execution. When the physical conservation residual is within the safety envelope and has no control command conflict with the isolated control node, the pending transaction status is atomically changed to the committed transaction status. Based on the control topology based on execution, it is sent to the CNC drive for execution to obtain control feedback information. The incremental factor elimination method is used to perform iterative processing on the control feedback information to eliminate error factors, correct the configuration parameters of control node signal reliability and control priority, and form the initial control topology for the next cycle.

[0007] As a preferred embodiment of the intelligent manufacturing production line optimization control method based on digital twins described in this invention, the production site operation status data includes spindle vibration voltage signal, positioning shaft temperature signal, and actual position deviation value.

[0008] As a preferred embodiment of the intelligent manufacturing production line optimization control method based on digital twins described in this invention, the control node refers to a logical object representing an independent control entity in the initial control topology, corresponding to the CNC control node, execution control node, and detection control node in the production site.

[0009] As a preferred embodiment of the intelligent manufacturing production line optimization control method based on digital twins described in this invention, the current execution basis is an initial control topology structure composed of control nodes and signal connection edges, constructed based on the production site operation status data.

[0010] As a preferred embodiment of the intelligent manufacturing production line optimization control method based on digital twins described in this invention, the specific steps for constructing the initial control topology and using it as the basis for current execution are as follows: By collecting and analyzing the operational status data of the production site in real time, the control node identifiers and signal connection relationships are obtained. The control node identifiers and signal connection relationships are reorganized and structurally mapped to construct an initial control topology consisting of control nodes and signal connection edges, which serves as the basis for current execution.

[0011] As a preferred embodiment of the intelligent manufacturing production line optimization control method based on digital twins described in this invention, the specific steps for forming the virtual verification object are as follows: Based on the current execution criteria, the execution status extraction process parses the running data of the control node to obtain the real-time status information of the node. The node's real-time status information is marked for transactions and reorganized with instructions to obtain control instructions marked with transactions to be committed. By marking control instructions that indicate the status of transactions to be committed, encapsulating the control instructions, and performing data aggregation and structuring, a virtual verification object is obtained.

[0012] As a preferred embodiment of the intelligent manufacturing production line optimization control method based on digital twins described in this invention, the specific steps for obtaining state deduction information and safety envelope are as follows: A digital twin is used to virtually execute a virtual verification object to obtain a predicted state data sequence. Based on the predicted state data sequence, the measured state data of the current control cycle is collected in real time. The difference between the predicted state data sequence corresponding to the current moment and the measured state data is calculated to obtain the physical conservation residual set. Define the safe range of the physical conservation residual set to obtain state deduction information and safety envelope.

[0013] As a preferred embodiment of the intelligent manufacturing production line optimization control method based on digital twins described in this invention, the specific steps of reconstructing the control topology as the execution basis are as follows: Based on state extrapolation information, the difference between the physical conservation residual and the boundary value of the safety envelope is determined, and the physical conservation residual exceeding the boundary value of the safety envelope is recorded as abnormal information to obtain abnormal deviation data. Based on the abnormal deviation data, control nodes that correspond to the abnormal deviation data and have continuous difference judgment information are identified as risk nodes, and signal connection edges that have signal connection relationships with risk nodes and correspond to abnormal deviation data are identified as risk connection edges. Risk nodes are marked as isolation objects and risk connection edges are marked as deletion objects to obtain reconstruction marking information. Based on the reconstructed labeling information, the initial control topology is deleting signal connection edges and isolating control nodes to form the control topology for execution.

[0014] As a preferred embodiment of the intelligent manufacturing production line optimization control method based on digital twins described in this invention, the specific steps for obtaining control feedback information are as follows: Based on the safety envelope, interval matching is performed on the physical conservation residuals within the current control cycle, and conflict detection is performed on the control instructions corresponding to the isolated control nodes to obtain transaction commit determination information. When the physical conservation residual characterization of the transaction commit determination information is within the safe envelope and there is no control instruction conflict, atomic state switching processing is performed on the control instructions marked as pending transaction state to change the pending transaction state to committed transaction state and obtain the set of committed control instructions. Based on the submitted set of control instructions, drive mapping is performed on the control topology of the execution basis to generate drive execution signals that match the corresponding control nodes and send them to the corresponding drive execution. During the execution of CNC drive, drive response data is collected in real time, and the collected operation response data is extracted and processed to obtain control feedback information.

[0015] As a preferred embodiment of the intelligent manufacturing production line optimization control method based on digital twins described in this invention, the specific steps for forming the initial control topology for the next cycle are as follows: Based on control feedback information, feedback feature data corresponding to the operating status, command execution status and drive response status of the control node are extracted. Incremental factor elimination method is used to iteratively remove error factors and filter noise in the feedback feature data to obtain feature data after eliminating interference. Based on the feature data after interference elimination, the configuration parameters of control node signal reliability and control priority are numerically corrected and updated to obtain the updated set of configuration parameters. Based on the updated set of configuration parameters, the signal connection relationships between control nodes are structurally mapped and reorganized to form the initial control topology for the next control cycle.

[0016] The beneficial effects of this invention are as follows: by comparing the physical conservation residuals of the deduced state and the measured state, a reliable and safe envelope is formed, which provides a basis for dynamic topology reconstruction, improves the control and adaptation capability of switching between different board types in subsequent PCB processing, reduces the risk of decreased board separation accuracy, increased burrs and missorting, and improves yield and production line stability. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating an optimized control method for a digital twin-based intelligent manufacturing production line.

[0019] Figure 2 This is a data flow diagram for a smart manufacturing production line based on digital twins.

[0020] Figure 3 This is a virtual execution flowchart for a smart manufacturing production line based on digital twins.

[0021] Figure 4 This is a flowchart of the topology reconstruction of a smart manufacturing production line based on digital twins. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a method for optimizing and controlling an intelligent manufacturing production line based on digital twins, comprising the following steps: S1: Obtain production site operation status data and construct an initial control topology consisting of control nodes and signal connection edges, which serves as the basis for current execution.

[0026] S1.1: Production site operation status data includes spindle vibration voltage signal, positioning axis temperature signal and actual position deviation value, including operating parameters such as spindle speed and feed rate, environmental monitoring values ​​such as machining area temperature, and signal interaction records such as command transmission and response timestamps between control nodes.

[0027] A control node is a logical object representing an independent control entity in the initial control topology, corresponding to the CNC control node, execution control node, and detection control node in the production site.

[0028] The current execution is based on the initial control topology structure, which is constructed from the production site operation status data and consists of control nodes and signal connection edges.

[0029] By collecting and analyzing the operational status data of the production site in real time, the control node identifiers and signal connection relationships are obtained.

[0030] Furthermore, the acquisition and communication interface analyzes the operating parameters, environmental monitoring values, and signal interaction records in the production site operation status data, identifies the independent control entities corresponding to the real-time working status of the physical entities in the production site, extracts the control node identifiers according to the correspondence between CNC control nodes, execution control nodes, and detection control nodes, organizes the signal interaction records, extracts the signal connection relationships between control nodes, and collects the control node identifiers and signal connection relationships to obtain a set of control node identifiers and signal connection relationships.

[0031] Specifically, the control node identifier is the information content that represents the identity attributes of each control node in the initial control topology. It comes from the independent control entity identified after the status analysis of the operating parameters and environmental monitoring values ​​in the production site operation status data.

[0032] The control node identifiers correspond to the PCB router spindle drive, conveyor belt execution drive, and AOI detection drive in the production site, respectively, distinguishing the categories, objects of action, and positional relationships of different control nodes in the control topology.

[0033] The signal connection relationship set is a set of information that characterizes the signal interaction and association between control nodes. It is the connection relationship content obtained after analyzing and organizing the signal interaction records in the production site operation status data.

[0034] The signal connection relationship set represents the signal transmission direction, connection correspondence, and interaction path between control nodes, as well as the structural relationship formed between control nodes through signal connection edges.

[0035] The correspondence between CNC control nodes, execution control nodes, and detection control nodes is based on the functional categories and signal interaction relationships of physical entities in the production site during the control process. The identified independent control entities are respectively mapped to the control node categories corresponding to CNC control nodes, execution control nodes, and detection control nodes.

[0036] Signal connection relationship refers to the signal transmission correspondence formed between control nodes based on signal interaction records, which represents the signal transmission direction, connection correspondence and interaction path between each control node.

[0037] S1.2: Reorganize and structurally map the set of control node identifiers and signal connection relationships to construct an initial control topology composed of control nodes and signal connection edges, which serves as the basis for current execution.

[0038] Furthermore, the control node identifiers and signal connection relationship sets are reorganized and structurally mapped. The control node identifiers that can represent independent control entities such as CNC control nodes, execution control nodes, and detection control nodes are mapped as control nodes. The signal connection relationship sets that exist between control nodes are mapped as signal connection edges. The corresponding connection relationship between control nodes and signal connection edges is determined based on the signal connection relationship set. An initial control topology composed of control nodes and signal connection edges is constructed as the basis for current execution.

[0039] Specifically, the reorganization and structural mapping process involves collecting and organizing the control node identifiers and signal connection relationship sets. This involves mapping the control node identifiers, representing the independent control entities of CNC control nodes, execution control nodes, and detection control nodes, to the control nodes themselves. The signal connection relationship set is then sorted out, mapping the signal transmission directions and connection correspondences between control nodes to signal connection edges. Based on the connection correspondences between control nodes in the signal connection relationship set, the control nodes and signal connection edges are connected, arranged, and structurally reorganized. The signal connection edges are then connected to the corresponding control nodes according to the signal transmission relationships, completing the initial control topology construction composed of control nodes and signal connection edges.

[0040] S2: Based on the current execution criteria, parse the control node status and perform signal encapsulation processing to generate control instructions marked with the status of transactions to be submitted, thus forming a virtual verification object.

[0041] S2.1: Based on the current execution criteria, the execution status extraction performs runtime data parsing on the control node to obtain the node's real-time status information.

[0042] Furthermore, based on the initial control topology composed of control nodes and signal connection edges in the current execution basis, the control nodes are analyzed according to the control node identifiers and signal connection relationship sets corresponding to the production site operation status data. The operation parameters, environmental monitoring values ​​and signal interaction records corresponding to the control nodes are extracted from the production site operation status data. Combined with the signal connection relationship corresponding to the signal connection edge, the status of the control nodes corresponding to the CNC control node, execution control node and detection control node is extracted and organized to obtain the real-time status information of the nodes.

[0043] S2.2: Perform transaction marking and instruction reorganization on the real-time status information of the nodes to obtain control instructions that mark transactions to be committed.

[0044] Furthermore, after extracting the state and obtaining the real-time state information of the nodes based on the current execution basis, the control instruction content (instruction information content of the control actions and signal interaction relationships of each control node) corresponding to each control node is extracted according to the control node identifier and signal connection relationship set. Based on the signal connection edges in the current execution basis, the control instruction content with signal interaction relationship is reorganized according to the signal connection relationship set to form a control instruction sequence corresponding to the initial control topology. Transaction marking processing is performed on the control instruction sequence, and the control instructions in the control instruction sequence are marked as pending transaction states to obtain the control instructions marked with pending transaction states.

[0045] It should be noted that in the subsequent processing of PCB boards, control instructions may include spindle control instructions for calling the corresponding router processing path for different board types, compensation displacement generated based on vision positioning, sorting and rejecting defective boards, and deceleration and buffering instructions generated for abnormal transmission cycle time.

[0046] S2.3: By marking control instructions that indicate the status of transactions to be committed, encapsulate the control instructions and perform data aggregation and structuring to obtain a virtual verification object.

[0047] Furthermore, by marking control instructions with pending transaction status, these instructions are associated with real-time node status information, control node identifiers, and signal connection relationship sets. The control instructions are then encapsulated based on the correspondence between control nodes and signal connection edges in the current execution basis. These instructions are then aggregated according to the control node identifiers and signal connection relationship sets. Reorganization and structural mapping are used to process the aggregated data information (representing the correspondence between control instructions and the overall organizational form). Finally, the control instructions with pending transaction status, real-time node status information, control node identifiers, and signal connection relationship sets are organized in an organizational form consistent with the current execution basis, thus obtaining a virtual verification object.

[0048] Specifically, the virtual verification object can correspond to a batch of PCB boards to be processed in the subsequent processing of the PCB boards. It includes the thickness, shape, processing path, positioning reference, spindle target speed, conveying cycle and sorting information of the batch of PCB boards. The subsequent processing of the batch of PCB boards is pre-verified in the digital twin environment.

[0049] S3: Use digital twins to perform virtual execution on the virtual verification object, calculate the physical conservation residual between the current inferred state and the current measured state, and obtain state inference information and safety envelope.

[0050] S3.1: Use digital twins to perform virtual execution on the virtual verification object to obtain a predicted state data sequence.

[0051] Furthermore, based on the current execution criteria, the control nodes, signal connection edges, and control instructions marked with pending transaction states in the virtual verification object are mapped to the digital twin. The state parsing results of the control nodes are processed sequentially, signal transmission and state updates are executed according to the signal connection relationships represented by the signal connection edges, and the control instructions marked with pending transaction states are virtually executed. The state changes of the control nodes within the current control cycle are continuously recorded, and a predicted state data sequence is formed according to the time sequence.

[0052] It should be noted that "virtual" refers to using digital twins to process each control command in sequence according to the arrangement of control commands in the virtual verification object, the signal connection relationship between control nodes, and the connection relationship corresponding to the initial control topology. Based on the virtual operation information of the previous control command, the state deduction of the next control command is continued. During the virtual operation of each control command, the state changes of the corresponding control nodes, signal transmission changes, and drive response changes are recorded synchronously until the virtual execution of all control commands in the virtual verification object is completed.

[0053] S3.2: Based on the predicted state data sequence, collect the measured state data of the current control cycle in real time, calculate the difference between the predicted state at the current moment and the measured state data in the predicted state data sequence, and obtain the physical conservation residual set.

[0054] Furthermore, based on the predicted state data sequence and in the time order of the current control cycle, the operating parameters, environmental monitoring values, and signal interaction records are acquired in real time through the acquisition and communication interface to obtain the measured state data of the current control cycle. Based on the control node identifier and signal connection relationship set, the predicted state and measured state data corresponding to the current moment in the predicted state data sequence are analyzed. The difference between the analyzed predicted state and measured state data is calculated to obtain the physical conservation residual set.

[0055] The formula for the aggregation of physical conservation residuals at control nodes is: ; in, Indicates the first The physical conservation residual value of each control node. Indicates the first The control node at the ... Predicted status data across monitoring dimensions Indicates the first The control node at the ... Actual status data across each monitoring dimension Indicates the first The calibration coefficients for each monitoring dimension This represents the total number of monitoring dimensions involved in the calculation. Indicates the control node identifier. This indicates a data dimension index.

[0056] It should be noted that, and Both are physical state quantities; both are voltage (unit: volts) and both are temperature (unit: degrees Celsius); they have the same unit. The units of the calculation results are the same as those of the calculation results. , Consistent, such as volts and degrees Celsius, It is a proportional value, without units. The numerator is the product of the unitless proportional value and the physical quantity with units. The unit of the result is the same as the physical quantity. The denominator is the sum of multiple unitless proportional values. The result is still unitless. The dimensionality of the control node physical conservation residual aggregation formula is unified.

[0057] S3.3: Define the safe range of the physical conservation residual set to obtain state deduction information and safe envelope.

[0058] Furthermore, based on the predicted state data sequence, the measured state data of the current control cycle is collected in real time, and the difference between the predicted state and the measured state data at the corresponding current moment is calculated to obtain the physical conservation residual set. The physical conservation residual set is then associated and organized with the predicted state and measured state data at the corresponding current moment according to the temporal relationship to form state prediction information. The safe range of the physical conservation residual set is defined to form a safe envelope.

[0059] For example, for PCBs with thinner board thickness, board type and processing path, stricter boundary values ​​for positioning deviation, vibration amplitude and path offset should be set to reduce the risk of board separation deviation and board edge burrs.

[0060] It should be noted that the definition of the safe range refers to the process of dividing the safe range corresponding to each physical conservation residual into boundaries based on the set of physical conservation residuals calculated from the difference between the predicted state and the measured state data at the current moment in the predicted state data sequence, and forming a safe envelope line to determine that the physical conservation residual is in a safe state.

[0061] The definition of the safety envelope mainly adopts dynamic boundary division based on process constraints and experience. It is based on the set of physical conservation residuals obtained by digital twin virtual execution, combined with the boundary values ​​of the safety range of the specific process parameters of the current production task (such as PCB board thickness, board type and processing path). For example, the residual allowable range will be configured for thinner boards and board types with complex structures. A dynamic safety envelope is formed to determine whether the residual is in a safe state. It matches and defines the physical conservation residuals with the process tolerance under specific processing scenarios.

[0062] State projection information refers to the data set formed by linking the predicted state data sequence with the measured state data collected in real time during the current control cycle in chronological order after the virtual verification object is executed virtually through digital twins. It includes the projected state of each control node at different times, the measured state at the corresponding time, and the set of physical conservation residuals obtained by difference calculation. It characterizes the state change trajectory of the control node in the production site during the virtual execution process and the deviation between the prediction and the actual situation.

[0063] S4: Based on state deduction information, physical conservation residuals exceeding the safety envelope boundary value are identified as abnormal deviation data, and risk nodes and connecting edges are identified to obtain reconstruction marker information. By deleting signal connecting edges and isolating control nodes, the initial control topology is dynamically reconstructed into the control topology based on execution.

[0064] S4.1: Based on the state deduction information, the difference between the physical conservation residual and the boundary value of the safety envelope is determined, and the physical conservation residual exceeding the boundary value of the safety envelope is recorded as abnormal information to obtain abnormal deviation data.

[0065] Furthermore, based on the state deduction information, according to the temporal arrangement of the physical conservation residual set within the current control cycle, the boundary values ​​of the safety envelope corresponding to the current moment in the state deduction information are obtained item by item, and the deviation comparison is performed between the physical conservation residual and the boundary values ​​of the safety envelope at the same moment. During the deviation comparison process, the difference is determined as to whether the physical conservation residual is within the safety envelope. The physical conservation residual within the safety envelope is recorded as normal information, and the physical conservation residual outside the safety envelope is recorded as abnormal information, thus obtaining abnormal deviation data.

[0066] The formula for calculating the safety compliance of the signal connection edge is: ; in, Indicates the first Security compliance score for each signal connection edge. Indicates the first The signal connection edge is at the first The deviation value at each time step Indicates the first The signal connection edge is at the first The envelope boundary values ​​at each time step. Indicates the first Sampling coefficients at each time step This indicates the total number of sampling time steps within the current control cycle. Indicates the signal connection edge identifier. Indicates the time step index.

[0067] It should be noted that, It is the physical conservation residual value, with units such as volts and degrees Celsius. These are envelope boundary values, and the units are volts and degrees Celsius. It involves dividing two physical quantities with the same unit; the units cancel each other out, and the result is a pure number without units. It's a ratio, it has no unit. and Both are additions of unitless values, and the result is still unitless. It is purely digital, and the dimensions of the calculation formula for the safety compliance of the signal connection edge are unified.

[0068] The difference determination is based on the state deduction information to obtain the safety envelope boundary value corresponding to the physical conservation residual time series. The physical conservation residual at each time is compared with the corresponding safety envelope boundary value to determine whether the physical conservation residual is within the safety envelope. The physical conservation residual that exceeds the safety envelope is identified, and the abnormal deviation data is processed.

[0069] S4.2: Based on the abnormal deviation data, the control nodes that correspond to the abnormal deviation data and have continuous difference judgment information are identified as risk nodes, and the signal connection edges that have signal connection relationships with the risk nodes and correspond to the abnormal deviation data are identified as risk connection edges. The risk nodes are marked as isolated objects and the risk connection edges are marked as deleted objects to obtain reconstruction marking information.

[0070] Furthermore, based on the abnormal deviation data, the abnormal deviation data is mapped to the control node identifiers and signal connection relationship sets in the initial control topology according to the physical conservation residual time sequence. The control nodes and signal connection edges corresponding to the abnormal deviation data are determined. Combined with the state inference information, the real-time state information of the control nodes, the control instructions marked with the state of pending transactions, the physical conservation residual set and the boundary values ​​of the safety envelope are correlated and compared. The control nodes that correspond to the abnormal deviation data and have a continuous difference judgment are identified as risk nodes. The signal connection edges that have a signal connection relationship with the risk nodes and correspond to the abnormal deviation data are marked in the initial control topology to obtain the reconstruction marking information.

[0071] For example, when the visual positioning deviation continuously exceeds the safety envelope, the visual positioning control node can be identified as a risk node. When the spindle vibration amplitude and tool wear value are continuously abnormal, the corresponding router spindle control node can be identified as a risk node. When the AOI detection results show that a batch of PCB boards has continuous board edge abnormalities, the corresponding detection node and associated sorting node can be marked as reconstruction objects.

[0072] S4.3: Based on the reconstruction marker information, the initial control topology is deleting signal connection edges and isolating control nodes to form the control topology for execution.

[0073] Furthermore, based on the reconstruction marking information, the topology reconstruction method is used to process the control nodes and signal connection edges in the initial control topology item by item. According to the risk node and connection edge identification information corresponding to the abnormal deviation data, the reconstruction is performed in the priority order of first isolating nodes, then deleting connection edges, and finally retaining unmarked relationships. The control nodes marked as isolated by the reconstruction marking information are isolated and all signal connection edges between the isolated control nodes and the other control nodes are cut off.

[0074] The signal connection edges marked as to be deleted by the reconstruction marking information are deleted, the specific signal connection relationships between non-isolated control nodes are removed, and the control nodes and signal connection edges not marked by the reconstruction marking information are retained. The structure mapping and reorganization of the initial control topology are completed, and the control topology based on the execution is formed.

[0075] Specifically, when the vision positioning control node is determined to be abnormal, the compensation signal connection edge between it and the router spindle control node can be deleted, and the corresponding PCB board can be suspended from entering the router and sorting station. When the tool wear value exceeds the normal value, the corresponding router spindle control node can be isolated, and the station can be switched to a standby station and the tool change process can be triggered. When the AOI detection node identifies a defective board, the diversion connection between the conveyor node and the sorting node can be retained, and the defective board can enter the designated buffer and rejection path.

[0076] Topology reconstruction method refers to a processing method that adjusts the structure of control nodes and signal connection edges in the initial control topology based on reconstruction marker information. Based on the reconstruction marker information, risk nodes and their corresponding signal connection edges are identified. The signal connection edges corresponding to the risk nodes are deleted to cut off the signal transmission path corresponding to abnormal deviation data. Control node isolation processing is performed on the risk nodes to remove them from the connection relationship in the current control topology, while retaining the connection correspondence between the signal connection edges that have not been deleted and the control nodes that have not been isolated.

[0077] S5: When the physical conservation residual is within the safety envelope and there is no control instruction conflict with the isolated control node, the atomic change of the pending transaction status to the committed transaction status is issued to the CNC drive based on the execution base, and control feedback information is obtained.

[0078] S5.1: Based on the safety envelope, perform interval matching judgment on the physical conservation residuals within the current control cycle, and perform conflict detection on the control instructions corresponding to the isolated control nodes to obtain transaction commit judgment information.

[0079] Furthermore, based on the safety envelope, the physical conservation residuals within the current control period are time-series matched to obtain the safety envelope boundary values ​​corresponding to the time sequence of the physical conservation residuals. The physical conservation residuals within the current control period are then matched with the safety envelope boundary values ​​to determine whether the physical conservation residuals are located within the safety envelope.

[0080] Conflict detection is performed on the control instructions corresponding to the isolated control nodes. The control instructions marked with the transaction status to be committed are compared with the isolated control nodes. The execution status of the control instructions in the current control cycle is checked to determine whether there is a control instruction conflict for the control instructions corresponding to the isolated control nodes. The determination information of whether the physical conservation residual is within the safe envelope is compared with the control instruction execution conflict detection result to obtain the transaction commit determination information.

[0081] Specifically, conflict detection involves iterating through the set of control commands to be submitted on the isolated control node, comparing the status of the signal connection edges of each control command with the isolated control topology. If the signal connection edge on which the command depends has been deleted and the target control node of the command is in an isolated state, it is determined that there is a control command conflict. At the same time, it detects the occupation of the same physical execution resource by multiple commands in the same time step. If resource competition and timing contradictions occur, they are marked as conflict commands, conflict detection information is output, and it is determined whether to allow the transaction state switch.

[0082] S5.2: When the physical conservation residual of the transaction commit determination information is within the safe envelope and there is no control instruction conflict, perform atomic state switching processing on the control instructions marked as pending transaction state, change the pending transaction state to committed transaction state, and obtain the set of committed control instructions.

[0083] Furthermore, when the transaction commit determination information indicates that the physical conservation residual is within the safe envelope and there is no control instruction conflict, the control instructions marked with the pending transaction status are filtered based on the transaction commit determination information. The control instructions marked with the pending transaction status that have no control instruction conflict with the isolated control node and whose corresponding physical conservation residual is within the safe envelope are identified as state switching objects. The state switching objects are then uniformly updated according to the atomic state switching process. In the same process, the pending transaction status of the state switching object is directly changed to the committed transaction status. The control instructions that have completed the change of the committed transaction status are collected to obtain the set of committed control instructions.

[0084] S5.3: Based on the submitted set of control instructions, perform drive mapping on the control topology of the execution basis, generate drive execution signals that match the corresponding control nodes, and send them to the corresponding drive execution.

[0085] Furthermore, based on the set of submitted control instructions, and according to the correspondence between control nodes and signal connection edges in the control topology of the execution basis, drive mapping processing is performed on each submitted control instruction in the set of submitted control instructions. The control node identifier, instruction execution status, and signal connection relationship corresponding to the submitted control instruction are matched, and the matching information is converted into CNC drive execution signals consistent with the corresponding control nodes. After the CNC drive execution signals are generated, the CNC drive execution signals are sent to the corresponding drive executions according to the drive mapping results.

[0086] Specifically, drive mapping is a process based on the set of submitted control instructions and the control topology of the execution basis. It matches the submitted control instructions corresponding to each control node according to the correspondence between the control node and the CNC drive, determines the CNC drive execution object and connection relationship corresponding to each submitted control instruction, and converts the submitted control instructions into the corresponding CNC drive execution signals.

[0087] S5.4: During the execution of CNC drive, drive response data is collected in real time, and the collected operation response data is extracted and processed to obtain control feedback information.

[0088] Furthermore, during the CNC drive execution process, the control topology of the execution basis is driven and mapped based on the submitted control instruction set to generate the corresponding CNC drive execution signal and send it to the corresponding drive execution. At the same time, the drive response data is collected in real time through the acquisition and communication interface. The running response data corresponding to the control node running status, instruction execution status and drive response status is extracted from the drive response data. The collected running response data is then processed to obtain control feedback information.

[0089] S6: The incremental factor elimination method is used to perform iterative processing on the control feedback information to eliminate error factors, correct the configuration parameters of control node signal reliability and control priority, and form the initial control topology for the next cycle.

[0090] S6.1: Based on control feedback information, extract feedback feature data corresponding to the operating status, command execution status and drive response status of the control node, and use incremental factor elimination method to iteratively remove error factors and filter noise in the feedback feature data to obtain feature data after eliminating interference.

[0091] Furthermore, based on control feedback information, the operating status of the control node, the execution order of control commands, and the drive response time are correlated and extracted according to the correspondence between control node identification, control command execution order, and drive response time to form feedback feature data. The feedback feature data is processed using incremental factor elimination. Combining the changes in control feedback information within a continuous control cycle, error factors in the feedback feature data are iteratively stripped off one by one. After each iteration of stripping, noise filtering is performed on the remaining feedback feature data, retaining the effective feature parts corresponding to the operating status of the control node, the command execution status, and the drive response status. The control cycle continues to advance, and the iterative stripping of error factors and noise filtering are repeatedly performed until the interference components in the feedback feature data are gradually eliminated, obtaining the feature data after interference elimination.

[0092] Specifically, feedback feature data refers to the feature information content of the control node's operating status, instruction execution status, and drive response status.

[0093] The incremental factor elimination method is based on the feedback feature data extracted from the control feedback information. According to the order of the control cycle, the error factors affecting the operating state, command execution state and drive response state of the newly added feedback feature data are gradually separated and iteratively stripped. In each iteration, the influence of error factors is eliminated and noise is filtered in combination with the existing feedback feature data, so as to retain the feature information that represents the actual feedback state.

[0094] The execution process of incremental factor elimination is to extract features based on the changes within the continuous control cycle, and then, in conjunction with the changes in the control feedback information within the continuous control cycle, identify the feature parts in the feedback feature data that correspond to the physical conservation residuals exceeding the boundary value of the safety envelope.

[0095] Error factors are removed based on abnormal deviation data. The feature parts whose corresponding physical conservation residuals exceed the boundary value of the safety envelope are identified as error factors and iteratively stripped from the feedback feature data.

[0096] Noise filtering is performed to retain valid features. After each iteration of stripping, noise filtering is performed on the remaining feedback feature data to retain the valid feature parts corresponding to the control node's running state, instruction execution state, and drive response state.

[0097] The process continues iteratively until the interference is eliminated. The control cycle continues to advance, and error factor iterative stripping and noise filtering are repeatedly performed until the interference components in the feedback feature data are gradually eliminated, thus obtaining the feature data after interference elimination.

[0098] S6.2: Based on the characteristic data after interference elimination, the configuration parameters of the control node signal reliability and control priority are numerically corrected and updated to obtain the updated configuration parameter set.

[0099] Furthermore, based on the feature data after interference elimination, the updated values ​​of control node signal reliability and control priority are directly calculated. The feature data after interference elimination is matched item by item with the configuration parameters of control node signal reliability and control priority. According to the control node operating status, command execution status and drive response status represented in the feature data after interference elimination, the configuration parameters of control node signal reliability and control priority are numerically corrected and updated to obtain the updated set of configuration parameters.

[0100] S6.3: Based on the updated set of configuration parameters, perform structural mapping and reorganization of the signal connection relationships between control nodes to form the initial control topology for the next control cycle.

[0101] Furthermore, based on the updated configuration parameter set, when performing structural mapping and reorganization of the signal connection relationship between control nodes, the updated configuration parameter set is obtained according to the feature data after interference elimination. The signal reliability and control priority of the control nodes in the updated configuration parameter set are used as the basis for structural mapping and reorganization processing. The updated configuration parameter set is mapped to the signal connection relationship between control nodes, and structural mapping and reorganization processing is performed on the control nodes and signal connection edges to form the initial control topology for the next control cycle.

[0102] In summary, this invention, by comparing the physical conservation residuals of the deduced state and the measured state, forms a reliable and secure envelope, providing a basis for dynamic topology reconstruction, improving the control and adaptation capability of switching between different board types in subsequent PCB processing, reducing the risk of decreased board separation accuracy, increased burrs and missorting, and improving yield and production line stability.

[0103] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for optimizing and controlling an intelligent manufacturing production line based on digital twins, characterized in that, include: Acquire production site operation status data and construct an initial control topology consisting of control nodes and signal connection edges, which serves as the basis for current execution; Based on the current execution criteria, the control node status is parsed and signal encapsulation processing is performed to generate control instructions marked with the status of transactions to be submitted, thus forming a virtual verification object; A digital twin is used to perform virtual execution on the virtual verification object, calculate the physical conservation residual between the current inferred state and the current measured state, and obtain the state inference information and the safety envelope. Based on state inference information, physical conservation residuals exceeding the safety envelope boundary value are identified as abnormal deviation data, and risk nodes and connecting edges are identified to obtain reconstruction marker information. By deleting signal connecting edges and isolating control nodes, the initial control topology is dynamically reconstructed into the control topology of the execution basis. When the physical conservation residual is within the safety envelope and does not conflict with the control instructions of the isolated control node, the atomic change of the pending transaction status to the committed transaction status is sent to the CNC drive execution based on the execution-based control topology to obtain control feedback information. Incremental factor elimination is used to iteratively process the control feedback information to eliminate error factors, correct the configuration parameters of control node signal reliability and control priority, and form the initial control topology for the next cycle.

2. The intelligent manufacturing production line optimization control method based on digital twin as described in claim 1, characterized in that, The production site operation status data includes spindle vibration voltage signal, positioning shaft temperature signal, and actual position deviation value.

3. The intelligent manufacturing production line optimization control method based on digital twin as described in claim 1, characterized in that, The control node refers to the logical object representing an independent control entity in the initial control topology, corresponding to the CNC control node, execution control node, and detection control node in the production site.

4. The intelligent manufacturing production line optimization control method based on digital twin as described in claim 1, characterized in that, The current execution basis is an initial control topology structure composed of control nodes and signal connection edges, constructed based on the production site operation status data.

5. The intelligent manufacturing production line optimization control method based on digital twin as described in claim 4, characterized in that, The specific steps for constructing the initial control topology and using it as the basis for current execution are as follows: By collecting and analyzing the operational status data of the production site in real time, the control node identifiers and signal connection relationships are obtained. The control node identifiers and signal connection relationships are reorganized and structurally mapped to construct an initial control topology consisting of control nodes and signal connection edges, which serves as the basis for current execution.

6. The intelligent manufacturing production line optimization control method based on digital twin as described in claim 1, characterized in that, The specific steps for creating the virtual verification object are as follows: Based on the current execution criteria, the execution status extraction process parses the running data of the control node to obtain the real-time status information of the node. The node's real-time status information is marked for transactions and reorganized with instructions to obtain control instructions marked with transactions to be committed. By marking control instructions that indicate the status of transactions to be committed, encapsulating the control instructions, and performing data aggregation and structuring, a virtual verification object is obtained.

7. The intelligent manufacturing production line optimization control method based on digital twin as described in claim 1, characterized in that, The specific steps for obtaining the state deduction information and the safety envelope are as follows: A digital twin is used to virtually execute a virtual verification object to obtain a predicted state data sequence. Based on the predicted state data sequence, the measured state data of the current control cycle is collected in real time. The difference between the predicted state data sequence corresponding to the current moment and the measured state data is calculated to obtain the physical conservation residual set. Define the safe range of the physical conservation residual set to obtain state deduction information and safety envelope.

8. The intelligent manufacturing production line optimization control method based on digital twin as described in claim 1, characterized in that, The reconstructing of the control topology as the execution basis involves the following steps: Based on state extrapolation information, the difference between the physical conservation residual and the boundary value of the safety envelope is determined, and the physical conservation residual exceeding the boundary value of the safety envelope is recorded as abnormal information to obtain abnormal deviation data. Based on the abnormal deviation data, control nodes that correspond to the abnormal deviation data and have continuous difference judgment information are identified as risk nodes, and signal connection edges that have signal connection relationships with risk nodes and correspond to abnormal deviation data are identified as risk connection edges. Risk nodes are marked as isolation objects and risk connection edges are marked as deletion objects to obtain reconstruction marking information. Based on the reconstructed labeling information, the initial control topology is deleting signal connection edges and isolating control nodes to form the control topology for execution.

9. The intelligent manufacturing production line optimization control method based on digital twin as described in claim 1, characterized in that, The specific steps for obtaining control feedback information are as follows: Based on the safety envelope, interval matching is performed on the physical conservation residuals within the current control cycle, and conflict detection is performed on the control instructions corresponding to the isolated control nodes to obtain transaction commit determination information. When the physical conservation residual characterization of the transaction commit determination information is within the safe envelope and there is no control instruction conflict, atomic state switching processing is performed on the control instructions marked as pending transaction state to change the pending transaction state to committed transaction state and obtain the set of committed control instructions. Based on the submitted set of control instructions, drive mapping is performed on the control topology of the execution basis to generate drive execution signals that match the corresponding control nodes and send them to the corresponding drive execution. During the execution of CNC drive, drive response data is collected in real time, and the collected operation response data is extracted and processed to obtain control feedback information.

10. The intelligent manufacturing production line optimization control method based on digital twin as described in claim 1, characterized in that, The specific steps for forming the initial control topology for the next cycle are as follows: Based on control feedback information, feedback feature data corresponding to the operating status, command execution status and drive response status of the control node are extracted. Incremental factor elimination method is used to iteratively remove error factors and filter noise in the feedback feature data to obtain feature data after eliminating interference. Based on the feature data after interference elimination, the configuration parameters of control node signal reliability and control priority are numerically corrected and updated to obtain the updated set of configuration parameters. Based on the updated set of configuration parameters, the signal connection relationships between control nodes are structurally mapped and reorganized to form the initial control topology for the next control cycle.

Images