Laser additive remanufacturing control method and system based on digital twinning

Abstract

The invention discloses a laser additive remanufacturing management and control method and system based on digital twinning, and relates to the technical field of data processing. The method comprises the steps that a laser additive remanufacturing associated data set is collected; building an additive remanufacturing twinborn task list, carrying out twinborn simulation integrated analysis on the laser additive remanufacturing associated data set, and building a laser additive remanufacturing digital twinborn body; performing simulation prediction to obtain part performance prediction parameters; and carrying out process strategy regulation and control on the part performance prediction parameters, determining target part process optimization parameters, and carrying out laser additive remanufacturing optimization management and control. The technical problem that in the prior art, the laser additive remanufacturing process lacks real-time process regulation and control, working condition changes are difficult to respond in time, and consequently the remanufacturing quality is unstable is solved, real-time mapping and prediction deduction are conducted on the whole additive remanufacturing process through digital twinning, self-adaptive optimization and adjustment of process parameters are achieved, and the laser additive remanufacturing quality is improved. The technical effect of improving the remanufacturing quality stability is achieved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a method and system for controlling laser additive remanufacturing based on digital twins. Background Technology

[0002] Laser additive remanufacturing, as a technology for repairing and restoring the performance of high-value metal parts, boasts advantages such as high processing precision, high material utilization, and the ability to reconstruct complex structures, and has been widely applied in aerospace, energy equipment, and other fields. However, the additive remanufacturing process inherently involves multi-physics coupling behaviors such as material cladding, heat input, stress evolution, and microstructure changes. The operating conditions are highly sensitive to process parameters; even slight changes can lead to forming defects, molten pool instability, or fluctuations in microstructure and properties. Existing processes mostly rely on experience-based parameter setting and post-processing adjustments based on offline detection results. This lack of real-time perception of the remanufacturing process and prediction of part performance evolution results in lag in process control, an inability to respond promptly to dynamic changes in operating conditions, and difficulty in ensuring stable and consistent remanufacturing quality. Summary of the Invention

[0003] This application provides a method and system for controlling laser additive remanufacturing based on digital twins, which solves the technical problem of unstable remanufacturing quality caused by the lack of real-time process control and difficulty in responding to changes in working conditions in the existing laser additive remanufacturing process.

[0004] The first aspect of this application provides a method for controlling laser additive remanufacturing based on digital twins, the method comprising:

[0005] A laser additive remanufacturing (LAR) associated dataset is collected, including material performance data, process parameter data, and equipment characteristic data. Based on the LAR RRA control objectives, a LAR RRA twin task list is constructed. The LAR RRA associated dataset is then subjected to twin simulation integration analysis based on the LAR RRA task list to build a LAR RRA digital twin. Real-time data streams of the part remanufacturing process are collected and mapped to the LAR RRA digital twin for simulation and prediction to obtain part performance prediction parameters. Process strategy adjustments are applied to these part performance prediction parameters to determine the target part's process optimization parameters. Based on these target part process optimization parameters, LAR RRA optimization control is implemented.

[0006] A second aspect of this application provides a digital twin-based laser additive remanufacturing control system, the system comprising:

[0007] Data Acquisition Module: Acquires a laser additive remanufacturing related dataset, which includes material performance data, process parameter data, and equipment characteristic data. Analysis Module: Based on the laser additive remanufacturing control objectives, constructs a laser additive remanufacturing twin task list, performs twin simulation integrated analysis on the laser additive remanufacturing related dataset based on the twin task list, and builds a laser additive remanufacturing digital twin. Simulation Prediction Module: Acquires real-time data streams of the part remanufacturing process, maps these data streams to the laser additive remanufacturing digital twin for simulation prediction, and obtains part performance prediction parameters. Control Module: Adjusts the process strategy based on the part performance prediction parameters, determines the target part's process optimization parameters, and performs laser additive remanufacturing optimization control based on these target part process optimization parameters.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] First, a laser additive remanufacturing (LAR) related dataset is collected, including material performance data, process parameter data, and equipment characteristic data. Next, based on the LAR remanufacturing control objectives, a LAR remanufacturing twin task list is constructed. Based on this task list, a twin simulation integration analysis is performed on the LAR remanufacturing related dataset to build a digital twin of the LAR remanufacturing process. Then, the remanufacturing process data stream is collected in real time and mapped to the LAR remanufacturing digital twin for simulation and prediction, yielding predicted performance parameters for the parts. Finally, process strategy adjustments are made based on the predicted performance parameters to determine the optimal process parameters for the target parts. Based on these optimal parameters, optimized control of the LAR remanufacturing process is implemented. This approach solves the technical problem of unstable remanufacturing quality caused by the lack of real-time process control and difficulty in responding to changes in operating conditions in existing LAR remanufacturing technologies. By using a digital twin to map and predict the entire LAR remanufacturing process in real time, adaptive optimization and adjustment of process parameters are achieved, resulting in improved remanufacturing quality stability. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0011] Figure 1 A schematic diagram of the process for controlling laser additive remanufacturing based on digital twins, provided for an embodiment of this application;

[0012] Figure 2 A schematic diagram of the structure of a laser additive remanufacturing control system based on digital twins provided in this application embodiment.

[0013] Figure labeling: Data acquisition module 11, analysis module 12, simulation and prediction module 13, control module 14. Detailed Implementation

[0014] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structure, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0015] Example 1, as Figure 1 As shown, this application provides a method for controlling laser additive remanufacturing based on digital twins, wherein the method includes:

[0016] Collect a laser additive remanufacturing related dataset, which includes material property data, process parameter data, and equipment characteristic data.

[0017] In this embodiment, a multi-dimensional data acquisition channel is established for the material properties, processing technology, and operating status of the additive manufacturing equipment for the parts to be remanufactured. Material performance data includes the chemical composition, particle size distribution, thermophysical parameters (such as thermal conductivity, specific heat capacity, and melting point), and mechanical performance parameters (such as yield strength, elastic modulus, and hardness) of the powder or filament raw materials, imported from raw material testing equipment or a material database. Process parameter data includes laser power, scanning speed, powder feed rate, layer thickness, scanning spacing, preheating temperature, and their time-varying records, acquired in real-time by the additive manufacturing equipment control system and external sensing modules. Equipment characteristic data includes laser output stability, spot size, energy density distribution, positioning accuracy of the motion platform, conveying efficiency of the powder feeding system, and flow characteristics of the gas protection system, obtained through the equipment's built-in status monitoring module and external calibration device. The aforementioned material performance data, process parameter data, and equipment characteristic data are timestamped and encoded according to a unified format to form a laser additive remanufacturing associated dataset, providing fundamental data support for subsequent digital twin modeling, process simulation prediction, and process optimization and control.

[0018] Based on the control objectives of laser additive remanufacturing, a twin task list for additive remanufacturing is constructed. Based on the twin task list, a twin simulation integration analysis is performed on the laser additive remanufacturing related dataset to build a digital twin of laser additive remanufacturing.

[0019] In this embodiment, after collecting the associated dataset for laser additive remanufacturing, an additive remanufacturing twin task list is constructed to drive digital twin modeling, based on the remanufacturing requirements of the target part and the overall additive process control objectives. The additive remanufacturing control objectives include forming quality stability objectives, geometric accuracy restoration objectives, microstructure performance improvement objectives, and process efficiency optimization objectives. According to different control objectives, the additive remanufacturing process is decomposed into several executable twin task units, such as equipment state simulation tasks, process parameter sensitivity analysis tasks, melt pool behavior prediction tasks, temperature field-stress field-flow field multiphysics coupling simulation tasks, and forming layer and overall part performance prediction tasks, thus forming the additive remanufacturing twin task list. Subsequently, driven by the additive remanufacturing twin task list, a twin simulation integration analysis was performed on the aforementioned laser additive remanufacturing related dataset. This included: formatting, converting, aligning, and mapping material performance data, process parameter data, and equipment characteristic data according to the input requirements of each twin task; calling physical models, data-driven models, or hybrid models corresponding to the task type to perform operating condition simulation, behavior prediction, and process response analysis; and performing unified model-driven coupling and data fusion processing on the simulation results of each twin task to generate multi-dimensional, multi-temporal digital twin ecosystem mapping data. Based on the above integration analysis process, the equipment 3D model, physical integration model, process rule model, and task model repository were consistently coupled to construct a laser additive remanufacturing digital twin with real-time response, iterative deduction, and dynamic updating capabilities.

[0020] Furthermore, building a digital twin for laser additive remanufacturing includes:

[0021] Based on the laser additive remanufacturing associated dataset, geometric and physical models are reconstructed to obtain a 3D model of the laser additive remanufacturing equipment and a laser additive remanufacturing physical integrated model. Process rules are mined from the laser additive remanufacturing associated dataset to establish a laser additive remanufacturing process rule model. Historical data is crawled and task models are trained based on the additive remanufacturing twin task list to obtain an additive remanufacturing task model repository. The 3D model of the laser additive remanufacturing equipment and the laser additive remanufacturing physical integrated model, as well as the laser additive remanufacturing process rule model and the additive remanufacturing task model repository, are coupled to drive the construction of a laser additive remanufacturing digital twin.

[0022] First, based on the laser additive remanufacturing associated dataset, structural information and geometric parameters of the additive manufacturing equipment are extracted and analyzed. Three-dimensional modeling is then performed using the equipment's structural features, motion unit information, and optical system parameters to establish an initial 3D model of the equipment. Simultaneously, multiphysics simulations of temperature, stress, and flow fields are performed, combining material performance data and process parameter data to obtain temperature field models, stress evolution models, and molten pool flow models describing the additive manufacturing behavior. In operational simulations, the physical models are driven by measured data for correction and parameter calibration, forming a laser additive remanufacturing physical integrated model that accurately reflects the dynamic characteristics of the additive manufacturing process. Subsequently, feature extraction, labeling, and process behavior identification are performed on the laser additive remanufacturing associated dataset. A process rule mining algorithm is used to generate a set of process rules reflecting the relationship between process settings, process response, and forming results. These rules are then filtered and dynamically updated based on their confidence level and applicability to construct a laser additive remanufacturing process rule model to support process decision-making. Simultaneously, based on the additive remanufacturing twin task list, data samples related to each twin task are automatically crawled from historical production records and additive experiment databases. Data-driven training is then performed using the corresponding model structure selected according to the task type, resulting in task sub-models such as weld bead shape prediction models, molten pool behavior prediction models, and process sensitivity analysis models, forming an additive remanufacturing task model repository. Finally, the equipment 3D model, physical integration model, process rule model, and task model repository are coupled according to the driving relationships required by the twin tasks. The inputs and outputs of geometric structures, physical behaviors, process rules, and task models are uniformly mapped. A core operating framework for the digital twin is established through data flow, event flow, and model-driven relationships, thereby constructing a laser additive remanufacturing digital twin.

[0023] Furthermore, the three-dimensional model of the laser additive remanufacturing equipment and the physical integration model of laser additive remanufacturing are obtained, including:

[0024] Equipment data is extracted and 3D reconstructed from the laser additive remanufacturing associated dataset to establish an initial 3D model of the equipment. A mesh size is preset according to the model's accuracy requirements, and the initial 3D model of the equipment is then subjected to mesh lightweighting processing according to the mesh size to obtain a 3D model of the laser additive remanufacturing equipment. Physical model simulation is performed based on the laser additive remanufacturing associated dataset to construct additive remanufacturing temperature field model, additive remanufacturing stress field model, and additive remanufacturing flow field model. Multiphysics coupling and simulation-driven correction are applied to the additive remanufacturing temperature field model, additive remanufacturing stress field model, and additive remanufacturing flow field model to obtain a physical integrated model of laser additive remanufacturing.

[0025] First, based on the laser additive remanufacturing associated dataset, data related to the equipment is extracted and processed, including equipment structural parameters, optical system parameters, motion axis parameters, powder feeding channel structural information, and protective gas flow channel information. By analyzing the geometric dimensions, component arrangement relationships, and motion chain characteristics of the equipment structure, a 3D reconstruction of the equipment structure is performed using 3D modeling tools, resulting in an initial 3D model of the equipment including optical components, motion platform, nozzle components, and powder feeding system. To ensure the efficiency of subsequent simulation and coupled calculations, a preset mesh size is used according to the model's accuracy requirements. The initial 3D model of the equipment is then meshed and lightweighted, including iso-meshing of complex surfaces, local mesh refinement of detailed areas, and simplification of non-critical structures, thereby generating a 3D model of the laser additive remanufacturing equipment that meets computational requirements. Subsequently, material property data, process parameter data, and equipment operation data from the laser additive remanufacturing associated dataset are used to simulate and analyze the physical behavior of the additive process. Specifically, a temperature field model for additive remanufacturing is constructed using heat conduction equations, heat source models, and boundary conditions. A stress field model is built based on temperature field distribution and material thermal stress behavior to describe the stress evolution of the cladding layer and heat-affected zone. A flow field model is constructed by combining molten pool dynamics, powder flow transport characteristics, and gas protection flow patterns to reflect the flow behavior inside and around the molten pool. Furthermore, a multiphysics coupling method is used to jointly solve the temperature field, stress field, and flow field, achieving a comprehensive simulation of the thermo-mechanical-fluid coupling process. Real-time data collected during the remanufacturing process drives the correction of the simulation results, including updating boundary conditions, calibrating heat source parameters, correcting flow viscosity, and adjusting thermophysical parameters, ensuring consistency between the simulation results and actual operating conditions. This results in a physically integrated laser additive remanufacturing model that truly reflects the dynamic characteristics of the additive process.

[0026] Furthermore, a rule model for laser additive remanufacturing processes is established, including:

[0027] Feature extraction and labeling are performed on the laser additive remanufacturing associated dataset to obtain a laser additive remanufacturing process feature set; process rule mining and confidence evaluation are performed on the laser additive remanufacturing process feature set to obtain a laser additive remanufacturing process rule set and a process rule confidence set; based on the process rule confidence set, the laser additive remanufacturing process rule set is filtered and dynamically updated to establish a laser additive remanufacturing process rule model.

[0028] First, feature extraction and labeling are performed on the laser additive remanufacturing associated dataset. This includes extracting thermophysical parameters, mechanical property indicators, and surface state parameters from material property data; extracting key process variables such as laser power, scanning speed, powder feed rate, layer thickness, scanning spacing, and gas flow rate from process parameter data; and extracting equipment behavior characteristics such as spot energy distribution, motion axis stability, and powder feed system response characteristics from equipment characteristic data. Combined with existing forming quality records, the geometric accuracy of the formed layer, melt pool stability, surface defects, and microstructure performance indicators during the additive manufacturing process are manually or automatically labeled, thus forming a laser additive remanufacturing process feature set that can be used for rule analysis. Subsequently, a process rule mining algorithm is used to analyze the process feature set, including mining the correlation between feature variables, the influencing factors of process results, and parameter combination patterns. A preliminary process rule set is generated through the statistical relationships or pattern associations between different process features and forming results. For example, the optimal melt pool width range under different combinations of laser power and scanning speed, the influence of preheating temperature on thermal stress distribution, and the linkage between powder feed rate and forming stability can be mined. Simultaneously, a confidence index is calculated for each process rule, including rule support, scope of application, and prediction consistency, to form a process rule confidence set. Finally, based on the process rule confidence set, the initially generated process rule set is screened and dynamically updated. Rules with low confidence, insufficient coverage, or significant stability deviations in new data are removed or corrected, and new effective rules are automatically introduced based on the latest additive manufacturing process data, achieving dynamic evolution of the rule set. Through the above rule screening, updating, and integration, a laser additive remanufacturing process rule model is constructed.

[0029] Furthermore, obtain a repository of additive remanufacturing task models, including:

[0030] Historical data is crawled based on the additive remanufacturing twin task list to obtain an additive remanufacturing twin task associated dataset; a twin task model structure set is selected according to the additive remanufacturing twin task list; task models are trained and labeled and integrated on the additive remanufacturing twin task associated dataset based on the twin task model structure set to obtain an additive remanufacturing task model repository.

[0031] First, based on the different types of twin tasks defined in the additive remanufacturing twin task list, such as melt pool behavior prediction, forming defect identification, process parameter sensitivity analysis, temperature field trend prediction, stress evolution trend prediction, and part performance prediction, corresponding historical additive manufacturing data is automatically crawled from historical production record databases, experimental databases, and equipment operation logs. Data cleaning, formatting, label alignment, and time-series rearrangement are then performed to form an additive remanufacturing twin task association dataset that corresponds one-to-one with each twin task. Subsequently, an appropriate twin task model structure set is selected for each type of task according to the twin task list. For example, a morphology regression model structure is selected for the melt pool morphology prediction task, a sequence prediction model structure is selected for the temperature and stress field trend analysis task, a multi-input multi-output parameter influence analysis model structure is selected for the process parameter sensitivity analysis task, and a performance mapping model structure based on historical forming data is selected for the part performance prediction task. Finally, based on the twin task model structure set, the additive remanufacturing twin task associated dataset is trained and processed, including feature extraction, network parameter update, model error calibration, and label mapping integration. The trained task models are classified and managed according to task type, and the model weights, input feature requirements, output prediction dimensions, and applicable conditions are uniformly identified and stored to form an additive remanufacturing task model repository that can be called by the digital twin.

[0032] Real-time acquisition of part remanufacturing process data stream, mapping of the part remanufacturing process data stream to the laser additive remanufacturing digital twin for simulation and prediction, and obtaining part performance prediction parameters.

[0033] In this embodiment, the system acquires data streams from the part remanufacturing process in real time. Specifically, multiple types of sensing modules are arranged at key locations in the additive remanufacturing equipment, including infrared sensors for acquiring molten pool temperature and surface radiation energy, a visual imaging module for monitoring the surface contour of the forming layer, a laser output monitoring module for acquiring laser spot energy fluctuations, a powder conveying sensor for acquiring powder feed rate changes, and displacement and velocity sensors for acquiring the equipment's motion status. Simultaneously, real-time process parameters such as laser power, scanning speed, layer thickness, powder feed rate, and protective gas flow rate are acquired through the equipment control system. The aforementioned multi-source sensor data are synchronized according to timestamps to form a continuous data stream, serving as the real-time driving input for the digital twin. The part remanufacturing process data stream is input into the constructed laser additive remanufacturing digital twin, and data mapping is performed through the digital twin's internal geometric model, physical integration model, process rule model, and task model repository. The mapping process includes: spatial mapping of real-time data to the equipment's 3D model; semantic mapping of process parameters to the process rule model; numerical mapping of temperature, stress, and flow-related sensor data to the physical integration model; and triggering the corresponding digital twin task model to perform predictive inference. Through real-time simulation of molten pool behavior, temperature field changes, stress evolution, and forming layer quality trends in virtual space using a digital twin, part performance prediction parameters are obtained, including geometric accuracy prediction parameters, surface defect prediction parameters, microstructure and performance prediction parameters, and mechanical property prediction parameters.

[0034] The performance prediction parameters of the parts are adjusted by process strategy to determine the process optimization parameters of the target parts, and the laser additive remanufacturing optimization and control is carried out based on the process optimization parameters of the target parts.

[0035] In this embodiment, after the digital twin generates part performance prediction parameters, the system adjusts the process strategy based on the prediction results to achieve adaptive optimization of the remanufacturing process. Specifically, the part performance prediction parameters are first compared and analyzed with the additive remanufacturing process-part performance mapping rule library. By identifying risk characteristics that may lead to geometric accuracy deviations, molten pool instability, insufficient interlayer bonding, excessive residual stress, or abnormal microstructure under the current process conditions, the corresponding process control direction is determined. For example, adjusting the laser power to stabilize the molten pool temperature, optimizing the scanning speed to improve the uniformity of the forming layer, or correcting the powder feeding amount to ensure stable deposition of the cladding layer. Subsequently, based on the process rule model and task model library, strategy comparison analysis is performed on each control direction to obtain feasible target part process control strategies. After determining the target control strategy, the system optimizes the performance prediction parameters of the part according to the strategy. This includes generating several candidate parameter sets for the process strategy within the allowable range of the strategy parameters. A digital twin then performs process effect simulation and performance prediction evaluation on each candidate parameter set, comparing the predicted melt pool stability index, temperature field distribution, stress evolution trend, and expected forming quality. The optimal result that meets the optimization target of additive remanufacturing is selected, thus obtaining the optimized process parameters for the target part. Finally, the system sends the optimized process parameters of the target part to the additive manufacturing equipment, adjusting key process variables such as laser power, scanning speed, powder feed rate, layer thickness setting, and protective gas flow rate in real time. This achieves optimized control of the laser additive remanufacturing process, ensuring stable forming quality and performance output under dynamic operating conditions.

[0036] Furthermore, determining the process optimization parameters for the target part includes:

[0037] Construct an additive remanufacturing process-part performance mapping rule base; perform control strategy analysis on the part performance prediction parameters based on the additive remanufacturing process-part performance mapping rule base to obtain the target part process control strategy; use the target part process control strategy to perform strategy optimization analysis on the part performance prediction parameters to determine the target part process optimization parameters.

[0038] The additive remanufacturing process-part performance mapping rule library is generated based on the established process rule model and historical additive remanufacturing data. It is used to describe the influence of different combinations of process parameters on key performance indicators such as part geometric accuracy, forming quality, melt pool stability, thermal stress distribution and microstructure properties. The rules are classified and structured according to the parameter sensitivity and performance response characteristics, forming a set of mapping rules that can be used for rapid reasoning and strategy generation.

[0039] After the additive remanufacturing process-part performance mapping rule base is constructed, the system inputs the part performance prediction parameters generated by the digital twin into the mapping rule base for control strategy analysis. Through rule matching, it identifies key process variables affecting the part performance prediction results, such as the influence of scanning speed on the molten pool width, the effect of powder feed rate on the stability of the formed layer thickness, and the ability of laser power changes to regulate temperature field balance. Based on the deviation of the prediction results from the target performance, the system automatically generates corresponding target part process control strategies, such as reducing the scanning speed to increase the molten pool aspect ratio, increasing the powder feed rate to improve the formed layer density, or adjusting the laser power to balance heat input. The system then uses the target part process control strategies to perform strategy optimization analysis on the part performance prediction parameters. Specifically, within the range of process parameters defined by the target control strategy, multiple candidate combinations of process parameters are generated. A digital twin is used to perform simulation predictions on each candidate combination to evaluate the corresponding molten pool behavior, temperature field and stress field variation trends, and forming layer quality indicators. Based on the additive remanufacturing optimization objectives (such as maximizing forming uniformity, minimizing residual stress, minimizing defect risk, etc.), the effects of each candidate parameter are optimized and screened to determine the process parameter combination that can achieve the best prediction performance, which is then used as the final process optimization parameters for the target part.

[0040] Furthermore, the target part process control strategy is used to perform strategy optimization analysis on the performance prediction parameters of the part, and the target part process optimization parameters are determined, including:

[0041] The target part process control strategy is used to analyze the performance prediction parameters of the part by strategy threshold analysis to obtain the part process strategy parameter threshold; multiple part process strategy parameters are randomly selected from the part process strategy parameter threshold; the multiple part process strategy parameters are evaluated and optimized based on the laser additive remanufacturing digital twin to determine the target part process optimization parameters.

[0042] First, based on the process adjustment direction indicated by the target part's process control strategy, the system performs strategy threshold analysis on the part's performance prediction parameters. This involves combining the additive remanufacturing process-part performance mapping rule base to calculate the allowable variation range of the current process variables within the performance target range. This yields threshold values ​​for multiple process control dimensions, including laser power, scanning speed, powder feed rate, layer thickness, scanning path, and protective gas flow rate. These threshold values ​​limit the search space for adjustable parameters, ensuring that the subsequent optimization process operates within a practically executable range. Next, within the range defined by the aforementioned strategy parameter thresholds, the system generates multiple combinations of part process strategy parameters using random sampling. The sampling process can employ uniform randomness, distributed perturbation randomness, or multi-round perturbation randomness to cover different process control possibilities, ensuring that the optimization process is both broad and maintains the effectiveness of the process parameters. The generated multiple candidate process strategy parameter combinations constitute a candidate control parameter set, serving as input for subsequent digital twin model evaluation. Further, the multiple part process strategy parameters are input into the laser additive remanufacturing digital twin, and simulation evaluation is performed through the twin's internal physical integration model, process rule model, and task model repository. The evaluation includes molten pool temperature distribution, molten pool morphology stability, temperature and stress field variation trends, formed layer surface quality, interlayer bonding effect, and final part performance prediction indicators. The system compares and analyzes the simulation results of each candidate strategy based on the set additive remanufacturing optimization objectives (such as maximizing forming quality stability, minimizing residual stress, minimizing defect risk, or improving performance consistency), and selects the combination of process parameters with the best overall effect as the final process optimization parameters for the target part.

[0043] Furthermore, based on the laser additive remanufacturing digital twin, the process strategy parameters of the multiple parts are evaluated and optimized to determine the process optimization parameters for the target part, including:

[0044] Based on the optimization objectives of additive remanufacturing, a set of evaluation indicators for part process effects is constructed; the process strategy parameters of multiple parts are simulated and evaluated based on the laser additive remanufacturing digital twin according to the set of evaluation indicators to obtain the effects of multiple part process parameters; the process strategy parameters of multiple parts are expanded and optimized based on the effects of the multiple part process parameters to determine the process optimization parameters of the target part.

[0045] First, a set of evaluation indicators for part process effects is constructed based on the optimization goals of additive remanufacturing. These optimization goals include multiple dimensions such as the geometric accuracy and stability of the forming layer, the consistency of the molten pool morphology, the interlayer bonding strength, the balance of temperature and stress fields, defect risk control, and the consistency of final part performance. Correspondingly, the evaluation indicator set may include indicators such as molten pool aspect ratio stability, forming layer thickness fluctuation, temperature gradient balance, residual stress distribution uniformity, surface defect probability, and microstructure performance prediction, used to measure the merits of candidate process strategies from multiple perspectives. Subsequently, based on the part process effect evaluation indicator set, the system inputs multiple randomly generated part process strategy parameters into a laser additive remanufacturing digital twin for effect simulation and evaluation. The digital twin, based on its internal equipment 3D model, physical integration model, process rule model, and task model repository, performs molten pool behavior simulation, temperature and stress field calculation, forming layer quality prediction, and performance response deduction for each set of process parameters, obtaining the corresponding process parameter effect results. Further, the system expands and optimizes multiple part process strategy parameters based on the effects of multiple part process parameters, that is, it performs optimal expansion processing on candidate process parameters based on the evaluation results. The expansion process includes: generating locally expanded parameter sets within the corresponding parameter ranges for high-performing process parameter combinations to broaden the search scope for potential optimal solutions; simultaneously, eliminating parameter combinations with poor evaluation results or that violate process constraints, forming a strategy expansion population centered on an efficient candidate set. Finally, the expanded strategy parameters are simulated and evaluated again using a digital twin, and the parameter combination with the best overall performance is selected as the final process optimization parameters for the target part based on the evaluation index results.

[0046] Furthermore, based on the effects of the process parameters of the multiple parts, the process strategy parameters of the multiple parts are expanded and optimized to determine the process optimization parameters of the target part, including:

[0047] Based on the effects of the process parameters of the multiple parts, the process strategy parameters of the multiple parts are optimized and expanded to generate a population of process strategy parameters for the parts; the laser additive remanufacturing digital twin is used to perform global evaluation and optimization on the population of process strategy parameters for the parts to determine the process optimization parameters for the target parts.

[0048] Based on the performance of multiple part process parameters obtained from the previous simulation, the system optimizes process strategy parameters that perform well and meet process stability requirements. Specifically, the system ranks candidate parameter combinations according to performance indicators such as melt pool behavior stability, temperature and stress field balance, forming layer quality, and defect risk, selecting process parameters within the optimization threshold as the basis for expansion. On this basis, the system performs local expansion on these optimized parameters in conjunction with their variation trends, generating refined parameter points within the allowable range of each parameter dimension. Through small perturbations, a new set of part process strategy parameter populations is formed, making the expanded parameter population more densely cover the potential optimal region while maintaining consistency with process constraints. Subsequently, the generated part process strategy parameter population is input into the laser additive remanufacturing digital twin for global evaluation and optimization. Based on its internal multiphysics coupling model, process rule model, and task model repository, the digital twin performs global simulation calculations on each set of process parameters in the population, including melt pool morphology evolution simulation, joint solution of temperature and stress fields, prediction of forming layer surface and morphology quality, and deduction of the final part performance response. The system comprehensively evaluates the simulation results based on the additive remanufacturing optimization objectives, and takes the process parameters with high forming quality stability, low performance prediction deviation, minimal defect risk, and best overall effect as the final result, thereby determining the process optimization parameters for the target part and realizing the global optimal control of the additive remanufacturing process.

[0049] In summary, the embodiments of this application have at least the following technical effects:

[0050] First, a laser additive remanufacturing (LAR) related dataset is collected, including material performance data, process parameter data, and equipment characteristic data. Next, based on the LAR remanufacturing control objectives, a LAR remanufacturing twin task list is constructed. Based on this task list, a twin simulation integration analysis is performed on the LAR remanufacturing related dataset to build a digital twin of the LAR remanufacturing process. Then, the remanufacturing process data stream is collected in real time and mapped to the LAR remanufacturing digital twin for simulation and prediction, yielding predicted performance parameters for the parts. Finally, process strategy adjustments are made based on the predicted performance parameters to determine the optimal process parameters for the target parts. Based on these optimal parameters, optimized control of the LAR remanufacturing process is implemented. This approach solves the technical problem of unstable remanufacturing quality caused by the lack of real-time process control and difficulty in responding to changes in operating conditions in existing LAR remanufacturing technologies. By using a digital twin to map and predict the entire LAR remanufacturing process in real time, adaptive optimization and adjustment of process parameters are achieved, resulting in improved remanufacturing quality stability.

[0051] Example 2, based on the same inventive concept as the laser additive remanufacturing control method based on digital twins in the foregoing examples, such as... Figure 2 As shown, this application provides a laser additive remanufacturing control system based on digital twins, wherein the system includes:

[0052] Data acquisition module 11: Acquires a laser additive remanufacturing related dataset, which includes material performance data, process parameter data, and equipment characteristic data; Analysis module 12: Constructs a laser additive remanufacturing twin task list based on the laser additive remanufacturing control objectives, performs twin simulation integrated analysis on the laser additive remanufacturing related dataset based on the laser additive remanufacturing twin task list, and builds a laser additive remanufacturing digital twin; Simulation prediction module 13: Acquires the part remanufacturing process data stream in real time, maps the part remanufacturing process data stream to the laser additive remanufacturing digital twin for simulation prediction, and obtains part performance prediction parameters; Control module 14: Adjusts the process strategy based on the part performance prediction parameters, determines the target part process optimization parameters, and performs laser additive remanufacturing optimization control based on the target part process optimization parameters.

[0053] Furthermore, the analysis module 12 is used to perform the following methods:

[0054] Based on the laser additive remanufacturing associated dataset, geometric and physical models are reconstructed to obtain a 3D model of the laser additive remanufacturing equipment and a laser additive remanufacturing physical integrated model. Process rules are mined from the laser additive remanufacturing associated dataset to establish a laser additive remanufacturing process rule model. Historical data is crawled and task models are trained based on the additive remanufacturing twin task list to obtain an additive remanufacturing task model repository. The 3D model of the laser additive remanufacturing equipment and the laser additive remanufacturing physical integrated model, as well as the laser additive remanufacturing process rule model and the additive remanufacturing task model repository, are coupled to drive the construction of a laser additive remanufacturing digital twin.

[0055] Furthermore, the analysis module 12 is used to perform the following methods:

[0056] Equipment data is extracted and 3D reconstructed from the laser additive remanufacturing associated dataset to establish an initial 3D model of the equipment. A mesh size is preset according to the model's accuracy requirements, and the initial 3D model of the equipment is then subjected to mesh lightweighting processing according to the mesh size to obtain a 3D model of the laser additive remanufacturing equipment. Physical model simulation is performed based on the laser additive remanufacturing associated dataset to construct additive remanufacturing temperature field model, additive remanufacturing stress field model, and additive remanufacturing flow field model. Multiphysics coupling and simulation-driven correction are applied to the additive remanufacturing temperature field model, additive remanufacturing stress field model, and additive remanufacturing flow field model to obtain a physical integrated model of laser additive remanufacturing.

[0057] Furthermore, the analysis module 12 is used to perform the following methods:

[0058] Feature extraction and labeling are performed on the laser additive remanufacturing associated dataset to obtain a laser additive remanufacturing process feature set; process rule mining and confidence evaluation are performed on the laser additive remanufacturing process feature set to obtain a laser additive remanufacturing process rule set and a process rule confidence set; based on the process rule confidence set, the laser additive remanufacturing process rule set is filtered and dynamically updated to establish a laser additive remanufacturing process rule model.

[0059] Furthermore, the analysis module 12 is used to perform the following methods:

[0060] Historical data is crawled based on the additive remanufacturing twin task list to obtain an additive remanufacturing twin task associated dataset; a twin task model structure set is selected according to the additive remanufacturing twin task list; task models are trained and labeled and integrated on the additive remanufacturing twin task associated dataset based on the twin task model structure set to obtain an additive remanufacturing task model repository.

[0061] Furthermore, the control module 14 is used to perform the following methods:

[0062] Construct an additive remanufacturing process-part performance mapping rule base; perform control strategy analysis on the part performance prediction parameters based on the additive remanufacturing process-part performance mapping rule base to obtain the target part process control strategy; use the target part process control strategy to perform strategy optimization analysis on the part performance prediction parameters to determine the target part process optimization parameters.

[0063] Furthermore, the control module 14 is used to perform the following methods:

[0064] The target part process control strategy is used to analyze the performance prediction parameters of the part by strategy threshold analysis to obtain the part process strategy parameter threshold; multiple part process strategy parameters are randomly selected from the part process strategy parameter threshold; the multiple part process strategy parameters are evaluated and optimized based on the laser additive remanufacturing digital twin to determine the target part process optimization parameters.

[0065] Furthermore, the control module 14 is used to perform the following methods:

[0066] Based on the optimization objectives of additive remanufacturing, a set of evaluation indicators for part process effects is constructed; the process strategy parameters of multiple parts are simulated and evaluated based on the laser additive remanufacturing digital twin according to the set of evaluation indicators to obtain the effects of multiple part process parameters; the process strategy parameters of multiple parts are expanded and optimized based on the effects of the multiple part process parameters to determine the process optimization parameters of the target part.

[0067] Furthermore, the control module 14 is used to perform the following methods:

[0068] Based on the effects of the process parameters of the multiple parts, the process strategy parameters of the multiple parts are optimized and expanded to generate a population of process strategy parameters for the parts; the laser additive remanufacturing digital twin is used to perform global evaluation and optimization on the population of process strategy parameters for the parts to determine the process optimization parameters for the target parts.

[0069] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A laser additive remanufacturing control method based on digital twins, characterized in that, The method includes: Collect a laser additive remanufacturing related dataset, which includes material property data, process parameter data, and equipment characteristic data; Based on the control objectives of laser additive remanufacturing, a twin task list for additive remanufacturing is constructed. Based on the twin task list, a twin simulation integration analysis is performed on the laser additive remanufacturing related dataset to build a digital twin of laser additive remanufacturing. Real-time acquisition of part remanufacturing process data stream, mapping of the part remanufacturing process data stream to the laser additive remanufacturing digital twin for simulation and prediction, and obtaining part performance prediction parameters; The performance prediction parameters of the parts are adjusted by process strategy to determine the process optimization parameters of the target parts, and the laser additive remanufacturing optimization and control is carried out based on the process optimization parameters of the target parts.

2. The laser additive remanufacturing control method based on digital twin as described in claim 1, characterized in that, Building a digital twin for laser additive remanufacturing includes: Based on the aforementioned laser additive remanufacturing associated dataset, geometric and physical models are reconstructed to obtain a three-dimensional model of the laser additive remanufacturing equipment and a physical integrated model of laser additive remanufacturing. Process rules are mined from the laser additive remanufacturing association dataset to establish a laser additive remanufacturing process rule model; Based on the aforementioned additive remanufacturing twin task list, historical data is crawled and task model training is performed to obtain an additive remanufacturing task model repository. By driving and coupling the 3D model of the laser additive remanufacturing equipment and the physical integration model of laser additive remanufacturing, as well as the laser additive remanufacturing process rule model and the additive remanufacturing task model repository, a digital twin of laser additive remanufacturing is built.

3. The laser additive remanufacturing control method based on digital twin as described in claim 2, characterized in that, The three-dimensional model of the laser additive remanufacturing equipment and the physical integration model of laser additive remanufacturing were obtained, including: Equipment data is extracted and 3D reconstruction is performed on the laser additive remanufacturing associated dataset to establish an initial 3D model of the equipment; Based on the model accuracy requirements, a mesh size is preset, and the initial three-dimensional model of the equipment is subjected to mesh lightweighting processing according to the mesh size to obtain a three-dimensional model of the laser additive remanufacturing equipment. Based on the aforementioned laser additive remanufacturing associated dataset, physical model simulations were performed to construct additive remanufacturing temperature field model, additive remanufacturing stress field model, and additive remanufacturing flow field model. Multiphysics coupling and simulation-driven correction are performed on the additive remanufacturing temperature field model, additive remanufacturing stress field model, and additive remanufacturing flow field model to obtain the physical integrated model of laser additive remanufacturing.

4. The laser additive remanufacturing control method based on digital twin as described in claim 2, characterized in that, Establish a rule model for laser additive remanufacturing processes, including: Feature extraction and labeling are performed on the laser additive remanufacturing associated dataset to obtain a laser additive remanufacturing process feature set; The laser additive remanufacturing process feature set is subjected to process rule mining and confidence evaluation to obtain the laser additive remanufacturing process rule set and process rule confidence set; Based on the confidence set of the process rules, the laser additive remanufacturing process rule set is filtered and dynamically updated to establish a laser additive remanufacturing process rule model.

5. The laser additive remanufacturing control method based on digital twin as described in claim 2, characterized in that, Obtain the additive remanufacturing task model repository, including: Based on the aforementioned list of additive remanufacturing twin tasks, historical data is crawled to obtain a dataset associated with additive remanufacturing twin tasks. Based on the additive remanufacturing twin task list, select the twin task model structure set; Based on the twin task model structure set, the additive remanufacturing twin task association dataset is used to train task models and integrate them to obtain an additive remanufacturing task model repository.

6. The laser additive remanufacturing control method based on digital twin as described in claim 1, characterized in that, Determine the process optimization parameters for the target part, including: Construct a rule base for mapping additive remanufacturing process to part performance; Based on the additive remanufacturing process-part performance mapping rule base, the control strategy analysis of the part performance prediction parameters is performed to obtain the target part process control strategy. The target part process control strategy is used to perform strategy optimization analysis on the performance prediction parameters of the part, and the process optimization parameters of the target part are determined.

7. The laser additive remanufacturing control method based on digital twin as described in claim 6, characterized in that, The target part's process control strategy is used to perform strategy optimization analysis on the performance prediction parameters of the part, and the target part's process optimization parameters are determined, including: The target part process control strategy is used to analyze the performance prediction parameters of the part by strategy threshold analysis to obtain the part process strategy parameter threshold. Randomly select multiple part process strategy parameters from the threshold values ​​of the part process strategy parameters; Based on the laser additive remanufacturing digital twin, the process strategy parameters of the multiple parts are evaluated and optimized to determine the process optimization parameters of the target part.

8. The laser additive remanufacturing control method based on digital twin as described in claim 7, characterized in that, Based on the laser additive remanufacturing digital twin, the process strategy parameters of the multiple parts are evaluated and optimized to determine the process optimization parameters of the target part, including: Based on the optimization goals of additive remanufacturing, a set of evaluation indicators for part process effects is constructed. Based on the laser additive remanufacturing digital twin, the process strategy parameters of the multiple parts are simulated and evaluated according to the set of evaluation indicators for the process effect of the parts, so as to obtain the effect of the process parameters of the multiple parts. Based on the effects of the process parameters of the multiple parts, the process strategy parameters of the multiple parts are expanded and optimized to determine the process optimization parameters of the target part.

9. The laser additive remanufacturing control method based on digital twin as described in claim 8, characterized in that, Based on the effects of the process parameters of the multiple parts, the process strategy parameters of the multiple parts are expanded and optimized to determine the process optimization parameters of the target part, including: Based on the effects of the process parameters of the multiple parts, the process strategy parameters of the multiple parts are optimized and expanded to generate a population of process strategy parameters for the parts. The laser additive remanufacturing digital twin is used to perform a global evaluation and optimization of the population of process strategy parameters for the part, and to determine the process optimization parameters for the target part.

10. A laser additive remanufacturing control system based on digital twins, characterized in that, For implementing the laser additive remanufacturing control method based on digital twins as described in any one of claims 1-9, the system comprises: Data acquisition module: Acquires laser additive remanufacturing related datasets, which include material property data, process parameter data, and equipment characteristic data; Analysis module: Based on the control objectives of laser additive remanufacturing, construct a twin task list for additive remanufacturing, perform twin simulation integration analysis on the laser additive remanufacturing related dataset based on the twin task list, and build a digital twin of laser additive remanufacturing. Simulation and prediction module: Real-time acquisition of data stream of part remanufacturing process, mapping of the data stream of part remanufacturing process to the laser additive remanufacturing digital twin for simulation and prediction, and obtaining part performance prediction parameters; Control module: Performs process strategy adjustment on the performance prediction parameters of the part, determines the process optimization parameters of the target part, and performs laser additive remanufacturing optimization control based on the process optimization parameters of the target part.

Images