Rotating spindle digital twin fusion modeling and dynamic optimization method and system

Abstract

The invention provides a rotating main shaft digital twinning fusion modeling and dynamic optimization method and system, and relates to the technical field of intelligent manufacturing and digital twinning, and the method comprises the steps: obtaining multi-source data of a numerical control machine tool main shaft system, including manufacturing parameters, working condition data and physical state data; a digital twinning mechanism model representing the mechanical-thermal-dynamic multi-field coupling effect is constructed, and a structured digital twinning working condition model is constructed through feature extraction and classification; fusing the digital twinning mechanism model and the digital twinning working condition model to form a dynamic fusion digital twinning model responding to variable working conditions; co-simulation and optimization are carried out, and weak design links of the main shaft system under variable working conditions are identified; on the basis of the identified weak design links, a multi-objective optimization model taking improvement of fatigue life and machining quality as objectives is constructed, and solving is carried out; and performing dynamic performance simulation verification on the optimized design parameters through a dynamic fusion digital twinborn model, and outputting an optimal design parameter set.

Description

Technical Field

[0001] This invention relates to the fields of intelligent manufacturing and digital twin technology, and in particular to a method and system for digital twin fusion modeling and dynamic optimization of a rotating spindle. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.

[0003] As the "mother machine" of modern manufacturing, the performance of the spindle system, the core functional component of CNC machine tools, directly determines the machining accuracy, efficiency, and reliability of the entire machine. In actual operation, the spindle system is constantly under complex working conditions of varying speed and load. At the same time, its performance is also affected by uncertainties such as geometric tolerances and assembly errors generated during the manufacturing stage. This results in its performance evolution exhibiting multi-temporal and spatial scales and strong nonlinear characteristics, which poses a huge challenge to its high-precision design and reliability optimization.

[0004] Digital twin technology provides a new paradigm for the design, operation, and optimization of complex equipment by constructing a bidirectional dynamic mapping between physical entities and virtual models. Theoretically, a high-fidelity digital twin model of a spindle system should be able to deeply integrate its multiphysics mechanisms, real-time operating conditions, and historical manufacturing information, thereby achieving accurate performance prediction and optimization. However, existing digital twin modeling methods still have significant limitations when facing the aforementioned challenges: First, existing models lack a deep fusion mechanism for multi-source heterogeneous information. Mechanistic models, real-time operating data, and manufacturing information are often isolated from each other, resulting in a "data silo" problem. Traditional modeling methods are mostly static parametric models, which are difficult to construct cross-domain mapping relationships from "geometry-physics-performance" and cannot accurately characterize the coupled impact of manufacturing errors (such as assembly gaps) and service damage (such as wear and thermal deformation) on dynamic performance. This leads to insufficient fidelity in digital twin models, making them unreliable as a basis for optimization design.

[0005] Secondly, even with a high-fidelity model, its high computational cost and the "curse of dimensionality" hinder its application in real-time optimization. The digital twin model of the spindle system involves strong coupling across multiple disciplines such as mechanics, thermodynamics, and dynamics, exhibiting high dimensionality and strong nonlinearity. Traditional simulation methods inherently contradict each other between "model accuracy" and "computational efficiency," making it difficult to meet the need for rapid, real-time identification of weak points under varying operating conditions. Furthermore, dynamic changes in operating conditions cause the sensitivity of design parameters to drift, further increasing the difficulty of quickly and accurately extracting key design flaw information from massive amounts of data.

[0006] In summary, these technical problems are not isolated but interconnected and mutually restrictive. Low model fusion leads to a weak foundation, while low computational efficiency makes model-based optimization and identification difficult to implement practically. Therefore, the core technical challenge that urgently needs to be overcome in this field is: how to construct a digital twin framework that can deeply integrate multi-source information (manufacturing, mechanism, working condition) and achieve efficient simulation calculation and intelligent optimization, thereby overcoming the challenge of accurate dynamic performance mapping and rapid optimization of CNC machine tool spindle systems throughout the entire "design-manufacturing-service" chain. Summary of the Invention

[0007] To address at least one of the technical problems mentioned above, this invention proposes a digital twin fusion modeling and dynamic optimization method and system for rotating spindles. Based on the concept of digital twin technology, it solves the challenges of fusion model construction, model simulation optimization, and multi-objective design optimization in the collaborative optimization research of CNC machine tool spindle systems that integrate variable working conditions and manufacturing information.

[0008] The first aspect of the present invention provides a method for digital twin fusion modeling and dynamic optimization of a rotating spindle, comprising: Acquire multi-source data from the CNC machine tool spindle system, including manufacturing parameters, operating condition data, and physical state data; Based on the manufacturing parameters and physical state data, a digital twin mechanism model characterizing the multi-field coupling effect of mechanical-thermal-dynamics is constructed; based on the operating condition data, a structured digital twin operating condition model is constructed through feature extraction and classification. The digital twin mechanism model and the digital twin operating condition model are fused together to form a dynamic fused digital twin model that responds to changing operating conditions. The dynamic fusion digital twin model is subjected to co-simulation and optimization to identify the weak design links of the spindle system under varying working conditions; based on the identified weak design links, a multi-objective optimization model is constructed with the goal of improving fatigue life and machining quality, and then solved. The optimized design parameters are dynamically simulated and verified using the dynamic fusion digital twin model, and the optimal design parameter set is output.

[0009] Furthermore, the construction steps of the digital twin mechanism model include: Based on the SysML system modeling language, the spindle system is decoupled into mechanical, thermodynamic, dynamic and control subsystems; The multi-field coupling equations of each subsystem are uniformly described using the Modelica language, and a multi-spatial-scale mechanism model is constructed. The manufacturing parameters are embedded as variables into the multi-spatial-scale mechanism model to form a parametric design optimization framework. The manufacturing parameters include any one or more of the following: geometric tolerances, surface roughness, and assembly clearance. A three-dimensional surface topography model of the spindle system, encompassing manufacturing parameters, is constructed based on fractal theory functions and characterization parameters. A Bayesian network is then used to dynamically update the fractal dimension D and characteristic scale coefficient G parameters, thereby enabling probabilistic evolution modeling of wear and thermal deformation of the system.

[0010] Furthermore, the construction steps of the digital twin working condition model include: The collected raw operating condition information is processed in an orderly manner, and state features are extracted and classified to establish an initial feature library of operating condition information. Hidden Markov Model and DS evidence theory algorithm are used to extract and simplify the working condition feature information; A multi-scale convolutional neural network model is constructed to extract transient operating condition features of high-frequency vibration and impact loads, thereby enabling the identification of high-frequency disturbance conditions. A Mealy state machine model is constructed to identify the slow-varying processes of low-frequency thermal drift and fatigue crack propagation, thereby obtaining low-frequency operating condition feature information. Based on the dynamic time warping algorithm, the similarity of working conditions is quantified, triggering an adaptive update mechanism for the initial feature library of the working condition information. Based on the XML Schema definition of the working condition feature tag system, the feature information corresponding to each typical working condition is stored as a structured XML file, and the collection of XML files constitutes the digital twin working condition model.

[0011] Furthermore, the digital twin mechanism model is integrated with the digital twin operating condition model, specifically including: The digital twin operating condition model is parsed using the Data Access command line, with real-time operating condition characteristics used as external stimulus terms. Dynamically inject the control equations of the digital twin mechanism model; Through constraint functions The stress limit and vibration amplitude threshold are directly embedded into the solution process of the digital twin mechanism model to form a set of coupled equations with multi-dimensional dynamic constraints. The dynamic loading of the operating condition threshold and excitation parameters is achieved by using dot-notation syntax, and the numerical decoupling of multiple time scales is completed by combining time step iterative algorithm. Finally, bidirectional data interaction and co-simulation between the mechanism model and the operating condition model are realized in a unified modeling environment of multiple domains.

[0012] Furthermore, the steps for co-simulation and optimization of the dynamically fused digital twin model specifically include: The Krylov subspace projection method is used to reduce the order of the dynamic fusion digital twin model to construct a dimensionality-reduced digital twin model. The accuracy of the reduced-order model is dynamically evaluated based on the time-domain maximum error limit evaluation mechanism, and its calculation formula is as follows: ; in, , The outputs are for the full model and the reduced-order model, respectively. As a preset error threshold, when At that time, dynamic optimization of the model is achieved by incrementally updating the reduced-order basis vectors; A cloud-edge collaborative computing architecture is constructed, in which the cloud is responsible for performing computationally intensive tasks such as model reduction and global optimization, while the edge is responsible for real-time data acquisition and local simulation. The architecture uses Kubernetes cluster management technology to achieve elastic scaling and dynamic allocation of computing resources, and the MQTT protocol and its QoS level guarantee mechanism to achieve efficient and reliable communication and data synchronization between the cloud and the edge.

[0013] Furthermore, the step of identifying weak design elements specifically includes: A universal identification model for mapping the relationship between design parameters and performance degradation is constructed based on a deep neural network, and trained by minimizing the cross-entropy loss function. A domain adaptation mechanism is introduced, using Maximum Mean Difference (MMD) constraint to ensure consistency of feature distributions between the source and target domains. The calculation formula is as follows: ; in, For kernel function mapping, To balance the factors, a grid search optimization was used to determine them. The target domain dataset is real-time simulation data. Source domain dataset, To design the eigenvectors of the parameters, Labeled as performance degradation; By jointly optimizing the classification loss and the distribution difference loss, the universal recognition model is dynamically adapted to the target working condition. The model outputs a performance sensitivity score based on the transfer optimization, and the key design parameter set is determined by a dynamic threshold, thereby realizing the dynamic identification of weak links.

[0014] A second aspect of the present invention provides a digital twin fusion modeling and dynamic optimization system for a rotating spindle, comprising: The multi-source data acquisition module is used to acquire multi-source data from the CNC machine tool spindle system, including manufacturing parameters, operating condition data, and physical state data. The digital twin model construction module is used to construct a digital twin mechanism model of the CNC machine tool spindle system that characterizes the multi-field coupling effect of mechanical-thermal-dynamic based on the manufacturing parameters and physical state data; and to construct a structured digital twin working condition model of the CNC machine tool spindle system through feature extraction and classification based on the working condition data. The model fusion module is used to fuse the digital twin mechanism model with the digital twin working condition model to form a dynamic fused digital twin model that can respond to changing working conditions. The simulation optimization and weak link identification module is used to perform collaborative simulation and optimization on the dynamic fusion digital twin model, identify the weak design links of the spindle system under varying working conditions, and construct and solve a multi-objective optimization model based on the identified weak design links to improve fatigue life and machining quality. The dynamic verification and output module is used to perform dynamic performance simulation verification on the optimized design parameters through the dynamic fusion digital twin model, and output the optimal design parameter set.

[0015] A third aspect of the present invention provides an electronic device including a memory, a processor, and a program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps in the rotating spindle digital twin fusion modeling and dynamic optimization method as described in the first aspect of the present invention.

[0016] A fourth aspect of the present invention provides a computer-readable storage medium having a program stored thereon that, when executed by a processor, implements the steps in the rotating spindle digital twin fusion modeling and dynamic optimization method as described in the first aspect of the present invention.

[0017] A fifth aspect of the present invention provides a computer program product comprising software code, wherein the program in the software code performs the steps of the rotating spindle digital twin fusion modeling and dynamic optimization method as described in the first aspect of the present invention.

[0018] Compared with existing technologies, the present invention provides a digital twin fusion modeling and dynamic optimization method and system for rotating spindles, which has the following advantages: (1) To address the technical problem of insufficient fusion of multi-source heterogeneous information in digital twin models, which makes it difficult to accurately characterize the coupling effect between manufacturing and service, this invention constructs a multi-field coupling mechanism model using SysML and Modelica and embeds manufacturing parameters. Simultaneously, it extracts multi-scale features from operating condition data using Hidden Markov Models, DS evidence theory, CNN, and Mealy state machines. Furthermore, it innovatively injects operating conditions as dynamic stimuli and constraints into the mechanism model solution process through Data Access command lines and dot-notation syntax. This technical solution achieves deep, structured fusion of the mechanism model, manufacturing information, and operating condition data within a unified framework, breaking down "data silos" and establishing a high-fidelity, dynamically updatable digital twin. This lays the foundation for accurately predicting performance evolution under the coupling effect of varying operating conditions and manufacturing deviations.

[0019] (2) To address the technical problem of high computational complexity in high-fidelity models and the difficulty in balancing simulation optimization efficiency and accuracy, this invention employs the Krylov subspace projection method for model order reduction and combines it with a time-domain maximum error limit mechanism to dynamically evaluate and optimize the order reduction accuracy. Simultaneously, a cloud-edge collaborative computing architecture based on Kubernetes and MQTT protocols is constructed. This technical solution significantly improves simulation efficiency while ensuring high fidelity of key model characteristics, effectively overcoming the computational bottleneck caused by the "curse of dimensionality," and providing a feasible technical path for achieving large-scale parameter optimization and rapid response.

[0020] (3) To address the technical problem of parameter sensitivity drift caused by dynamic changes in operating conditions, making it difficult to quickly and accurately identify weak links, this invention introduces a transfer learning mechanism. By constraining the consistency of feature distribution between the source and target domains through maximum mean difference (MMD), the classification loss and distribution difference loss are jointly optimized. This technical solution enables the identification model to have adaptive capabilities across operating conditions, effectively overcomes the data distribution offset problem, and significantly improves the accuracy and robustness of dynamically identifying weak design links of the spindle system under varying operating conditions.

[0021] (4) Combining the above technical solutions, this invention constructs a complete technical closed loop from "multi-source fusion modeling" to "cooperative simulation optimization" and then to "multi-objective decision-making". The method and system ultimately realize the dynamic, multi-objective collaborative optimization of the design parameters of the CNC machine tool spindle system. While shortening the optimization iteration cycle, it can significantly improve the fatigue life and machining quality of the spindle system, providing an effective theoretical tool and engineering practice solution for the reliability design and intelligent operation and maintenance of high-end CNC equipment. Attached Figure Description

[0022] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure.

[0023] Figure 1 This is a flowchart of the digital twin fusion modeling and dynamic optimization method for rotating spindles provided in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the overall technical route of the method provided in Embodiment 1 of the present invention; Figure 3 This is a flowchart of the method for constructing a digital twin working condition model of a CNC machine tool spindle system according to Embodiment 1 of the present invention; Figure 4 This is a flowchart of the dynamic identification method for weak links driven by collaborative simulation optimization of digital twin models provided in Embodiment 1 of the present invention; Figure 5 This is a schematic diagram of the multi-objective intelligent optimization technology route for CNC machine tool spindle systems provided in Embodiment 1 of the present invention; Figure 6 This is a schematic diagram of the rotating spindle digital twin fusion modeling and dynamic optimization system provided in Embodiment 2 of the present invention. Detailed Implementation

[0024] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0025] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form as well. Furthermore, it should be understood that the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0026] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0027] All data acquisition in this embodiment is carried out in accordance with laws and regulations and with user consent, and the data is used legally.

[0028] Example 1 like Figure 1This embodiment provides a method for digital twin fusion modeling and dynamic optimization of a rotating spindle, including: Acquire multi-source data from the CNC machine tool spindle system, including manufacturing parameters, operating condition data, and physical state data; Based on the manufacturing parameters and physical state data, a digital twin mechanism model characterizing the multi-field coupling effect of mechanical-thermal-dynamics is constructed; based on the operating condition data, a structured digital twin operating condition model is constructed through feature extraction and classification. The digital twin mechanism model and the digital twin operating condition model are fused together to form a dynamic fused digital twin model that responds to changing operating conditions. The dynamic fusion digital twin model is subjected to co-simulation and optimization to identify the weak design links of the spindle system under varying working conditions; based on the identified weak design links, a multi-objective optimization model is constructed with the goal of improving fatigue life and machining quality, and then solved. The optimized design parameters are dynamically simulated and verified using the dynamic fusion digital twin model, and the optimal design parameter set is output.

[0029] Specifically, the steps for constructing the digital twin mechanism model include: Based on the SysML system modeling language, the spindle system is decoupled into mechanical, thermodynamic, dynamic and control subsystems; The multi-field coupling equations of each subsystem are uniformly described using the Modelica language, and a multi-spatial-scale mechanism model is constructed. The manufacturing parameters are embedded as variables into the multi-spatial-scale mechanism model to form a parametric design optimization framework. The manufacturing parameters include any one or more of the following: geometric tolerances, surface roughness, and assembly clearance. A three-dimensional surface topography model of the spindle system, encompassing manufacturing parameters, is constructed based on fractal theory functions and characterization parameters. A Bayesian network is then used to dynamically update the fractal dimension D and characteristic scale coefficient G parameters, thereby enabling probabilistic evolution modeling of wear and thermal deformation of the system.

[0030] This embodiment specifically defines how to construct a high-fidelity digital twin mechanism model. Parameterization is achieved through multi-domain unified modeling (SysML / Modelica) and embedding manufacturing parameters. Fractal theory and Bayesian networks are introduced to enable dynamic model updates. This ensures that the mechanism model not only describes the physical laws under ideal conditions but also reflects the initial impact of manufacturing deviations and evolves with the wear and deformation of the physical entity. This significantly improves the model's realism and accuracy, serving as a prerequisite and foundation for achieving precise optimization.

[0031] In one specific embodiment, the process of constructing a multi-scale digital twin mechanism model for a CNC machine tool spindle system is as follows: Based on the SysML system modeling language, the spindle system is decoupled into mechanical, thermodynamic, dynamic, and control subsystems. The Modelica language is used to achieve a unified description of the multi-field coupling equations of the spindle system, including mechanical and thermal aspects, thus enabling the construction of a multi-scale mechanistic model of the spindle system. Simultaneously, key manufacturing parameters such as geometric tolerances, surface roughness, and assembly clearances from the design optimization stage are embedded into the mechanistic model, forming a parametric design optimization framework.

[0032] To accurately reflect the dynamic changes in the performance of a CNC machine tool spindle system and maintain consistency between the digital twin mechanism model and its physical entity's performance state, this invention proposes to employ algorithms and tools such as Bayesian network models, fatigue damage physics, and fractal theory to effectively update the mechanism model. For example, fractal theory functions, characterization parameters (fractal dimension D and feature scale coefficient G), and surface fusion mechanisms (technical routes such as...) can be used. Figure 2 As shown, a three-dimensional simulation of the surface morphology of the spindle system, which includes manufacturing information, can be established. Then, a Bayesian network is used to dynamically update parameters such as D and G to complete the probabilistic evolution modeling of wear and thermal deformation of the system.

[0033] The above technical means can be used to construct a multi-scale digital twin mechanism model of the spindle system. The modular architecture supports the traceability of design parameters and provides a high-fidelity model that approximates a physical prototype for design optimization and other work.

[0034] Specifically, the construction steps of the digital twin working condition model include: The collected raw operating condition information is processed in an orderly manner, and state features are extracted and classified to establish an initial feature library of operating condition information. Hidden Markov Model and DS evidence theory algorithm are used to extract and simplify the working condition feature information; A multi-scale convolutional neural network model is constructed to extract transient operating condition features of high-frequency vibration and impact loads, thereby enabling the identification of high-frequency disturbance conditions. A Mealy state machine model is constructed to identify the slow-varying processes of low-frequency thermal drift and fatigue crack propagation, thereby obtaining low-frequency operating condition feature information. Based on the dynamic time warping algorithm, the similarity of working conditions is quantified, triggering an adaptive update mechanism for the initial feature library of the working condition information. Based on the XML Schema definition of the working condition feature tag system, the feature information corresponding to each typical working condition is stored as a structured XML file, and the collection of XML files constitutes the digital twin working condition model.

[0035] This embodiment specifically defines how to construct a structured operating condition model from messy, raw operating condition data. Multiple algorithms (HMM, DS, CNN, Mealy, DTW) are comprehensively used for feature extraction, classification, recognition, and updating, and standardized storage is performed using XML. This transforms real-time, variable, and unstructured operating condition data into structured knowledge that is computer-recognizable, processable, and queryable, providing high-quality, standardized data input for subsequent integration with mechanistic models.

[0036] In one specific embodiment, the process of constructing a digital twin working condition model of a CNC machine tool spindle system is as follows: Methods for constructing digital twin working condition models of CNC machine tool spindle systems, such as Figure 3 As shown. First, a method for orderly processing of spindle system operating condition information is studied. The operating condition information is characterized by state feature identification, extraction and optimization. Then, an initial feature library of operating condition information is established according to the corresponding form of monitoring location, operating condition features and other information.

[0037] This study investigates the extraction and simplification optimization of working condition feature information by combining Hidden Markov Models (HMMs) and the DS algorithm. A multi-scale convolutional neural network is then designed to extract transient features such as high-frequency vibration and impact loads, enabling the identification of high-frequency disturbance conditions. A Mealy state machine model is used to identify slow-varying processes such as low-frequency thermal drift and fatigue crack propagation, obtaining low-frequency working condition feature information for the CNC machine tool spindle system. Finally, a dynamic time warping algorithm is used to quantify the similarity of working conditions, triggering adaptive updates to the feature library.

[0038] Based on the XML Schema definition of the working condition feature tag system, the working condition information is stored in a structured manner. Each typical working condition corresponds to a structured XML file, and the collection of the above XML files constitutes a digital twin working condition model that accurately reflects the operating status of the CNC machine tool spindle system.

[0039] Specifically, the fusion of the digital twin mechanism model and the digital twin operating condition model includes: The digital twin operating condition model is parsed using the Data Access command line, with real-time operating condition characteristics used as external stimulus terms. Dynamically inject the control equations of the digital twin mechanism model; Through constraint functions The stress limit and vibration amplitude threshold are directly embedded into the solution process of the digital twin mechanism model to form a set of coupled equations with multi-dimensional dynamic constraints. The dynamic loading of the operating condition threshold and excitation parameters is achieved by using dot-notation syntax, and the numerical decoupling of multiple time scales is completed by combining time step iterative algorithm. Finally, bidirectional data interaction and co-simulation between the mechanism model and the operating condition model are realized in a unified modeling environment of multiple domains.

[0040] This embodiment specifically defines how to deeply integrate the mechanistic model and the operating condition model. Through Data Access and dot-notation, operating condition data is embedded into the control equations and solution process of the mechanistic model in two forms: "external excitation" and "dynamic constraints." This enables the virtual model to respond to changes in the physical world in real time, realizing the direct driving and constraint of operating conditions on mechanistic behavior, thus allowing the simulation results to realistically reflect the actual state of the spindle under specific operating conditions.

[0041] In one specific embodiment, the method for fusing the mechanism model and working condition model of the CNC machine tool spindle system is as follows: The operating condition model is parsed based on the Data Access command line, and the operating condition characteristics are injected into the external excitation terms of the mechanism model in real time. At the same time, through constraint functions By directly embedding stress limits, vibration amplitudes, and other operating condition thresholds into the solution process, a coupled set of equations with multi-dimensional dynamic constraints is formed. Dynamic parameter loading is achieved using dot-notation syntax, and numerical decoupling across multiple time scales is accomplished by combining time-step iterative algorithms. Finally, bidirectional data interaction and co-simulation between the mechanistic model and the operating condition model are realized in a unified multi-domain modeling environment, along with improved dynamic performance prediction accuracy under varying operating conditions. This provides a high-fidelity, strongly constrained co-simulation foundation for multi-objective optimization of the spindle system.

[0042] Specifically, the steps for co-simulation and optimization of the dynamically fused digital twin model include: The Krylov subspace projection method is used to reduce the order of the dynamic fusion digital twin model to construct a dimensionality-reduced digital twin model. Dynamically evaluate the accuracy of the reduced-order model based on the time-domain maximum error limit evaluation mechanism; A cloud-edge collaborative computing architecture is constructed, in which the cloud is responsible for performing computationally intensive tasks such as model reduction and global optimization, while the edge is responsible for real-time data acquisition and local simulation. The architecture uses Kubernetes cluster management technology to achieve elastic scaling and dynamic allocation of computing resources, and the MQTT protocol and its QoS level guarantee mechanism to achieve efficient and reliable communication and data synchronization between the cloud and the edge.

[0043] This embodiment specifically defines how to solve the problem of high computational complexity. It employs Krylov model reduction to shrink the model size and utilizes a cloud-edge collaborative architecture (Kubernetes + MQTT) to distribute computationally intensive tasks. While maintaining accuracy, it significantly improves computational efficiency, enabling real-time simulation and optimization based on high-fidelity models. This resolves the "accuracy-efficiency" contradiction and is crucial for the practical application of the entire method.

[0044] In one specific embodiment, the dynamic identification process of weak links driven by the co-simulation optimization of the digital twin model of the spindle system is as follows: Identifying parameters that significantly impact the dynamic performance of CNC machine tool spindle systems is crucial for recognizing weak design elements. However, digital twin models of CNC machine tool spindle systems exhibit multidisciplinary coupling and strong nonlinearity, resulting in high simulation computation costs. Therefore, this invention investigates a dual-driven model simulation optimization strategy of "model order reduction - cloud-edge collaboration" to rapidly identify weak elements. The technical approach is as follows: Figure 4 As shown.

[0045] Simulation Optimization of Digital Twin Model of Spindle System Based on Cloud-Edge Collaboration and Model Reduction An adaptive order reduction framework is constructed using the Krylov subspace projection method as the core and combined with the time-domain maximum error limit evaluation mechanism. This framework is then encapsulated into a dynamic link library to autonomously optimize and reduce the order of the described model. During model simulation optimization, the inherent characteristic parameters such as the mass matrix and stiffness matrix are first extracted through the model decoupling module. Then, the Krylov subspace projection method is used to construct reduced-order basis vectors, generating a reduced-dimensional digital twin model. Finally, the accuracy of the reduced-order model is dynamically evaluated based on the time-domain maximum error limit shown in formula (1). While preserving the necessary behavioral characteristics and dominant effects of the system, the solution speed of the simulation model is optimized.

[0046] (1) in, and The outputs are for the full model and the reduced-order model, respectively. This is a preset error threshold. When... At the same time, the model is dynamically optimized by incrementally updating the reduced-order basis vectors to ensure the high-fidelity characteristics of the reduced-order model in key frequency bands.

[0047] A cloud-edge collaborative computing platform is built based on Kubernetes cluster management technology. The cloud handles computationally intensive tasks such as high-load model reduction and global optimization, while the edge handles real-time data acquisition, local simulation, and lightweight parameter identification. Kubernetes' elastic scaling and load balancing mechanisms dynamically allocate GPU cluster computing resources, prioritizing computationally intensive tasks such as model reduction and thermal coupling simulation, achieving real-time synchronization of digital threads between the cloud and edge. The lightweight MQTT protocol enables efficient communication between the cloud and edge. The edge uploads pre-processed vibration, temperature, and other operating condition data to the cloud's HBase database in real time via MQTT's publish-subscribe model; the cloud feeds back the reduced model parameters and optimization results to the edge, forming a closed-loop control. The robustness of industrial field communication is enhanced through MQTT's QoS level guarantee mechanism and will message functionality, constructing a closed-loop feedback system that allows the twin model to dynamically correct spindle thermal deformation errors and other information, supporting spindle system health status prediction and manufacturing parameter optimization.

[0048] The above collaborative optimization framework can effectively solve the problem of high computational complexity in the multi-parameter coupled optimization of CNC machine tool spindle systems, and build a fast solution paradigm for dynamic balance between accuracy and speed.

[0049] Specifically, the steps for identifying weak design elements include: A universal identification model for mapping the relationship between design parameters and performance degradation is constructed based on a deep neural network, and trained by minimizing the cross-entropy loss function. A domain adaptation mechanism is introduced to constrain the consistency of feature distributions between the source and target domains using the maximum mean difference (MMD) constraint. By jointly optimizing the classification loss and the distribution difference loss, the universal recognition model is dynamically adapted to the target working condition. The model outputs a performance sensitivity score based on the transfer optimization, and the key design parameter set is determined by a dynamic threshold, thereby realizing the dynamic identification of weak links.

[0050] This embodiment specifically defines how to intelligently identify weak points. It employs a combination of deep learning and transfer learning (MMD) to enable the model to adapt to changing operating conditions and accurately identify key design parameters leading to performance degradation. This improves the intelligence level and cross-condition robustness of the identification process. It can automatically and accurately locate defects in complex systems, providing precise targets and directions for subsequent optimization and avoiding blind optimization.

[0051] In one specific embodiment, the dynamic identification and sensitivity analysis of weak points in the spindle system are as follows: First, a universal identification model is constructed based on a deep neural network, and it is jointly trained using historical monitoring data and public datasets. This is achieved by minimizing the cross-entropy loss function. First, establish the initial mapping relationship between design parameters and performance degradation; second, introduce a domain adaptation mechanism, as shown in formula (2), to constrain the consistency of feature distribution between the source domain (historical data) and the target domain (real-time simulation data), and jointly optimize the classification loss and distribution difference loss (objective function: This enables dynamic adaptation of the model to the target operating condition; ultimately, a performance sensitivity score is output based on the transfer optimization model. This method determines the key parameter set through dynamic thresholding. It integrates transfer learning and digital twin technologies, and can overcome data distribution bias through a two-stage training strategy, improving the generalization and robustness of parameter identification under complex and varying working conditions.

[0052] (2) in, For kernel function mapping, To balance the factors, a grid search optimization was used to determine them. The target domain dataset is real-time simulation data. Source domain dataset, To design the eigenvectors of the parameters, This is a performance degradation label.

[0053] Research methods for multi-objective intelligent design and optimization of CNC machine tool spindle systems based on digital twins, such as... Figure 5 As shown, the detailed technical solution is as follows: (a) Multi-objective optimization model construction module for weak links of spindle system The core of optimizing CNC machine tool spindle systems lies in establishing a mathematical model that can comprehensively reflect multi-dimensional performance indicators. Firstly, based on a fatigue life prediction model driven by digital twins, and combining material physical properties with the dynamic load spectrum of the working conditions, a time-varying fatigue life objective function is constructed. By extracting multi-source features such as spindle vibration and temperature using a deep convolutional neural network, and introducing an attention mechanism to enhance the sensitivity of key parameters such as spindle stiffness K and bearing preload Fp, a system is established. ,in To correct for crack propagation based on the Paris formula, the model is designed to achieve both data fitting accuracy and physical interpretability. Then, fractal theory is used to quantify the surface morphology, and a random forest algorithm is employed to select the weights of cutting parameters on surface roughness Ra, thus constructing a manufacturing quality objective function. ( This represents the assembly error disturbance term, enabling an explicit characterization of manufacturing process uncertainties. Based on the constructed... , The model is used as the objective function of the multi-objective optimization model for the weak link of the main shaft system.

[0054] Based on the real-time analysis of the impact load Fpeak, temperature rise ΔT, and other variable working condition information on the design boundary using a fused digital twin model, a dynamic constraint function is established: (3) in For time-varying von Mises stress, dynamic updates are achieved through finite element simulation and sensor data fusion. A multi-objective optimization model is established with the objectives of minimizing fatigue life loss and maximizing manufacturing quality. (4) in, , These are weighting coefficients, dynamically adjusted using domain expert experience and data-driven methods. and These represent the disturbance terms of operating condition fluctuations and manufacturing uncertainties to the target, respectively.

[0055] (b) Multi-objective optimization solution module for weak links in CNC machine tool spindle system An improved particle swarm optimization (PSO) algorithm (deployed in the cloud, performing high-dimensional design space search) integrating multiple strategies is applied to find the Pareto optimal solution for a multi-objective optimization model of key design information of a CNC machine tool spindle system. Based on the traditional PSO algorithm, a random decision number Rd is introduced to control population diversity. When the fitness value exceeds Rd, selection, crossover, and mutation operations are performed on the population, constructing a new particle swarm through an elite retention strategy. The Pareto optimal design solution set is output when the stopping criterion is met. Each set of key design information in this solution set satisfies the multi-objective constraints and objectives.

[0056] (c) Dynamic performance analysis and verification module of CNC machine tool spindle system The entropy weight method is used to perform multi-objective decision-making on the Pareto solution set. j Entropy value of the item index With weight The calculation is as follows: (5) in, n The formula for calculating the weights is: (The formula is missing from the original text.) The optimal solution is selected through weighted scoring to define the optimal design information. Dynamic feedback of optimization parameters is achieved via the MQTT protocol, and this feedback is based on the optimal design parameters and... Figure 3The model update strategy shown implements the construction of an optimized digital twin model of the CNC machine tool spindle system with accompanying manufacturing information. Abnormal operating condition data is loaded into this model for dynamic performance simulation analysis. The dynamic performance data of the weak points corresponding to the abnormal operating conditions are observed to determine whether the optimized design scheme meets the design requirements.

[0057] Digital Twin System Construction and Application Verification A multi-field coupled digital twin model of the CNC machine tool spindle system was constructed by integrating information from multiple sources such as manufacturing and operating conditions. Based on cloud-edge collaboration and model reduction technology, the model was optimized through simulation. The model was then applied and verified on the VNM1100 high-precision vertical machining center of Shandong Weida Precision Intelligent Equipment Co., Ltd. The digital twin model was deployed on-site and tested online. After identifying the weak design links of the equipment, the reliability optimization design of the spindle system was quickly carried out. This provides a technical demonstration for the practical engineering transformation of the design optimization theory of CNC machine tool spindle systems.

[0058] Example 2 like Figure 6 As shown, this embodiment provides a digital twin fusion modeling and dynamic optimization system for a rotating spindle, including: The multi-source data acquisition module is used to acquire multi-source data from the CNC machine tool spindle system, including manufacturing parameters, operating condition data, and physical state data. The digital twin model construction module is used to construct a digital twin mechanism model of the CNC machine tool spindle system that characterizes the multi-field coupling effect of mechanical-thermal-dynamic based on the manufacturing parameters and physical state data; and to construct a structured digital twin working condition model of the CNC machine tool spindle system through feature extraction and classification based on the working condition data. The model fusion module is used to fuse the digital twin mechanism model with the digital twin working condition model to form a dynamic fused digital twin model that can respond to changing working conditions. The simulation optimization and weak link identification module includes a simulation optimization module and a weak link identification module, which are used to perform collaborative simulation and optimization on the dynamic fusion digital twin model, identify the weak design links of the spindle system under variable working conditions; based on the identified weak design links, a multi-objective optimization model with the goal of improving fatigue life and machining quality is constructed and solved; The dynamic verification and output module is used to perform dynamic performance simulation verification on the optimized design parameters through the dynamic fusion digital twin model, and output the optimal design parameter set.

[0059] Example 3 Embodiment 3 of the present invention provides an electronic device.

[0060] An electronic device includes a memory, a processor, and a program stored in the memory and running on the processor. When the processor executes the program, it implements the steps in the rotating spindle digital twin fusion modeling and dynamic optimization method as described in Embodiment 1 of the present invention.

[0061] The detailed steps are the same as those of the rotating spindle digital twin fusion modeling and dynamic optimization method provided in Example 1, and will not be repeated here.

[0062] Example 4 Embodiment 4 of the present invention provides a computer-readable storage medium.

[0063] A computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps in the rotating spindle digital twin fusion modeling and dynamic optimization method as described in Embodiment 1 of the present invention.

[0064] The detailed steps are the same as those of the rotating spindle digital twin fusion modeling and dynamic optimization method provided in Example 1, and will not be repeated here.

[0065] Example 5 Embodiment 5 of the present invention provides a computer program product.

[0066] A computer program product includes software code, wherein the program in the software code performs the steps of the rotating spindle digital twin fusion modeling and dynamic optimization method as described in Embodiment 1 of the present invention.

[0067] The detailed steps are the same as those of the rotating spindle digital twin fusion modeling and dynamic optimization method provided in Example 1, and will not be repeated here.

[0068] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0069] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0070] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0071] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0072] The above description is merely a preferred embodiment of this practice and is not intended to limit the scope of this practice. Various modifications and variations can be made to this practice by those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of this practice should be included within the protection scope of this practice.

Claims

1. A method for digital twin fusion modeling and dynamic optimization of a rotating spindle, characterized in that, include: Acquire multi-source data from the CNC machine tool spindle system, including manufacturing parameters, operating condition data, and physical state data; Based on the manufacturing parameters and physical state data, a digital twin mechanism model characterizing the multi-field coupling effect of mechanical-thermal-dynamics is constructed; based on the operating condition data, a structured digital twin operating condition model is constructed through feature extraction and classification. The digital twin mechanism model and the digital twin operating condition model are fused together to form a dynamic fused digital twin model that responds to changing operating conditions. The dynamic fusion digital twin model is subjected to co-simulation and optimization to identify the weak design links of the spindle system under varying working conditions. Based on the identified weak design elements, a multi-objective optimization model is constructed with the goal of improving fatigue life and processing quality, and then solved. The optimized design parameters are dynamically simulated and verified using the dynamic fusion digital twin model, and the optimal design parameter set is output.

2. The method as described in claim 1, characterized in that, The steps for constructing the digital twin mechanism model include: Based on the SysML system modeling language, the spindle system is decoupled into mechanical, thermodynamic, dynamic and control subsystems; The multi-field coupling equations of each subsystem are uniformly described using the Modelica language, and a multi-spatial-scale mechanism model is constructed. The manufacturing parameters are embedded as variables into the multi-spatial-scale mechanism model to form a parametric design optimization framework. The manufacturing parameters include any one or more of the following: geometric tolerances, surface roughness, and assembly clearance. A three-dimensional surface topography model of the spindle system, encompassing manufacturing parameters, is constructed based on fractal theory functions and characterization parameters. A Bayesian network is then used to dynamically update the fractal dimension D and characteristic scale coefficient G parameters, thereby enabling probabilistic evolution modeling of wear and thermal deformation of the system.

3. The method as described in claim 1, characterized in that, The steps for constructing the digital twin operating condition model include: The collected raw operating condition information is processed in an orderly manner, and state features are extracted and classified to establish an initial feature library of operating condition information. Hidden Markov Model and DS evidence theory algorithm are used to extract and simplify the working condition feature information; A multi-scale convolutional neural network model is constructed to extract transient operating condition features of high-frequency vibration and impact loads, thereby enabling the identification of high-frequency disturbance conditions. A Mealy state machine model is constructed to identify the slow-varying processes of low-frequency thermal drift and fatigue crack propagation, thereby obtaining low-frequency operating condition feature information. Based on the dynamic time warping algorithm, the similarity of working conditions is quantified, triggering an adaptive update mechanism for the initial feature library of the working condition information. Based on the XML Schema definition of the working condition feature tag system, the feature information corresponding to each typical working condition is stored as a structured XML file, and the collection of XML files constitutes the digital twin working condition model.

4. The method as described in claim 1, characterized in that, The fusion of the digital twin mechanism model and the digital twin operating condition model specifically includes: The digital twin operating condition model is parsed using the Data Access command line, with real-time operating condition characteristics used as external stimulus terms. Dynamically inject the control equations of the digital twin mechanism model; Through constraint functions The stress limit and vibration amplitude threshold are directly embedded into the solution process of the digital twin mechanism model to form a set of coupled equations with multi-dimensional dynamic constraints. The dynamic loading of the operating condition threshold and excitation parameters is achieved by using dot-notation syntax, and the numerical decoupling of multiple time scales is completed by combining time step iterative algorithm. Finally, bidirectional data interaction and co-simulation between the mechanism model and the operating condition model are realized in a unified modeling environment of multiple domains.

5. The method as described in claim 1, characterized in that, The specific steps for co-simulating and optimizing the dynamic fused digital twin model include: The Krylov subspace projection method is used to reduce the order of the dynamic fusion digital twin model to construct a dimensionality-reduced digital twin model. The accuracy of the reduced-order model is dynamically evaluated based on the time-domain maximum error limit evaluation mechanism, and its calculation formula is as follows: ； in, , The outputs are for the full model and the reduced-order model, respectively. As a preset error threshold, when At that time, dynamic optimization of the model is achieved by incrementally updating the reduced-order basis vectors; A cloud-edge collaborative computing architecture is constructed, in which the cloud is responsible for performing computationally intensive tasks such as model reduction and global optimization, while the edge is responsible for real-time data acquisition and local simulation. The architecture uses Kubernetes cluster management technology to achieve elastic scaling and dynamic allocation of computing resources, and the MQTT protocol and its QoS level guarantee mechanism to achieve efficient and reliable communication and data synchronization between the cloud and the edge.

6. The method as described in claim 1, characterized in that, The steps for identifying weak design elements specifically include: A universal identification model for mapping the relationship between design parameters and performance degradation is constructed based on a deep neural network, and trained by minimizing the cross-entropy loss function. A domain adaptation mechanism is introduced, using Maximum Mean Difference (MMD) constraint to ensure consistency of feature distributions between the source and target domains. The calculation formula is as follows: ； in, For kernel function mapping, To balance the factors, a grid search optimization was used to determine them. The target domain dataset is real-time simulation data. Source domain dataset, To design the eigenvectors of the parameters, Labeled as performance degradation; By jointly optimizing the classification loss and the distribution difference loss, the universal recognition model is dynamically adapted to the target working condition. The model outputs a performance sensitivity score based on the transfer optimization, and the key design parameter set is determined by a dynamic threshold, thereby realizing the dynamic identification of weak links.

7. A digital twin fusion modeling and dynamic optimization system for a rotating spindle, characterized in that, include: The multi-source data acquisition module is used to acquire multi-source data from the CNC machine tool spindle system, including manufacturing parameters, operating condition data, and physical state data. The digital twin model construction module is used to construct a digital twin mechanism model of the CNC machine tool spindle system that characterizes the multi-field coupling effect of mechanical-thermal-dynamic based on the manufacturing parameters and physical state data; and to construct a structured digital twin working condition model of the CNC machine tool spindle system through feature extraction and classification based on the working condition data. The model fusion module is used to fuse the digital twin mechanism model with the digital twin working condition model to form a dynamic fused digital twin model that can respond to changing working conditions. The simulation optimization and weak link identification module is used to perform collaborative simulation and optimization on the dynamic fusion digital twin model and identify the weak design links of the spindle system under variable working conditions. Based on the identified weak design elements, a multi-objective optimization model is constructed with the goal of improving fatigue life and processing quality, and then solved. The dynamic verification and output module is used to perform dynamic performance simulation verification on the optimized design parameters through the dynamic fusion digital twin model, and output the optimal design parameter set.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the program, it implements the steps of the rotating spindle digital twin fusion modeling and dynamic optimization method as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the rotating spindle digital twin fusion modeling and dynamic optimization method as described in any one of claims 1 to 6.

10. A computer program product, comprising software code, characterized in that, The program in the software code executes the steps of the rotating spindle digital twin fusion modeling and dynamic optimization method as described in any one of claims 1 to 6.

Images