A spatial environment welding digital twin method based on multi-scale dynamic coupling

Abstract

This invention discloses a digital twin method for welding in a space environment based on multi-scale dynamic coupling, comprising: S1: collecting key data during the welding process and establishing a basic numerical model of the welding process and microstructure changes by combining experimental data; S2: constructing a digital twin model of the welding process using numerical simulation methods based on the basic numerical model; S3: establishing a real-time data interaction mechanism between the digital model and the actual welding equipment by utilizing the virtual-real synchronization model in the digital twin model; S4: automatically adjusting the welding process parameters through the intelligent optimization model in the digital twin model based on the feedback data obtained by the real-time data interaction mechanism, so that the microstructure changes in the welding process match the target prediction results; S5: establishing a welding process data management database to store key data and adjusted process parameters, and continuously training and optimizing the prediction accuracy of the digital twin model based on historical data in the database.

Description

Technical Field

[0001] This invention relates to the field of aerospace manufacturing technology, and in particular to a digital twin method for welding in a space environment based on multi-scale dynamic coupling. Background Technology

[0002] With the rapid development of aerospace technology and the continuous expansion of deep space exploration missions, the welding quality requirements for spacecraft structural components are becoming increasingly stringent. Welding operations in the space environment face extreme challenges not found in the terrestrial environment, including the coupled effects of multiple complex factors such as large-scale temperature cycling, microgravity conditions, and atomic oxygen corrosion. These factors have a significant impact on the thermal cycling, molten pool behavior, microstructure evolution, and joint performance of the welding process.

[0003] Traditional welding process parameter design and quality control methods are mainly based on ground environmental conditions, and their theoretical models and experimental data are difficult to directly apply to the space environment. Existing welding process modeling methods usually adopt single-scale analysis, which cannot effectively describe the multi-scale coupling mechanism between material phase transformation at the atomic scale and structural deformation at the macroscale, resulting in insufficient prediction accuracy of welding processes in the space environment.

[0004] Therefore, there is an urgent need for a digital twin method for welding in a space environment based on multi-scale dynamic coupling. Summary of the Invention

[0005] This invention provides a digital twin method for welding in a space environment based on multi-scale dynamic coupling, in order to solve the above-mentioned problems in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A digital twin method for welding in a space environment based on multi-scale dynamic coupling includes:

[0008] S1: Collect key data during the welding process and establish a basic numerical model of the welding process and microstructure changes by combining experimental data;

[0009] S2: Based on the basic numerical model, a digital twin model of the welding process is constructed using numerical simulation methods. The digital twin model integrates a dynamic model of the spatial environment, a multi-scale coupling model, an intelligent optimization model, and a virtual-real synchronization model.

[0010] S3: Utilize the virtual-real synchronization model in the digital twin model to establish a real-time data interaction mechanism between the digital model and the actual welding equipment;

[0011] S4: Based on the feedback data obtained through the real-time data interaction mechanism, the welding process parameters are automatically adjusted through the intelligent optimization model in the digital twin model so that the changes in the microstructure during the welding process match the target prediction results.

[0012] S5: Establish a welding process data management database to store key data and adjusted process parameters, and continuously train and optimize the prediction accuracy of the digital twin model based on historical data in the database.

[0013] Step S1 includes:

[0014] S11: Through a radiation-resistant temperature sensor array, a microgravity accelerometer, and an atomic oxygen mass spectrometer, key data on temperature field, microgravity disturbance, and material corrosion during the space welding process are collected in real time at a preset sampling frequency.

[0015] S12: A wavelet transform-Kalman filter hybrid algorithm is used to preprocess key data, eliminate interference, and align multi-source data;

[0016] S13: Extract three core parameters from the preprocessed key data: periodic temperature difference, microgravity disturbance curve, and material corrosion rate;

[0017] S14: Establish a basic numerical model of welding process and microstructure changes based on experimental data, including temperature difference sub-model, microgravity disturbance sub-model and atomic oxygen corrosion sub-model.

[0018] Step S2 includes:

[0019] S21: Based on the basic numerical model, a dynamic model of the space environment is constructed, integrating the temperature difference cycle sub-model, the microgravity disturbance sub-model, and the atomic oxygen corrosion sub-model to form a dynamic model framework describing the influence of the space environment on the welding process.

[0020] S22: Based on the dynamic model of the space environment, a multi-scale coupled model is constructed. Through thermo-mechanical coupling analysis, the multi-physics field coupling of temperature field, stress field and material microstructure evolution during the welding process is realized, and a coupling mechanism between macroscopic welding process and microstructure change is established.

[0021] S23: Integrate the multi-scale coupling model with the intelligent optimization model and the virtual-real synchronization model to form a digital twin model with real-time simulation, parameter optimization and virtual-real mapping functions.

[0022] Step S3 includes:

[0023] S31: Using a digital twin model, a data mapping relationship between the digital model and the actual welding equipment is established through a virtual-real synchronization model;

[0024] S32: Utilizes a multiphysics coupling mechanism to calculate temperature, stress, and corrosion field data in the digital model in real time, and transmits them to the actual welding equipment through a shared memory pool;

[0025] S33: The actual welding equipment adjusts the welding parameters based on the received data and feeds back the real-time welding data to the shared memory pool for use in updating the digital model, thus completing the establishment of a real-time data interaction mechanism.

[0026] Step S4 includes:

[0027] S41: The welding process parameters are dynamically adjusted based on real-time feedback data through the reinforcement learning algorithm in the intelligent optimization model.

[0028] S42: Use reward functions to evaluate the degree of matching between the microstructure changes during the welding process and the target prediction results, and guide the optimization and adjustment of process parameters;

[0029] S43: Uses convolutional neural networks to analyze thermal imaging data during the welding process, identify potential welding defects, and automatically correct the welding path and process parameters based on the identification results.

[0030] Step S5 includes:

[0031] S51: Establish a welding process data management database to store key data, adjusted process parameters, and microstructure change data during the welding process;

[0032] S52: Analyze and mine historical data in the database, construct a knowledge graph, and generate optimization suggestions;

[0033] S53: Based on historical data and optimization suggestions, continuously train and optimize the prediction accuracy of the digital twin model and the parameter adjustment strategy of the intelligent optimization model.

[0034] Step S22 includes:

[0035] S221: Establish a thermo-mechanical coupling model, use the finite element analysis method, and calculate the dynamic evolution of the temperature field and stress field during the welding process through the coupled equation set;

[0036] S222: Establish an austenite-martensite phase transformation model and predict the microstructure evolution based on the phase field method using phase transformation equations;

[0037] S223: Through thermo-mechanical-chemical coupling analysis, combined with preset temperature difference cycling conditions, real-time data exchange of temperature field, stress field, phase transformation field and corrosion field is realized, and a multi-scale coupling mechanism between macroscopic and microscopic dimensions is established.

[0038] Step S31 includes:

[0039] S311: Real-time data on extreme temperature difference cycling, microgravity disturbance, and atomic oxygen corrosion are acquired through a radiation-resistant temperature sensor array, a microgravity accelerometer, and an atomic oxygen mass spectrometer.

[0040] S312: Utilizing a data mapping engine, the digital model is synchronized with the actual welding equipment via the OPC UA communication protocol, transmitting thermo-mechanical coupling calculation results and phase change prediction data;

[0041] S313: The thermo-mechanical coupling optimization instructions and phase change control strategies in the digital twin model are transmitted to the actual welding equipment through the feedback optimization module, forming a closed-loop control that takes into account extreme space environments.

[0042] Step S14 includes:

[0043] S141: Establish an extreme temperature difference cycle sub-model and use an unsteady heat conduction equation to describe the heat conduction behavior of the welding material under a preset temperature difference cycle;

[0044] S142: Establish a microgravity perturbation sub-model and predict the influence of the microgravity environment on the weld pool morphology and austenite-martensite phase transformation dynamics through orbital dynamics equations;

[0045] S143: Establish an atomic oxygen corrosion sub-model, use reaction kinetic equations to simulate the corrosion effect of atomic oxygen on the surface of welding materials, and combine it with a phase transformation model to predict the impact of corrosion on microstructure evolution.

[0046] Step S12 includes:

[0047] S121: Wavelet transform is used to perform multi-scale denoising on key data containing extreme temperature difference cycles to eliminate sudden interference from solar flares;

[0048] S122: Kalman filtering is used to smooth the denoised temperature field, microgravity and corrosion data to ensure the accuracy of data for thermo-mechanical coupling calculations and phase transition predictions.

[0049] S123: Align multi-source heterogeneous data through spatiotemporal registration methods to establish a unified time reference to support synchronous calculations for thermo-mechanical coupling analysis and austenitic-martensite phase transformation simulation.

[0050] Compared with the prior art, the present invention has the following advantages:

[0051] A digital twin method for welding in a space environment based on multi-scale dynamic coupling includes: S1: collecting key data during the welding process and establishing a basic numerical model of the welding process and microstructure changes based on experimental data; S2: constructing a digital twin model of the welding process using numerical simulation methods based on the basic numerical model, integrating a space environment dynamic model, a multi-scale coupling model, an intelligent optimization model, and a virtual-real synchronization model; S3: establishing a real-time data interaction mechanism between the digital model and the actual welding equipment using the virtual-real synchronization model in the digital twin model; S4: automatically adjusting welding process parameters based on feedback data obtained from the real-time data interaction mechanism through the intelligent optimization model in the digital twin model, so that the microstructure changes during the welding process match the target prediction results; S5: establishing a welding process data management database to store key data and adjusted process parameters, and continuously training and optimizing the prediction accuracy of the digital twin model based on historical data in the database. This method fills the gap in existing technologies that cannot accurately simulate the coupling effect between the dynamic space environment (such as temperature difference cycling from -150℃ to 200℃) and material phase transformation.

[0052] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention.

[0053] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0054] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0055] Figure 1 This is a flowchart of a space environment welding digital twin method based on multi-scale dynamic coupling in an embodiment of the present invention;

[0056] Figure 2 This is a flowchart illustrating the basic numerical model for establishing the welding process and microstructure changes in this embodiment of the invention. Detailed Implementation

[0057] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0058] The embodiments of the present invention provide, as follows Figure 1 As shown, a digital twin method for welding in a space environment based on multi-scale dynamic coupling includes:

[0059] S1: Collect key data during the welding process and establish a basic numerical model of the welding process and microstructure changes by combining experimental data;

[0060] S2: Based on the basic numerical model, a digital twin model of the welding process is constructed using numerical simulation methods. The digital twin model integrates a dynamic model of the spatial environment, a multi-scale coupling model, an intelligent optimization model, and a virtual-real synchronization model.

[0061] S3: Utilize the virtual-real synchronization model in the digital twin model to establish a real-time data interaction mechanism between the digital model and the actual welding equipment;

[0062] S4: Based on the feedback data obtained through the real-time data interaction mechanism, the welding process parameters are automatically adjusted through the intelligent optimization model in the digital twin model so that the changes in the microstructure during the welding process match the target prediction results.

[0063] S5: Establish a welding process data management database to store key data and adjusted process parameters, and continuously train and optimize the prediction accuracy of the digital twin model based on historical data in the database.

[0064] The working principle of the above technical solution is as follows: S1: The system uses specialized equipment such as a radiation-resistant temperature sensor array, a microgravity accelerometer, and an atomic oxygen mass spectrometer to collect key parameters in the space welding environment in real time, with a sampling frequency of 100Hz, ensuring the accuracy of extreme temperature differences from -150℃ to 200℃ and 10 -6 Complete acquisition of data including microgravity perturbations was achieved. Among these, a radiation-resistant temperature sensor array is a temperature measurement device capable of operating normally in strong radiation environments; a microgravity accelerometer is used to measure minute gravity changes in near-zero gravity conditions; and an atomic oxygen mass spectrometer is used to detect the concentration of active atomic oxygen in the upper atmosphere. The acquired raw data underwent preprocessing using a wavelet transform-Kalman filter hybrid algorithm to eliminate sudden interference such as solar flares, and multi-source heterogeneous data were aligned using a spatiotemporal registration module, with time synchronization accuracy controlled within 1 μs. Based on the processed data, combined with experimental data and numerical simulation results, a fundamental numerical model was established to describe key physical phenomena such as temperature evolution, material microstructure changes, and stress distribution during the welding process.

[0065] S2: The system constructs a digital twin model comprising four core sub-models. The space environment dynamic model includes a temperature difference cycle sub-model, a microgravity disturbance sub-model, and an atomic oxygen corrosion sub-model, used to simulate the periodic temperature changes caused by day-night cycles during orbital operation, the impact of microgravity on material fluidity, and the oxidative corrosion of material surfaces by atomic oxygen, respectively. The multi-scale coupled model establishes a macroscopic-scale thermo-mechanical coupled model using finite element analysis to calculate the dynamic evolution of temperature and stress fields. Simultaneously, it uses phase field simulation to predict the microscopic-scale austenite-martensite phase transformation process, where austenite is a stable face-centered cubic crystal structure at high temperatures, and martensite is a body-centered cubic or body-centered tetragonal crystal structure formed by rapid cooling. The intelligent optimization model integrates reinforcement learning algorithms and convolutional neural networks. The reinforcement learning algorithm dynamically adjusts welding current, voltage, and speed parameters by designing a reward function, while the convolutional neural network analyzes infrared thermal images to identify potential welding defects. The virtual-real synchronization model establishes a real-time communication bridge between the digital model and physical devices through the OPC UA protocol. OPC UA is an industrial automation standard communication protocol that supports cross-platform device interconnection.

[0066] S3: The system utilizes the data mapping engine in the virtual-real synchronization model to achieve bidirectional real-time data interaction between the digital twin model and the actual welding equipment via the OPC UA protocol. The real-time data acquisition module obtains real-time operating parameters such as temperature, pressure, position, and current from various sensors of the physical welding equipment, with data transmission latency controlled at the millisecond level. The data mapping engine converts the measured data from the physical equipment into a standardized format recognizable by the digital model, while establishing a data quality assessment mechanism to automatically identify and eliminate abnormal data. The feedback optimization module transmits the optimization instructions calculated by the digital twin model to the control system of the welding equipment through the actuator interface, forming a closed-loop control circuit. The entire data interaction process ensures the consistency of the virtual and physical worlds through a spatiotemporal synchronization algorithm, achieving true "virtual-real synchronization."

[0067] S4: Based on real-time collected feedback data, the system automatically adjusts welding process parameters through intelligent optimization algorithms in a digital twin model, ensuring that the microstructure changes during the actual welding process match the target prediction results. Reinforcement learning algorithms dynamically optimize key parameters such as welding current, voltage, and welding speed according to the current welding state and environmental conditions. The algorithm's reward function comprehensively considers multiple quality indicators such as temperature deviation, residual stress, and welding defect rate. Convolutional neural networks analyze infrared thermal imaging images of the welding process in real time to identify potential defects such as porosity and cracks. Once an abnormal pattern is detected, the system automatically adjusts the welding path and process parameters for correction. A multi-scale coupled model synchronously calculates the macroscopic temperature field, stress field distribution, and microscopic phase transformation process, providing a comprehensive physical basis for parameter optimization. The entire optimization process employs an adaptive control strategy, dynamically adjusting the parameter weights of the optimization algorithm based on real-time welding quality evaluation results.

[0068] S5: The system establishes a dedicated welding process data management database, comprising three core components: an experimental data storage unit, a simulation result storage unit, and a knowledge graph construction module. The experimental data storage unit stores raw data from ground and space welding experiments, including sensor measurements, process parameter settings, and welding quality assessment results. The simulation result storage unit archives numerical simulation results, including simulation data such as temperature field evolution, stress distribution, and microstructure changes. The knowledge graph construction module analyzes the correlations between historical data using data mining techniques and automatically generates process optimization suggestions. The database adopts a distributed storage architecture, supporting efficient storage and retrieval of large-scale data. Based on the historical data accumulated in the database, the system uses machine learning algorithms to continuously train and optimize the prediction accuracy of the digital twin model, including updating neural network weights, adjusting physical model parameters, and refining optimization algorithm strategies. Through this continuous learning mechanism, the digital twin model can continuously adapt to new welding conditions and process requirements, improving prediction accuracy and optimization effectiveness.

[0069] The beneficial effects of the above technical solution are as follows: Welding digital twin technology, which considers a multi-scale coupling model during the welding process, can achieve intelligent evaluation and optimization of the welding process, improving welding quality and efficiency. This will bring significant technological breakthroughs to the development of the welding industry, promoting the development of welding processes towards intelligence and efficiency.

[0070] In another embodiment, such as Figure 2 As shown, step S1 includes:

[0071] S11: Through a radiation-resistant temperature sensor array, a microgravity accelerometer, and an atomic oxygen mass spectrometer, key data on temperature field, microgravity disturbance, and material corrosion during the space welding process are collected in real time at a preset sampling frequency.

[0072] S12: A wavelet transform-Kalman filter hybrid algorithm is used to preprocess key data, eliminate interference, and align multi-source data;

[0073] S13: Extract three core parameters from the preprocessed key data: periodic temperature difference, microgravity disturbance curve, and material corrosion rate;

[0074] S14: Establish a basic numerical model of welding process and microstructure changes based on experimental data, including temperature difference sub-model, microgravity disturbance sub-model and atomic oxygen corrosion sub-model.

[0075] The working principle of the above technical solution is as follows: S11: Key environmental data during the space welding process are collected in real time through a dedicated sensor array; among which, a radiation-resistant temperature sensor array is used to capture extreme temperature differences from -150℃ to 200℃, and a microgravity accelerometer monitors 10 -6 Gravitational disturbances at the g level are used to measure the degree of oxidation and corrosion on the material surface using an atomic oxygen mass spectrometer; the preset sampling frequency is 100Hz to ensure complete recording of the dynamic changes in the space environment.

[0076] S12: A wavelet transform-Kalman filter hybrid algorithm is used to preprocess key data; wavelet transform is used for multi-scale denoising to effectively eliminate the impact of sudden electromagnetic interference such as solar flares on sensor data; Kalman filtering is used to smooth the denoised temperature field, microgravity, and corrosion data to improve data accuracy; time synchronization of multi-source heterogeneous data is achieved through spatiotemporal registration, with synchronization accuracy controlled within 1μs;

[0077] S13: Extract core environmental parameters from the preprocessed key data; among them, the three types of core parameters extracted include: the periodic temperature difference function ΔT(t) reflects the influence of day-night alternation and orbital β angle changes, the microgravity perturbation curve G(t) describes the changes in the gravitational field caused by orbital dynamics, and the material corrosion rate O(t) quantifies the degree of erosion of the material surface by atomic oxygen; these parameters constitute the environmental characteristic tensor E(t) = [ΔT(t), G(t), O(t)];

[0078] S14: Based on experimental data, establish a basic numerical model containing three sub-models; among them, the temperature difference sub-model uses the following unsteady-state heat conduction equation to describe the heat conduction behavior under complex temperature difference cycles:

[0079]

[0080] Where ρ is the material density, and c p ρ is the specific heat capacity, k is the thermal conductivity, and Q is the specific heat capacity. r For radiative heat source items;

[0081] The microgravity perturbation sub-model predicts the impact of microgravity on the welding process using orbital dynamics equations:

[0082]

[0083] Where G is the gravitational constant, M is the mass of the Earth, R is the orbital radius, ω is the orbital angular velocity, and g0 is the standard gravitational acceleration;

[0084] The atomic oxygen corrosion sub-model uses reaction kinetic equations to establish the relationship between corrosion rate and environmental parameters:

[0085]

[0086] Where R c For corrosion rate, k c The reaction rate constant, F AO For atomic oxygen flux, E a R is the activation energy, R is the gas constant, and T is the temperature.

[0087] The beneficial effects of the above technical solution are: it can fill the gap in the existing technology that cannot accurately simulate the coupling effect between the dynamic space environment (such as the temperature difference cycle of -150℃ to 200℃) and the phase change of materials.

[0088] In another embodiment, step S2 includes:

[0089] S21: Based on the basic numerical model, a dynamic model of the space environment is constructed, integrating the temperature difference cycle sub-model, the microgravity disturbance sub-model, and the atomic oxygen corrosion sub-model to form a dynamic model framework describing the influence of the space environment on the welding process.

[0090] S22: Based on the dynamic model of the space environment, a multi-scale coupled model is constructed. Through thermo-mechanical coupling analysis, the multi-physics field coupling of temperature field, stress field and material microstructure evolution during the welding process is realized, and a coupling mechanism between macroscopic welding process and microstructure change is established.

[0091] S23: Integrate the multi-scale coupling model with the intelligent optimization model and the virtual-real synchronization model to form a digital twin model with real-time simulation, parameter optimization and virtual-real mapping functions.

[0092] The working principle of the above technical solution is as follows: S21: Construct a dynamic model of the space environment based on the basic numerical model; wherein, the temperature difference cycle sub-model, the microgravity disturbance sub-model and the atomic oxygen corrosion sub-model are systematically integrated to form a unified dynamic model framework of the space environment; this framework realizes data interaction between the sub-models through a shared memory pool, and can comprehensively describe the coupled influence of -150℃~200℃ temperature difference cycle, microgravity disturbance and atomic oxygen corrosion on the welding process;

[0093] S22: Based on the dynamic model of the space environment, a multi-scale coupled model is constructed, and multi-physics coupling is achieved through thermo-mechanical coupling analysis; among them, the multi-scale coupled model uses the finite element analysis method to take the welding heat source as the heat input to establish the thermo-mechanical coupling model:

[0094] [K T ]{T}={Q},[K S ]{σ}={F}

[0095] Where [K] T [K] represents the thermal conductivity stiffness matrix. S {} represents the structural stiffness matrix, {T} represents the temperature field, {σ} represents the stress field, {Q} represents the radiative heat source term, and {F} represents the external force;

[0096] Corrosion field data are obtained through the corrosion rate calculation formula:

[0097]

[0098] Where R c For corrosion rate, k c The reaction rate constant, F AO For atomic oxygen flux, E a To activate the energy; the calculation results are exchanged in real time through a shared memory pool to realize the exchange of temperature field, stress field and corrosion field. Each sub-model realizes data interaction through the shared memory pool and sets an adaptive adjustment strategy for coupling time step.

[0099] The coupling between the macroscopic welding process and the microscopic changes in microstructure is achieved through the phase-field method, establishing an austenite-martensite phase transformation model:

[0100]

[0101] in The phase field variable describes the material microstructure, M is the mobility, and F is the free energy function, including chemical free energy, elastic strain energy, and interface energy. The multiphysics coupling mechanism realizes real-time data exchange of temperature field, stress field, and corrosion field by setting an adaptive adjustment strategy for coupling time step, and predicts the dynamic changes of microstructure during welding, such as grain growth and phase transformation processes.

[0102] S23: Integrate the multi-scale coupled model with the intelligent optimization model and the virtual-real synchronization model into a system; the system integration achieves seamless connection between the models through standardized data interfaces and communication protocols; forming a complete digital twin model with real-time simulation, parameter optimization and virtual-real mapping functions; the model can respond to the state changes of the physical welding process in real time and provide predictive process optimization suggestions.

[0103] The beneficial effects of the above technical solution are as follows: the integrated digital twin model has real-time simulation, parameter optimization and virtual-real mapping functions, and can realize intelligent evaluation and optimized control of the welding process in complex space environment.

[0104] In another embodiment, step S3 includes:

[0105] S31: Using a digital twin model, a data mapping relationship between the digital model and the actual welding equipment is established through a virtual-real synchronization model;

[0106] S32: Utilizes a multiphysics coupling mechanism to calculate temperature, stress, and corrosion field data in the digital model in real time, and transmits them to the actual welding equipment through a shared memory pool;

[0107] S33: The actual welding equipment adjusts the welding parameters based on the received data and feeds back the real-time welding data to the shared memory pool for use in updating the digital model, thus completing the establishment of a real-time data interaction mechanism.

[0108] The working principle of the above technical solution is as follows: S31: Establish a data mapping relationship between the digital model and the actual welding equipment using a virtual-real synchronization model; wherein, the data mapping relationship is established through a data mapping engine, which defines the correspondence between virtual sensors in the digital space and actual sensors in the physical space; at the same time, establish a control mapping between virtual actuators and physical actuators to ensure that the calculation results of the digital model can be accurately transmitted to the physical equipment;

[0109] S32: Real-time calculation and transmission of field data is achieved using a multiphysics coupling mechanism. This mechanism solves for the distribution of temperature, stress, and corrosion fields in real time, and the results are transmitted to the actual welding equipment via a shared memory pool. The shared memory pool employs high-speed data buffering technology to ensure synchronous updates and fast access to multiphysics data.

[0110] S33: The actual welding equipment adjusts parameters based on the received data and feeds back real-time data; among which, the actual welding equipment automatically adjusts process parameters such as welding current, voltage, and speed based on the received temperature field and stress field data; at the same time, it feeds back the actual operating status of the equipment and sensor measurement data to the shared memory pool in real time for use in updating and correcting the digital model, forming a complete real-time data interaction closed loop.

[0111] The beneficial effects of the above technical solution are as follows: real-time welding data, including actual temperature distribution, stress changes and material microstructure evolution information, are fed back to a shared memory pool for use in updating the digital model, forming a closed-loop control system and realizing intelligent optimization of welding parameters.

[0112] In another embodiment, step S4 includes:

[0113] S41: The welding process parameters are dynamically adjusted based on real-time feedback data through the reinforcement learning algorithm in the intelligent optimization model.

[0114] S42: Use reward functions to evaluate the degree of matching between the microstructure changes during the welding process and the target prediction results, and guide the optimization and adjustment of process parameters;

[0115] S43: Uses convolutional neural networks to analyze thermal imaging data during the welding process, identify potential welding defects, and automatically correct the welding path and process parameters based on the identification results.

[0116] The working principle of the above technical solution is as follows: S41: The welding process parameters are dynamically adjusted through a reinforcement learning algorithm; wherein, the reinforcement learning algorithm takes the current welding state as input and outputs the adjustment amount of welding current, voltage and speed through a policy network; the algorithm learns the optimal parameter adjustment strategy through interaction with the environment and can adapt to the complex and ever-changing spatial welding environment.

[0117] S42: Utilize a reward function to evaluate the degree of matching between the welding process and the target; whereby, the reward function is designed as follows:

[0118] R = -(w1·ΔT + w2·σ) r +w3·D f )

[0119] Where ΔT is the temperature deviation, σ r For residual stress, D f The welding defect rate is represented by w1, w2, and w3, which are weighting parameters.

[0120] The reward function comprehensively considers three key indicators: temperature deviation, residual stress, and welding defect rate; the weight parameters w1, w2, and w3 are adjusted according to specific welding quality requirements; the calculation results of the reward function guide the strategy update of the reinforcement learning algorithm, thereby achieving continuous optimization of process parameters.

[0121] S43: The system uses a convolutional neural network to analyze thermal imaging data and automatically correct process parameters. The convolutional neural network performs real-time analysis of infrared thermal images during the welding process to identify potential welding defects such as porosity and cracks. Based on the defect identification results, the system automatically calculates the optimal welding path correction scheme and process parameter adjustment strategy to achieve adaptive control of the welding process.

[0122] The beneficial effects of the above technical solution are as follows: the execution of parameter adjustment aims to meet the welding quality requirements in the space environment. The execution results are then fed back into the reinforcement learning algorithm through the feedback optimization module to continuously optimize the decision-making strategy, ensuring that the organizational changes in the welding process are highly matched with the prediction results of the digital twin model, and ultimately realizing intelligent evaluation and dynamic optimization of welding quality in the space environment.

[0123] In another embodiment, step S5 includes:

[0124] S51: Establish a welding process data management database to store key data, adjusted process parameters, and microstructure change data during the welding process;

[0125] S52: Analyze and mine historical data in the database, construct a knowledge graph, and generate optimization suggestions;

[0126] S53: Based on historical data and optimization suggestions, continuously train and optimize the prediction accuracy of the digital twin model and the parameter adjustment strategy of the intelligent optimization model.

[0127] The working principle of the above technical solution is as follows: S51: Establish a comprehensive welding process data management database; wherein, the database adopts a layered storage architecture, including a real-time data caching layer, a historical data storage layer, and a knowledge data management layer; the stored content covers key environmental data, optimized process parameters, and microstructure change data during the welding process; the database supports high-concurrency access and fast retrieval of large amounts of data;

[0128] S52: Analyze and mine historical data in the database to construct a knowledge graph; among which, data mining algorithms are used to analyze the correlation between welding parameters, environmental conditions and welding quality; the knowledge graph construction module transforms the mined patterns into structured knowledge representations and generates process optimization suggestions for different welding conditions; the knowledge graph supports semantic query and reasoning, providing intelligent support for welding process decisions;

[0129] S53: Continuously train and optimize the digital twin model based on historical data; in this process, the prediction algorithm of the digital twin model is retrained using accumulated historical data to improve the model's prediction accuracy and generalization ability; at the same time, the parameter adjustment strategy in the intelligent optimization model is optimized so that it can better adapt to different welding scenarios; thus forming a data-driven model continuous improvement mechanism.

[0130] The beneficial effects of the above technical solution are as follows: the purpose of continuous optimization is to enable the digital twin model to more accurately predict the changes in the microstructure of the welding process in the space environment, and the intelligent optimization algorithm to more effectively adapt to the complex space environment and welding conditions.

[0131] In another embodiment, step S22 includes:

[0132] S221: Establish a thermo-mechanical coupling model, use the finite element analysis method, and calculate the dynamic evolution of the temperature field and stress field during the welding process through the coupled equation set;

[0133] S222: Establish an austenite-martensite phase transformation model and predict the microstructure evolution based on the phase field method using phase transformation equations;

[0134] S223: Through thermo-mechanical-chemical coupling analysis, combined with preset temperature difference cycling conditions, real-time data exchange of temperature field, stress field, phase transformation field and corrosion field is realized, and a multi-scale coupling mechanism between macroscopic and microscopic dimensions is established.

[0135] The working principle of the above technical solution is as follows: S221: Establish a thermo-mechanical coupling model for dynamic evolution calculation; wherein, the finite element analysis method is adopted to calculate the spatiotemporal distribution of temperature field and stress field during welding by solving the coupled heat conduction equation and structural mechanics equation; the coupled solution takes into account the change of material thermal properties with temperature and the reaction of thermal stress on temperature field;

[0136] S222: Establishing an austenite-martensite phase transformation model to predict microstructure evolution; among which, the phase transformation model is constructed based on the phase field method:

[0137]

[0138] Phase field variables It describes the microstructure of materials, where M is the mobility and F is the free energy function, including chemical free energy, elastic strain energy, and interfacial energy; it predicts microstructural changes such as grain growth and phase transformation during the welding process by solving the phase-field equation.

[0139] S223: A multi-scale coupling mechanism between macroscopic and microscopic dimensions is established through multi-physics field coupling analysis. In this mechanism, real-time data exchange between temperature field, stress field, phase transformation field and corrosion field is realized by combining preset temperature difference cycling conditions. The multi-scale coupling mechanism transmits macroscopic temperature and stress information to the microscopic phase transformation model through a scale bridging algorithm, while feeding back the influence of microstructure changes on material properties to the macroscopic model.

[0140] The beneficial effects of the above technical solution are as follows: by setting an adaptive adjustment strategy for the coupling time step, real-time data exchange is achieved between the temperature field from the thermo-mechanical coupling model in step one, the stress field from the structural mechanics calculation, the phase transformation field from the phase field model in step two, and the corrosion field from the atomic oxygen corrosion sub-model, thereby establishing a multi-scale dynamic coupling calculation framework from the macroscopic welding process to the microscopic microstructure evolution.

[0141] In another embodiment, step S31 includes:

[0142] S311: Real-time data on extreme temperature difference cycling, microgravity disturbance, and atomic oxygen corrosion are acquired through a radiation-resistant temperature sensor array, a microgravity accelerometer, and an atomic oxygen mass spectrometer.

[0143] S312: Utilizing a data mapping engine, the digital model is synchronized with the actual welding equipment via the OPC UA communication protocol, transmitting thermo-mechanical coupling calculation results and phase change prediction data;

[0144] S313: The thermo-mechanical coupling optimization instructions and phase change control strategies in the digital twin model are transmitted to the actual welding equipment through the feedback optimization module, forming a closed-loop control that takes into account extreme space environments.

[0145] The working principle of the above technical solution is as follows: S311: Real-time data of the space environment is acquired through dedicated sensors; among them, the radiation-resistant temperature sensor array measures extreme temperature difference cycle data, the microgravity accelerometer monitors gravity disturbance information, and the atomic oxygen mass spectrometer detects the corrosion status of the material surface; the sensor data is transmitted in real time to the data mapping engine for processing and format conversion;

[0146] S312: Utilizes a data mapping engine to achieve data synchronization between the digital model and the actual equipment; specifically, a standardized data exchange channel is established through the OPC UA communication protocol to transmit thermo-mechanical coupling calculation results and phase change prediction data; the data mapping engine is responsible for the conversion and synchronization between different data formats to ensure data consistency between the digital model and the physical equipment;

[0147] S313: A closed-loop control system considering extreme space environments is formed through a feedback optimization module. The feedback optimization module transmits the thermo-mechanical coupling optimization instructions and phase change control strategies calculated by the digital twin model to the actual welding equipment. After the equipment executes the optimization instructions, it feeds back the results to the digital model, forming a complete closed-loop control system to achieve adaptive control of extreme space environment conditions.

[0148] The beneficial effects of the above technical solution are as follows: closed-loop control continuously monitors the deviation between the actual welding effect and the prediction results of the digital model, and dynamically corrects process parameters such as welding current, voltage and speed, so as to ensure the stability and controllability of welding quality in extreme space environments.

[0149] In another embodiment, step S14 includes:

[0150] S141: Establish an extreme temperature difference cycle sub-model and use an unsteady heat conduction equation to describe the heat conduction behavior of the welding material under a preset temperature difference cycle;

[0151] S142: Establish a microgravity perturbation sub-model and predict the influence of the microgravity environment on the weld pool morphology and austenite-martensite phase transformation dynamics through orbital dynamics equations;

[0152] S143: Establish an atomic oxygen corrosion sub-model, use reaction kinetic equations to simulate the corrosion effect of atomic oxygen on the surface of welding materials, and combine it with a phase transformation model to predict the impact of corrosion on microstructure evolution.

[0153] The working principle of the above technical solution is as follows: S141: Establish an extreme temperature difference cycle sub-model to describe the heat conduction behavior;

[0154] The heat transfer process of the welding material under temperature difference cycling of -150℃ to 200℃ is described by an unsteady-state heat conduction equation:

[0155]

[0156] Where ρ is the material density, and c p ρ is the specific heat capacity, k is the thermal conductivity, and Q is the specific heat capacity. r For radiative heat source items;

[0157] Since there is no air conduction in the space environment, a radiation heat transfer equation is introduced to calculate heat exchange, namely:

[0158]

[0159] Where ε is the surface emissivity of the material, and σ is the Stefan-Boltzmann constant. The background temperature is used; the microgravity perturbation sub-model uses the following formula to calculate the orbital microgravity effect:

[0160] S142: Establishing a microgravity perturbation sub-model to predict the impact on the welding process; among which, the orbital dynamics equations are used to predict the impact on the welding process.

[0161]

[0162] Where G is the gravitational constant, M is the mass of the Earth, R is the orbital radius, ω is the orbital angular velocity, and g0 is the standard gravitational acceleration;

[0163] The model predicts the effects of microgravity on weld pool morphology, material flow, and austenite-martensite phase transformation kinetics.

[0164] S143: Establish an atomic oxygen corrosion sub-model to simulate corrosion; the atomic oxygen corrosion sub-model uses the following chemical reaction kinetic equation to describe the corrosion process:

[0165]

[0166] Where R c For corrosion rate, k c The reaction rate constant, F AO For atomic oxygen flux, E a The three sub-models, with activation energy R, gas constant R, and temperature T, interact in real time through a shared memory pool, forming a complete numerical simulation foundation for welding processes in a space environment. Combined with a phase transition model, the influence of corrosion on material microstructure evolution is predicted, providing corrosion field data for multiphysics coupled analysis.

[0167] The beneficial effects of the above technical solution are as follows: the model combines phase transformation model to analyze the influence of corrosion on the evolution of material microstructure, predicts the corrosion rate and oxide layer thickness changes, and provides accurate corrosion field data input for multiphysics coupling analysis.

[0168] In another embodiment, step S12 includes:

[0169] S121: Wavelet transform is used to perform multi-scale denoising on key data containing extreme temperature difference cycles to eliminate sudden interference from solar flares;

[0170] S122: Kalman filtering is used to smooth the denoised temperature field, microgravity and corrosion data to ensure the accuracy of data for thermo-mechanical coupling calculations and phase transition predictions.

[0171] S123: Align multi-source heterogeneous data through spatiotemporal registration methods to establish a unified time reference to support synchronous calculations for thermo-mechanical coupling analysis and austenitic-martensite phase transformation simulation.

[0172] The working principle of the above technical solution is as follows: S121: Wavelet transform is used for multi-scale denoising processing; among which, the key data mainly comes from the real-time acquisition of radiation-resistant temperature sensor array, microgravity accelerometer and atomic oxygen mass spectrometer, with a sampling frequency of 100Hz, covering extreme temperature difference cycle data of -150℃~200℃, microgravity disturbance data of 10^-6g level and atomic oxygen corrosion related data; wavelet transform is a multi-resolution time-frequency analysis method that can decompose signals at different time scales and frequency scales, effectively identify and separate noise interference of different frequency components; solar flares are explosive phenomena that occur on the surface of the sun, which will generate strong electromagnetic radiation and particle streams, causing sudden interference to electronic equipment and sensors in the space environment. Through the multi-scale decomposition characteristics of wavelet transform, the high-frequency sudden interference signal generated by solar flares can be separated from the low-frequency periodic signal of temperature difference cycle, ensuring the accuracy of environmental parameters such as temperature difference function ΔT(t), microgravity curve G(t) and corrosion rate O(t);

[0173] S122: Data smoothing is performed using Kalman filtering; Kalman filtering is a recursive filtering algorithm based on a state-space model, which uses a prediction-correction mechanism to optimally estimate the state of a dynamic system, effectively handling noisy measurement data; the temperature field data is based on the heat conduction-radiation coupling equation:

[0174]

[0175] To describe the thermal conduction behavior of welding materials, where ρ is the material density and c is the thermal conductivity. p ρ is the specific heat capacity, k is the thermal conductivity, and Q is the specific heat capacity. rFor the radiative heat source term, since there is no air conduction in the space environment, a radiative heat transfer equation is introduced to calculate the heat exchange, namely:

[0176]

[0177] Where ε is the surface emissivity of the material, and σ is the Stefan-Boltzmann constant. The background temperature is used; the microgravity perturbation sub-model uses the following formula to calculate the orbital microgravity effect:

[0178]

[0179] Where G is the gravitational constant, M is the Earth's mass, R is the orbital radius, ω is the orbital angular velocity, and g0 is the standard gravitational acceleration; the atomic oxygen corrosion sub-model uses the following chemical reaction kinetic equations to describe the corrosion process: Corrosion data are based on the surface chemical reaction kinetic model:

[0180]

[0181] Where R c For corrosion rate, k c The reaction rate constant, F AO For atomic oxygen flux, E a R is the activation energy, T is the gas constant, and T is the temperature. The three sub-models achieve real-time data interaction through a shared memory pool, forming a complete numerical simulation foundation for the welding process in a space environment. A Kalman filter-based state estimation and error covariance update mechanism ensure the thermo-mechanical coupling calculation model.

[0182] [K T ]{T}={Q},[K S ]{σ}={F}

[0183] Where [K] T [K] represents the thermal conductivity stiffness matrix. S Let {{}} be the structural stiffness matrix, {T} be the temperature field, {σ} be the stress field, {Q} be the radiative heat source term, and {F} be the external force; the coupling between the macroscopic welding process and the microstructure changes is achieved through the phase field method, establishing an austenite-martensite phase transformation model:

[0184]

[0185] in The phase field variable describes the material microstructure, M is the mobility, and F is the free energy function, including chemical free energy, elastic strain energy, and interface energy. The multiphysics coupling mechanism realizes real-time data exchange of temperature field, stress field, and corrosion field by setting an adaptive adjustment strategy for coupling time step, and predicts the dynamic changes of microstructure during welding, such as grain growth and phase transformation processes.

[0186] S123: Multi-source data alignment is achieved through spatiotemporal registration. Spatiotemporal registration refers to the technique of accurately aligning heterogeneous data from different sensors, sampling times, and spatial locations in both time and space. The heterogeneous data includes thermal field data from temperature sensor arrays, microgravity data from accelerometers, and atomic oxygen data from mass spectrometers, which have different sampling frequencies, data formats, and timestamps. The registration method employs time synchronization technology based on feature matching and interpolation algorithms to unify the timestamps of data from each sensor to the same reference time system, with synchronization accuracy controlled within 1 μs. The establishment of a unified time reference ensures the temporal consistency of each component in the environmental feature tensor E(t) = [ΔT(t), G(t), O(t)].

[0187] The beneficial effects of the above technical solution are: it provides a synchronous calculation basis for subsequent thermo-mechanical coupling analysis and microstructure phase transformation simulation, enabling temperature field, stress field and corrosion field to be coupled and solved at the same time point, realizing real-time data exchange and collaborative calculation of multi-physics fields.

[0188] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from the spirit and scope of this invention.

Claims

1. A multi-scale dynamic coupling based spatial environment welding digital twin method, characterized in that, include: S1: Collect key data during the welding process and establish a basic numerical model of the welding process and microstructure changes by combining experimental data; S2: Based on the basic numerical model, a digital twin model of the welding process is constructed using numerical simulation methods. The digital twin model integrates a dynamic model of the spatial environment, a multi-scale coupling model, an intelligent optimization model, and a virtual-real synchronization model. S3: Utilize the virtual-real synchronization model in the digital twin model to establish a real-time data interaction mechanism between the digital model and the actual welding equipment; S4: Based on the feedback data obtained through the real-time data interaction mechanism, the welding process parameters are automatically adjusted through the intelligent optimization model in the digital twin model so that the changes in the microstructure during the welding process match the target prediction results. S5: Establish a welding process data management database to store key data and adjusted process parameters, and continuously train and optimize the prediction accuracy of the digital twin model based on historical data in the database; Step S1 includes: S11: Through a radiation-resistant temperature sensor array, a microgravity accelerometer, and an atomic oxygen mass spectrometer, key data on temperature field, microgravity disturbance, and material corrosion during the space welding process are collected in real time at a preset sampling frequency. S12: A wavelet transform-Kalman filter hybrid algorithm is used to preprocess key data, eliminate interference, and align multi-source data; S13: Extract three core parameters from the preprocessed key data: periodic temperature difference, microgravity disturbance curve, and material corrosion rate; S14: Establish a basic numerical model of welding process and microstructure changes based on experimental data, including temperature difference sub-model, microgravity disturbance sub-model and atomic oxygen corrosion sub-model.

2. The multi-scale dynamic coupling based spatial environment welding digital twin method of claim 1, wherein, Step S2 includes: S21: Based on the basic numerical model, a dynamic model of the space environment is constructed, integrating the temperature difference cycle sub-model, the microgravity disturbance sub-model, and the atomic oxygen corrosion sub-model to form a dynamic model framework describing the influence of the space environment on the welding process. S22: Based on the dynamic model of the space environment, a multi-scale coupled model is constructed. Through thermo-mechanical coupling analysis, the multi-physics field coupling of temperature field, stress field and material microstructure evolution during the welding process is realized, and a coupling mechanism between macroscopic welding process and microstructure change is established. S23: Integrate the multi-scale coupling model with the intelligent optimization model and the virtual-real synchronization model to form a digital twin model with real-time simulation, parameter optimization and virtual-real mapping functions.

3. The spatial environment welding digital twin method based on multi-scale dynamic coupling according to claim 1, characterized in that, Step S3 includes: S31: Using a digital twin model, a data mapping relationship between the digital model and the actual welding equipment is established through a virtual-real synchronization model; S32: Utilizes a multiphysics coupling mechanism to calculate temperature, stress, and corrosion field data in the digital model in real time, and transmits them to the actual welding equipment through a shared memory pool; S33: The actual welding equipment adjusts the welding parameters based on the received data and feeds back the real-time welding data to the shared memory pool for use in updating the digital model, thus completing the establishment of a real-time data interaction mechanism.

4. The spatial environment welding digital twin method based on multi-scale dynamic coupling according to claim 1, characterized in that, Step S4 includes: S41: The welding process parameters are dynamically adjusted based on real-time feedback data through the reinforcement learning algorithm in the intelligent optimization model. S42: Use reward functions to evaluate the degree of matching between the microstructure changes during the welding process and the target prediction results, and guide the optimization and adjustment of process parameters; S43: Uses convolutional neural networks to analyze thermal imaging data during the welding process, identify potential welding defects, and automatically correct welding paths and process parameters based on the identification results.

5. The method for welding digital twins in a space environment based on multi-scale dynamic coupling according to claim 1, characterized in that, The S5 steps include: S51: Establish a welding process data management database to store key data, adjusted process parameters, and microstructure change data during the welding process; S52: Analyze and mine historical data in the database, construct a knowledge graph, and generate optimization suggestions; S53: Based on historical data and optimization suggestions, continuously train and optimize the prediction accuracy of the digital twin model and the parameter adjustment strategy of the intelligent optimization model.

6. The spatial environment welding digital twin method based on multi-scale dynamic coupling according to claim 2, characterized in that, Step S22 includes: S221: Establish a thermo-mechanical coupling model, use the finite element analysis method, and calculate the dynamic evolution of the temperature field and stress field during the welding process through the coupled equation set; S222: Establish an austenite-martensite phase transformation model and predict the microstructure evolution based on the phase field method using phase transformation equations; S223: Through thermo-mechanical-chemical coupling analysis, combined with preset temperature difference cycling conditions, real-time data exchange of temperature field, stress field, phase transformation field and corrosion field is realized, and a multi-scale coupling mechanism between macroscopic and microscopic dimensions is established.

7. The spatial environment welding digital twin method based on multi-scale dynamic coupling according to claim 3, characterized in that, Step S31 includes: S311: Real-time data on extreme temperature difference cycling, microgravity disturbance, and atomic oxygen corrosion are acquired through a radiation-resistant temperature sensor array, a microgravity accelerometer, and an atomic oxygen mass spectrometer. S312: Utilizing a data mapping engine, the digital model and the actual welding equipment are synchronized via the OPC UA communication protocol, transmitting thermo-mechanical coupling calculation results and phase change prediction data; S313: The thermo-mechanical coupling optimization instructions and phase change control strategies in the digital twin model are transmitted to the actual welding equipment through the feedback optimization module, forming a closed-loop control that takes into account extreme space environments.

8. The method for welding digital twins in a space environment based on multi-scale dynamic coupling according to claim 1, characterized in that, Step S14 includes: S141: Establish an extreme temperature difference cycle sub-model and use an unsteady heat conduction equation to describe the heat conduction behavior of the welding material under a preset temperature difference cycle; S142: Establish a microgravity perturbation sub-model and predict the influence of the microgravity environment on the weld pool morphology and austenite-martensite phase transformation dynamics through orbital dynamics equations; S143: Establish an atomic oxygen corrosion sub-model, use reaction kinetic equations to simulate the corrosion effect of atomic oxygen on the surface of welding materials, and combine it with a phase transformation model to predict the impact of corrosion on microstructure evolution.

9. The method for welding digital twins in a space environment based on multi-scale dynamic coupling according to claim 1, characterized in that, Step S12 includes: S121: Wavelet transform is used to perform multi-scale denoising on key data containing extreme temperature difference cycles to eliminate sudden interference from solar flares; S122: Kalman filtering is used to smooth the denoised temperature field, microgravity and corrosion data to ensure the accuracy of data for thermo-mechanical coupling calculations and phase transition predictions. S123: Align multi-source heterogeneous data through spatiotemporal registration methods to establish a unified time reference to support synchronous calculations for thermo-mechanical coupling analysis and austenitic-martensite phase transformation simulation.

Images