Machine tool welding intelligent control system based on digital twinning

Abstract

The invention discloses a machine tool welding intelligent control system based on digital twinning, and belongs to the technical field of machine tool welding control, a simulation module establishes a corresponding digital twinning model according to machine tool welding information, and a target welding scheme is determined based on the digital twinning model; welding monitoring data sent by the monitoring module are obtained in real time, analogue simulation is conducted on the welding monitoring data through a digital twin model, a welding control instruction is obtained, and the welding control instruction is sent to the control analysis module after being stamped with a timestamp; the monitoring module is used for conducting real-time monitoring in the welding process, obtaining welding monitoring data and sending the welding monitoring data to the simulation module of the cloud end. The control analysis module is used for performing control analysis on welding, receiving the welding control instruction sent by the simulation module, identifying a timestamp corresponding to the welding control instruction, performing aging analysis according to the timestamp, and generating a control execution instruction according to an aging analysis result; the welding control module is used for conducting welding control according to the control execution instruction.

Description

Technical Field

[0001] This invention belongs to the field of machine tool welding control technology, specifically a machine tool welding intelligent control system based on digital twins. Background Technology

[0002] Machine tool welding, as a core process in high-end equipment manufacturing, directly determines the structural strength, service life, and reliability of mechanical products. Traditional welding control methods mainly rely on manual experience to set parameters or open-loop control based on offline simulation. These methods struggle to respond in real time to complex physical phenomena such as dynamic changes in the molten pool, thermal stress accumulation, and nonlinear material deformation during welding. This results in poor weld uniformity and frequent defects such as porosity and cracks. Especially in high-value-added fields such as aerospace and new energy vehicle battery trays, unstable welding quality has become a key bottleneck restricting industrial upgrading.

[0003] In existing technologies, some companies have attempted to introduce closed-loop feedback control technology, constructing real-time adjustment systems based on PID algorithms or fuzzy control by arranging sensors such as temperature, current, and displacement on the welding torch or workpiece. However, such solutions have significant limitations: on the one hand, a single sensor can only acquire local physical quantities and cannot comprehensively reflect the multi-physics coupling state of the molten pool, base material, and environment, leading to one-sided control decisions; on the other hand, traditional control algorithms are not adaptable to nonlinear and time-varying systems, and are prone to overshoot or oscillation when facing conditions such as sudden changes in welding speed or material thickness, which can exacerbate quality fluctuations. In addition, the disconnect between offline simulation and actual production remains unresolved. Parameter deviations between the simulation model and the real welding environment (such as changes in material thermophysical parameters with temperature) are directly transmitted to the control loop, forming a vicious cycle of "model distortion - control failure". Summary of the Invention

[0004] To address the problems of the above solutions, this invention provides a machine tool welding intelligent control system based on digital twins.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A machine tool welding intelligent control system based on digital twins, including cloud and equipment terminals;

[0007] The cloud platform includes a simulation module;

[0008] The simulation module establishes a corresponding digital twin model based on the machine tool welding information, determines the target welding scheme based on the digital twin model, and sends the target welding scheme to the equipment.

[0009] Welding monitoring data sent by the equipment is acquired in real time. The welding monitoring data is simulated using a digital twin model to obtain corresponding welding control commands. The welding control commands are then timestamped and sent to the equipment.

[0010] Furthermore, the target welding scheme is determined based on the digital twin model, including:

[0011] Step SA1: Obtain welding target information and match the corresponding initial welding scheme based on the welding target information;

[0012] Step SA2: Simulate the initial welding scheme using a digital twin model to obtain the simulation results of the initial welding scheme;

[0013] Step SA3: Evaluate whether the simulation results of the proposed scheme meet the welding standards;

[0014] When the simulation results of the evaluation scheme meet the welding standards, the initial welding scheme is sent to the user for confirmation. The initial welding scheme confirmed by the user is marked as the target welding scheme, and the analysis ends.

[0015] When the simulation results of the evaluation scheme do not meet the welding standards, the initial welding scheme is adjusted according to the simulation results to obtain a new initial welding scheme, and then the process returns to step SA2.

[0016] Furthermore, a corresponding initial welding scheme is matched based on the welding target information, including:

[0017] Based on machine tool welding information and user product information, identify various potential welding targets of the user; obtain information on each potential welding target; based on machine tool welding information and potential welding target information, identify various welding schemes and mark them as reserve welding schemes for the corresponding potential welding targets; establish a reserve database based on the reserve welding schemes corresponding to each potential welding target.

[0018] Identify welding target information, input the welding target information into the reserve database for matching, obtain several reserve welding schemes, evaluate the priority of the matched reserve welding schemes, and mark the reserve welding scheme with the highest priority as the initial welding scheme.

[0019] Furthermore, the simulation results of the evaluation scheme are assessed to determine whether they meet welding standards, including:

[0020] Obtain welding standards, and collect features from the simulation results of the scheme based on the welding standards to obtain simulation feature data of the corresponding welding points;

[0021] A quality assessment model is established, and the expression of the quality assessment model is as follows:

[0022] ;

[0023] In the formula: (s, BZ) are the input data, s are the simulation feature data, and BZ are the welding standard; s→BZ indicates that the corresponding simulation feature data meets the welding standard; the output data is the quality assessment value ZP(s, BZ), and the quality assessment value is 1 or 0.

[0024] The simulation feature data and welding standards are integrated into corresponding input data and input into the quality assessment model for analysis to obtain the quality assessment value of the corresponding weld.

[0025] The individual quality assessment values ​​are summarized into a set of quality assessment values ​​for the simulation results of the scheme.

[0026] When the quality assessment value set contains a quality assessment value of 0, the simulation results of the assessment scheme do not meet the welding standards.

[0027] When there is no quality assessment value of 0 in the set of quality assessment values, the simulation results of the assessment scheme meet the welding standards.

[0028] The equipment includes a monitoring module, a control and analysis module, and a welding control module;

[0029] The monitoring module is used to perform real-time monitoring during the welding process, obtain corresponding welding monitoring data, and send the welding monitoring data to the simulation module in the cloud.

[0030] The control and analysis module is used to control and analyze the welding process. It receives welding control commands sent from the cloud in real time, identifies the timestamps corresponding to the welding control commands, sets a time axis, and marks the corresponding command points on the time axis according to the timestamps of each welding control command. It marks the command point closest to the current time on the time axis as the base point. It acquires welding time points in real time and marks the welding time point closest to the current time on the time axis. It marks the current time axis as the time command axis.

[0031] Based on the time command axis, the timestamps of the welding control commands received in real time are analyzed to obtain the time analysis results of the corresponding welding control commands. Based on the time analysis results, the corresponding control execution commands are generated and sent to the welding control module.

[0032] Furthermore, a timeliness analysis is performed on the timestamps of the welding control commands received in real time based on the time command axis, including:

[0033] Set aging standards based on user welding requirements; establish an aging analysis model, the expression of which is:

[0034] ;

[0035] In the formula: (TB, BS) are the input data, TB is the aging feature identified according to the time command axis, the aging feature includes the welding time point and the command aging value, BS is the aging standard, TB→BS means that the corresponding aging feature meets the aging standard; the output data is the aging analysis value SA(TB, BS), the aging analysis value is 1 or 0;

[0036] Based on the time command axis, the corresponding aging characteristics are identified in real time. The aging characteristics and aging standards are integrated into the input data and input into the aging analysis model for analysis to obtain the aging analysis value of the corresponding welding control command.

[0037] When the aging analysis value is 1, the aging analysis result is that the aging is qualified.

[0038] When the aging analysis value is 0, the corresponding instruction point in the time instruction axis is marked as a time delay instruction point, and the aging analysis result is that the aging is unqualified.

[0039] Furthermore, based on the timeliness analysis results, corresponding control execution instructions are generated, including:

[0040] When the timeliness analysis result is that the timeliness is qualified, the control execution instruction will be marked as a control execution instruction;

[0041] When the aging analysis result is that the aging is unqualified, the welding monitoring data is analyzed according to the quality assessment model to obtain the quality assessment value; when the quality assessment value is 1, the control execution command is generated according to the target welding plan; when the quality assessment value is 0, the preset emergency handling method is used.

[0042] The welding control module is used to perform welding control according to the received control execution instructions.

[0043] Compared with the prior art, the beneficial effects of the present invention are:

[0044] Digital twin technology deeply integrates physical entities with digital space by creating a virtual mapping model of welding equipment, process parameters, material conditions, and environmental conditions, forming a dynamic data closed loop covering the entire welding lifecycle. This model can record and analyze the physical state changes at each stage of the welding process in real time, providing complete data chain support for quality traceability. It improves the accuracy of locating the root cause of welding defects to the process parameter level, completely changing the passive "result detection - rework" mode of traditional quality control, and promoting the transformation of the manufacturing process towards proactive quality management of "prevention - prediction - pre-control". Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. 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.

[0046] Figure 1 This is a block diagram illustrating the principle of the present invention. Detailed Implementation

[0047] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0048] like Figure 1 As shown, a machine tool welding intelligent control system based on digital twins includes a cloud and an equipment terminal; the cloud and the equipment terminal are connected for communication.

[0049] The cloud platform is built upon existing cloud technologies, including simulation modules;

[0050] The simulation module establishes a corresponding digital twin model based on the machine tool welding information, which includes relevant information such as machine tool equipment and systems. It also deploys corresponding sensors at the equipment according to the simulation requirements of the digital twin model to achieve data monitoring and transmission. Based on the digital twin model, it determines the target welding scheme and sends the target welding scheme to the equipment.

[0051] Welding monitoring data sent by the equipment is acquired in real time. The welding monitoring data is simulated using a digital twin model to obtain corresponding welding control commands. The welding control commands are then timestamped and sent to the equipment.

[0052] In one embodiment, the digital twin model is built based on existing digital twin technology to simulate the welding system of the machine tool. Example:

[0053] 3D model construction:

[0054] Use CAD tools (such as SolidWorks and AutoCAD) to create 3D models of key components such as the machine tool body, welding torch, and fixtures, retain core structures such as motion axes, grinding wheels, and workpieces, and simplify non-critical components such as cooling systems and hydraulic devices to reduce computational load.

[0055] Rendering scenes can be built using engines such as Unity3D, enabling environmental settings such as lighting, camera perspective, and model interaction, and supporting visual monitoring across multiple terminals (PC, tablet, VR devices).

[0056] Motion relationship binding:

[0057] The relative motion logic between components is defined based on assembly constraints, such as the coupling relationship between the rotary axis and the translation axis of a five-axis linkage machine tool, and the relative trajectory between the welding torch and the workpiece.

[0058] By expressing motion logic through topological chain structures, the virtual model can accurately simulate the dynamic behavior of the physical system.

[0059] Logical modeling: Dynamic simulation and control of welding processes;

[0060] Logical modeling focuses on the dynamic logic of the welding process. Through modules such as virtual CNC system, coordinate system setting, and CNC code parsing, it realizes synchronous simulation and offline verification of the physical welding process by the virtual model.

[0061] Virtual CNC system construction:

[0062] The CNC system of a physical machine tool (such as FANUC or SIEMENS) is projected into a virtual space to create a virtual panel, which supports the remote issuance and execution feedback of operation commands.

[0063] Implement key information management functions, such as real-time display and historical record query of welding parameters (current, voltage, speed).

[0064] Coordinate system and motion logic definition:

[0065] Establish absolute coordinate system, machine tool coordinate system, workpiece coordinate system and component coordinate system, clarify the transformation relationship between each coordinate system, and ensure that the motion trajectory of the virtual model is consistent with the physical system.

[0066] For welding process characteristics (such as dynamic changes in the molten pool), a geometric abstract model of the grinding wheel (or welding gun) is defined. A general cross-sectional model is generated through parameterized input (such as grinding wheel radius and rotation speed), and a three-dimensional solid is generated by rotation and stretching.

[0067] CNC code analysis and verification:

[0068] Develop a CNC code parser that supports syntax recognition and instruction conversion for different CNC systems (G code, ISO code) to ensure that the virtual model can accurately simulate the machining process of the physical machine tool.

[0069] Offline simulation verifies the correctness of CNC programs, detects errors such as collisions and overcuts in advance, and reduces trial and error costs in actual welding.

[0070] Data modeling: Fusion and dynamic mapping of multi-source data;

[0071] Data modeling is the core of digital twins. By collecting and fusing sensor data, historical logs, and external system data, a real-time data-driven dynamic model is constructed to achieve synchronous updates between the physical system and the virtual model.

[0072] Data acquisition and transmission:

[0073] Deploy a sensor network (temperature, current, vibration, displacement sensors) to collect physical data during the welding process in real time, and upload it to the cloud or edge computing nodes through IoT platforms (such as AWS IoT, Azure Digital Twins) or industrial protocols (MQTT, OPC UA).

[0074] Integrate historical data from existing systems (ERP, SCADA) and supplement it with structured and unstructured data such as equipment maintenance records and process parameter adjustment logs.

[0075] Data cleaning and fusion:

[0076] The collected data is cleaned to remove noise (such as outliers caused by sensor malfunctions), fill in missing values ​​(such as using interpolation to generate temporary data), and standardize the format (unify timestamps and units) to ensure data quality.

[0077] Data is preprocessed by edge computing nodes (such as industrial gateways) to reduce transmission latency, and multi-source heterogeneous data is integrated using data fusion algorithms (such as Kalman filtering) to generate comprehensive evaluation indicators of component status (such as the Equipment Health Index HI).

[0078] Real-time data-driven model updates:

[0079] The fused data is then bound to the corresponding components of the virtual model, driving dynamic updates to the model. For example, the 3D morphology model of the molten pool can be adjusted based on real-time temperature data, and the load state model of the machine tool spindle can be updated based on vibration data.

[0080] Establish a historical database to store static attribute data (equipment parameters, workpiece materials) and real-time processing data, supporting model training and optimization analysis (such as grinding wheel wear prediction and welding defect diagnosis).

[0081] Hybrid modeling: synergistic optimization of mechanism and data-driven approaches;

[0082] To improve the accuracy and adaptability of the model, it is necessary to combine mechanistic models and data-driven models to build a hybrid modeling framework.

[0083] Mechanism model:

[0084] Simulation models of weld pool flow and thermal stress distribution are constructed based on physical laws (such as thermodynamic equations and Newtonian mechanics), and high-precision simulations are achieved using tools such as ANSYS and MATLAB / Simulink.

[0085] The transfer function is established using the system identification toolbox to describe the dynamic response characteristics of the welding system (such as step response overshoot and settling time).

[0086] Data-driven model:

[0087] Behavioral models are trained using machine learning algorithms (LSTM, random forest) to predict welding quality indicators (such as weld reinforcement and porosity) or equipment failures (such as spindle overheating and sensor failure).

[0088] By extracting key features (such as the frequency domain energy of the current waveform and the gradient distribution of the temperature field) through feature engineering, the complexity of the model can be reduced and the prediction accuracy can be improved.

[0089] Multi-model fusion and iterative optimization:

[0090] Digital threading technology is used to achieve dynamic coupling between the mechanistic model and the data-driven model. For example, the data model can be used to correct the parameter deviations of the mechanistic model (such as the thermal conductivity of the molten pool), or the mechanistic model can be used to constrain the output range of the data model (such as preventing the predicted weld reinforcement value from exceeding the physical limit).

[0091] The model parameters are continuously updated through online learning mechanisms (such as federated learning) to adapt to changes in welding processes (such as material changes and environmental temperature fluctuations), ensuring the long-term effectiveness of the model.

[0092] Validation and Deployment: From Virtual Simulation to Actual Production;

[0093] Model validation:

[0094] Static verification: Check whether the geometric model dimensional error (<0.1%) and material properties meet the standards (such as ASTM specifications).

[0095] Dynamic validation: The model overshoot error (<5%) is evaluated through step response testing, and the prediction error (RMS≤3σ) is validated through backtesting on historical data.

[0096] Fault injection testing: Simulate abnormal scenarios such as sensor failure and network latency to test the robustness and fault tolerance of the model.

[0097] Continuous Evolution:

[0098] By using the Digital Twin as a Service (DTaaS) model, design changes to physical entities (such as new 3D models after equipment upgrades) are synchronized regularly, and the models are retrained with new data to prevent model drift (such as a continuous increase in prediction errors).

[0099] By introducing cutting-edge technologies such as quantum-inspired algorithms and neural symbolic systems, we can improve the efficiency of solving combinatorial optimization problems and expand the application scenarios of the model (such as cross-production line collaborative optimization).

[0100] In one embodiment, determining the target welding scheme based on a digital twin model includes:

[0101] Step SA1: Obtain welding target information and match the corresponding initial welding scheme based on the welding target information;

[0102] Step SA2: Simulate the initial welding scheme using a digital twin model to obtain the corresponding scheme simulation results. That is, perform welding simulation on the welding target according to the initial welding scheme to obtain the corresponding welding simulation results, and integrate them into the scheme simulation results.

[0103] Step SA3: Evaluate whether the simulation results of the scheme meet the welding standards;

[0104] When the simulation results of the evaluation scheme meet the welding standards, the initial welding scheme is sent to the user for confirmation. The initial welding scheme confirmed by the user is marked as the target welding scheme. If the user adjusts the received initial welding scheme, it is used as a new initial welding scheme for re-verification. However, if the verification is successful, the user does not need to confirm and it is directly used as the target welding scheme. However, if the scheme is adjusted, the user needs to confirm. The analysis ends.

[0105] When the simulation results of the evaluation scheme do not meet the welding standards, the initial welding scheme is adjusted according to the simulation results to obtain a new initial welding scheme, and then the process returns to step SA2.

[0106] In one embodiment, matching an appropriate initial welding scheme based on the welding target information can be performed in an existing manner to determine the applicable initial welding scheme, or the user-defined welding scheme can be directly used as the initial welding scheme for matching.

[0107] In one embodiment, matching a corresponding initial welding scheme based on welding target information includes:

[0108] Based on machine tool welding information and user product information, identify various welding targets that the user may have and mark them as potential welding targets. Alternatively, users can directly provide various potential welding targets. Obtain information on each potential welding target. Based on machine tool welding information and potential welding target information, identify various welding schemes and mark them as reserve welding schemes for the corresponding potential welding targets. Establish a reserve database based on the reserve welding schemes corresponding to each potential welding target.

[0109] Identify welding target information, input the welding target information into the reserve database for matching, obtain several reserve welding schemes, evaluate the priority of the matched reserve welding schemes, and mark the reserve welding scheme with the highest priority as the initial welding scheme.

[0110] In one embodiment, the matching reserve welding schemes are prioritized. In addition to prioritizing according to existing priority assessment algorithms, simulations can also be performed based on digital twin models, and the priorities are determined based on the simulation results.

[0111] In one embodiment, the initial welding scheme is adjusted based on the simulation results. This involves adjusting welding parameters that do not meet the welding standards based on a digital twin model, so that they meet the welding standards.

[0112] In one embodiment, whether the simulation results of the evaluation scheme meet the welding standards can be determined based on existing methods.

[0113] In one embodiment, evaluating whether the simulation results of the scheme meet the welding standards includes:

[0114] Welding standards are typically provided by the user, or conventional industry standards can be used, or welding standards can be determined based on the user's customer requirements. Based on the welding standards, the simulation results of the solution are used to collect features and obtain the simulation feature data of the corresponding weld. For example, if a requirement of the welding standard is that the diameter of a single pore is ≤0.3σ (σ is the thickness of the base material), then the corresponding pore diameter feature is extracted for standard evaluation.

[0115] A quality assessment model is established, and the expression of the quality assessment model is as follows:

[0116] ;

[0117] In the formula: (s, BZ) are the input data, s are the simulation feature data, and BZ are the welding standard; s→BZ indicates that the corresponding simulation feature data meets the welding standard; the output data is the quality assessment value ZP(s, BZ), and the quality assessment value is 1 or 0; the corresponding normal state is determined by training the corresponding historical data labeled training set.

[0118] The simulation feature data and welding standards are integrated into corresponding input data and input into the quality assessment model for analysis to obtain the quality assessment value of the corresponding weld.

[0119] A set of quality assessment values ​​corresponding to the simulation results of the summary scheme;

[0120] When the quality assessment value set contains a quality assessment value of 0, the simulation results of the assessment scheme do not meet the welding standards.

[0121] When there is no quality assessment value of 0 in the set of quality assessment values, the simulation results of the assessment scheme meet the welding standards.

[0122] In one embodiment, welding monitoring data is simulated using a digital twin model. Based on existing methods, the simulation is performed to determine welding parameters and then generate corresponding welding control commands.

[0123] For example, the digital twin model is adjusted according to the welding monitoring data, and the adjusted digital twin model is used to perform simulation analysis according to the target welding scheme to determine the welding parameters and then generate the corresponding welding control commands.

[0124] The equipment includes a monitoring module, a control and analysis module, and a welding control module;

[0125] The monitoring module is used to perform real-time monitoring during the welding process, obtain corresponding welding monitoring data, and send the welding monitoring data to the simulation module in the cloud.

[0126] In one embodiment, to improve analysis efficiency, a corresponding edge device can be set to preprocess the welding monitoring data and send the preprocessed welding monitoring data to the cloud.

[0127] The control and analysis module is used to control and analyze the welding process. It receives welding control commands sent from the cloud in real time, identifies the timestamps corresponding to the welding control commands, sets a time axis, and marks the corresponding command points (i.e., the corresponding time points) on the time axis according to the timestamps of each welding control command. It marks the command point closest to the current time on the time axis as the base point. It acquires the welding time point in real time, i.e., the time corresponding to the actual welding, and marks the welding time point closest to the current time on the time axis. The current time axis is marked as the time command axis.

[0128] Based on the time command axis, the timestamps of the welding control commands received in real time are analyzed to obtain the corresponding time analysis results. Based on the time analysis results, the corresponding control execution commands are generated and sent to the welding control module.

[0129] In one embodiment, timeliness analysis is performed on the timestamps of the welding control commands received in real time based on the time command axis, including:

[0130] Set the timeliness standard according to the user's welding requirements, that is, the acceptable instruction delay duration;

[0131] Establish a timeliness analysis model. The expression for the timeliness analysis model is as follows:

[0132] ;

[0133] In the formula: (TB, BS) are the input data, TB is the timeliness feature identified according to the time command axis, the timeliness feature includes the welding time point and the command timeliness value, the welding time point refers to the welding time point closest to the current time, and the command timeliness value is the time period between the base point and the current time; BS is the timeliness standard, TB→BS means that the corresponding timeliness feature meets the timeliness standard; the output data is the timeliness analysis value SA(TB, BS), the timeliness analysis value is 1 or 0; if the command timeliness value is not less than the timeliness standard or the base point is not greater than the welding time point, it is considered that the timeliness standard is not met. Among them, the base point is not greater than the welding time point, which means that the comparison is made in the order of the time axis, such as t=1 and 2. It can also be understood that the base point is the historical time corresponding to the welding time point.

[0134] Based on the time command axis, the corresponding timeliness characteristics are identified in real time, and the timeliness characteristics and timeliness standards are integrated into the input data and input into the timeliness analysis model for analysis to obtain the corresponding timeliness analysis value.

[0135] When the aging analysis value is 1, the aging analysis result is that the aging is qualified.

[0136] When the aging analysis value is 0, the corresponding instruction point in the time instruction axis is marked as a time delay instruction point, and the aging analysis result is that the aging is unqualified.

[0137] In one embodiment, corresponding control execution instructions are generated based on the timeliness analysis results, including:

[0138] When the timeliness analysis result is that the timeliness is qualified, the control execution instruction will be marked as a control execution instruction;

[0139] When the aging analysis result is that the aging is unqualified, the welding monitoring data is analyzed according to the quality assessment model to obtain the corresponding quality assessment value. When the quality assessment value is 1, the corresponding control execution command is generated according to the target welding plan, that is, the corresponding control execution command is executed according to the target welding plan. When the quality assessment value is 0, the preset emergency handling method is used, which is generally to stop welding. If other emergency strategies are set, the corresponding strategies are used.

[0140] The welding control module is used to perform welding control according to the received control execution instructions.

[0141] The above formulas are all numerical calculations after removing dimensions. The formulas are obtained by software simulation based on a large amount of data and are closest to the real situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained by simulation based on a large amount of data.

[0142] The above embodiments are only used to illustrate the technical methods 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 methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A digital-twin-based intelligent control system for machine tool welding, characterized by, The cloud end and the device end are included; The cloud end includes a simulation module; the device end includes a monitoring module, a control analysis module, and a welding control module; The simulation module establishes a corresponding digital twin model according to machine tool welding information, determines a target welding scheme based on the digital twin model, and sends the target welding scheme to the device end; Real-time welding monitoring data sent by the device end is acquired, and welding control instructions are obtained by simulating the welding monitoring data through the digital twin model; the welding control instructions are sent to the device end after being marked with corresponding time stamps; The monitoring module is used for real-time monitoring during welding to obtain corresponding welding monitoring data, and the welding monitoring data is sent to the simulation module of the cloud end; The control analysis module is used for control analysis of welding, real-time reception of welding control instructions sent by the cloud end, identification of time stamps corresponding to the welding control instructions, setting of a time axis, marking of corresponding instruction points on the time axis according to the time stamps of the welding control instructions, marking of the instruction point closest to the current time in the time axis as a base point, real-time acquisition of welding time points, corresponding marking of the welding time points closest to the current time in the time axis, and marking of the current time axis as a time instruction axis; Time-effect analysis is performed on the time stamps of the welding control instructions received in real time according to the time instruction axis, time-effect analysis results of the corresponding welding control instructions are obtained, control execution instructions are generated according to the time-effect analysis results, and the control execution instructions are sent to the welding control module; The welding control module is used for welding control according to the received control execution instructions.

2. The machine tool welding intelligent control system based on digital twinning according to claim 1, characterized in that, Determining a target welding scheme based on a digital twin model includes: Step SA1: acquiring welding target information and matching a corresponding initial welding scheme according to the welding target information; Step SA2: simulating the initial welding scheme through a digital twin model to obtain a scheme simulation result of the initial welding scheme; Step SA3: evaluating whether the scheme simulation result meets welding standards; When the evaluation scheme simulation result meets the welding standards, the initial welding scheme is sent to a user for confirmation, the initial welding scheme confirmed by the user is marked as a target welding scheme, and the analysis is ended; When the evaluation scheme simulation result does not meet the welding standards, the initial welding scheme is adjusted according to the scheme simulation result, a new initial welding scheme is obtained, and the process returns to step SA2.

3. The digital-twin-based intelligent control system for machine tool welding according to claim 2, characterized in that, Matching a corresponding initial welding scheme according to welding target information includes: Determining various potential welding targets that a user has according to machine tool welding information and user product information, acquiring various potential welding target information, determining various welding schemes according to the machine tool welding information and the potential welding target information, marking the welding schemes as reserve welding schemes of corresponding potential welding targets, and establishing a reserve library according to the reserve welding schemes corresponding to each potential welding target; Identifying welding target information, inputting the welding target information into the reserve library for matching, obtaining a plurality of reserve welding schemes, performing priority evaluation on the matched reserve welding schemes, and marking the reserve welding scheme with the highest priority as an initial welding scheme.

4. The machine tool welding intelligent control system based on digital twinning of claim 2, wherein, Evaluating whether a scheme simulation result meets welding standards includes: Obtaining welding standards, collecting features according to the welding standards, and obtaining simulation feature data of the corresponding welding position; Establishing a quality evaluation model, Integrating the simulation feature data and the welding standards into corresponding input data and inputting the input data into the quality evaluation model for analysis to obtain a quality evaluation value of the corresponding welding position; Summarizing the quality evaluation values into a quality evaluation value set of the simulation result; When the quality evaluation value set has a quality evaluation value of 0, the simulation result does not meet the welding standards; When the quality evaluation value set has no quality evaluation value of 0, the simulation result meets the welding standards.

5. The digital-twin-based intelligent control system for machine tool welding according to claim 4, characterized in that, The expression of the quality evaluation model is: ； In the formula: (s, BZ) is input data, s is simulation feature data, and BZ is a welding standard; s→BZ indicates that the corresponding simulation feature data meets the welding standard; the output data is a quality evaluation value ZP(s, BZ), and the quality evaluation value is 1 or 0.

6. The digital-twin-based intelligent control system for machine tool welding according to claim 1, wherein, According to the time instruction axis, the time stamp of the welding control instruction received in real time is subjected to time limit analysis, including: According to the user's welding requirements, a time limit standard is set; a time limit analysis model is established, and the expression of the time limit analysis model is: ； In the formula: (TB, BS) is input data, TB is a time limit feature identified according to the time instruction axis, the time limit feature includes a welding time point and an instruction time limit value, BS is a time limit standard, TB→BS indicates that the corresponding time limit feature meets the time limit standard; the output data is a time limit analysis value SA(TB, BS), and the time limit analysis value is 1 or 0; According to the time instruction axis, the corresponding time limit feature is identified in real time, the time limit feature and the time limit standard are integrated into input data, the input data is input into the time limit analysis model for analysis, and a time limit analysis value of the corresponding welding control instruction is obtained; When the time limit analysis value is 1, the time limit analysis result is time limit qualified; When the time limit analysis value is 0, the corresponding instruction point in the time instruction axis is marked as a time delay instruction point, and the time limit analysis result is time limit unqualified.

7. The digital-twin-based intelligent control system for a welding machine according to claim 5, wherein, According to the time limit analysis result, corresponding control execution instructions are generated, including: When the time limit analysis result is time limit qualified, the control execution instruction is marked as a control execution instruction; When the time limit analysis result is time limit unqualified, the welding monitoring data is analyzed according to the quality evaluation model to obtain a quality evaluation value; when the quality evaluation value is 1, a control execution instruction is generated according to the target welding scheme; and when the quality evaluation value is 0, a preset emergency treatment method is used for treatment.

Images