Method and system for optimizing bolt surface treatment process parameters under digital twin environment

Abstract

This invention discloses a method and system for optimizing bolt surface treatment process parameters in a digital twin environment, comprising: S1: collecting multi-dimensional information of the bolt surface; constructing a high-fidelity digital twin model based on the multi-dimensional information of the bolt surface to obtain a digital feature matrix of the bolt surface; S2: establishing a mapping function for the influence of surface treatment process parameters on the bolt surface state based on the digital feature matrix; and constructing a constraint space for surface treatment process parameters; S3: designing a multi-objective adaptive particle swarm optimization algorithm that integrates bolt surface quality prediction based on the mapping function; S4: iteratively solving within the constraint space of surface treatment process parameters based on the multi-objective adaptive particle swarm optimization algorithm to obtain a set of candidate process parameter combinations. This invention can solve the technical problems of low accuracy, poor efficiency, and difficulty in multi-objective coordination in traditional bolt surface treatment process parameter optimization methods.

Description

Technical Field

[0001] This invention relates to the technical field of surface treatment processes, and in particular to a method and system for optimizing bolt surface treatment process parameters in a digital twin environment. Background Technology

[0002] Bolts, as critical fasteners in mechanical connections, are widely used in important industrial sectors such as aerospace, automotive manufacturing, petrochemicals, and nuclear power equipment. With the increasing demands for reliability and safety in modern industry, the surface treatment quality of bolts has become a key factor affecting their service performance and lifespan. High-quality bolt surface treatment not only improves their corrosion resistance and fatigue strength but also significantly enhances their surface friction characteristics and appearance, which is crucial for ensuring the long-term stable operation of mechanical equipment.

[0003] Currently, bolt surface treatment processes mainly include multiple steps such as sandblasting, chemical treatment, heat treatment, and coating, each involving the precise control of multiple process parameters. However, traditional process parameter optimization methods generally suffer from low optimization accuracy, poor efficiency, and difficulties in coordinating multiple objectives. Most existing methods still rely on empirical trial-and-error methods or simple single-factor experiments, lacking accurate modeling of the complex nonlinear relationship between process parameters and surface conditions, resulting in optimization results that often fail to achieve the expected results.

[0004] Against the backdrop of the rapid development of digital manufacturing technology, digital twin technology offers a new approach to solving the problem of optimizing traditional process parameters. For example, existing technology CN202311234412.7 discloses a digital twin process for laser wire-filling welding. This technology is based on the COMSOL Multiphysic multiphysics simulation platform and achieves digital simulation of the welding process through steps such as establishing a geometric model, setting physical field parameters, and performing mesh generation. However, this technology is mainly limited to specific laser welding processes and can only handle parameter settings for a single process, making it unsuitable for complex process systems involving multiple continuous steps, such as bolt surface treatment.

[0005] More importantly, existing digital twin technologies have significant shortcomings in the design of process parameter optimization algorithms. Traditional methods typically employ optimization algorithms with fixed parameters, lacking adaptive adjustment capabilities and prone to getting trapped in local optima when facing complex multi-objective optimization problems. Furthermore, existing technologies often neglect the coordination and balance between surface quality, process efficiency, and processing costs, failing to provide comprehensive optimization solutions for engineering practice. In addition, traditional optimization algorithms lack effective utilization of actual surface condition information, and the optimization process lacks specificity and foresight, making it difficult to achieve high-precision process parameter optimization.

[0006] Therefore, there is an urgent need to develop a method for optimizing bolt surface treatment process parameters based on a digital twin environment. This method involves constructing a high-fidelity digital twin model, establishing precise process parameter mapping relationships, and designing an intelligent multi-objective optimization algorithm to achieve accurate optimization of bolt surface treatment process parameters, thereby meeting the urgent needs of modern industry for high-quality bolt products. Summary of the Invention

[0007] In view of this, the present invention provides a method and system for optimizing bolt surface treatment process parameters in a digital twin environment, aiming to solve the technical problems of low accuracy, poor efficiency, and difficulty in coordinating multiple objectives in traditional bolt surface treatment process parameter optimization methods.

[0008] To achieve the above objectives, the present invention provides a method for optimizing bolt surface treatment process parameters in a digital twin environment, comprising the following steps:

[0009] S1: Collect multi-dimensional information about the bolt surface; based on the multi-dimensional information about the bolt surface, construct a high-fidelity digital twin model to obtain the digital feature matrix of the bolt surface;

[0010] S2: Based on the digital feature matrix, establish a mapping function for the influence of surface treatment process parameters on the surface condition of bolts; and construct a constraint space for surface treatment process parameters;

[0011] S3: Based on the mapping function, design a multi-objective adaptive particle swarm optimization algorithm that integrates bolt surface quality prediction;

[0012] S4: Based on the multi-objective adaptive particle swarm optimization algorithm, iterative solutions are performed within the constraint space of surface treatment process parameters to obtain a set of candidate process parameter combinations.

[0013] As a further improvement of the present invention:

[0014] Optionally, step S1 includes:

[0015] S11: Collect multi-dimensional information about the bolt surface:

[0016] Multi-dimensional information of the bolt surface was acquired using a laser scanner, a high-resolution industrial camera, and a surface roughness measuring instrument. This multi-dimensional information included geometric point cloud data of the bolt surface. Surface texture images and roughness distribution data Among them, geometric point cloud data Record the three-dimensional coordinates of each spatial position on the bolt surface , The horizontal coordinate representing spatial location. The vertical coordinate represents the spatial location. The height coordinates represent spatial locations;

[0017] S12: Generate a high-fidelity digital twin model:

[0018] The multi-dimensional information of the bolt surface collected in step S11 is spatially registered and fused to construct a high-fidelity digital twin model; the construction process includes: firstly, using the iterative nearest point algorithm to process the geometric point cloud data. Surface texture images and roughness distribution data Spatial alignment is performed, and then a digital feature matrix of the bolt surface is constructed. Each element of this digitized feature matrix It includes the three-dimensional coordinates of each spatial location on the bolt surface, the pixel values ​​of the three-dimensional coordinates of the surface texture image, and the data values ​​of the three-dimensional coordinates of the roughness distribution data. This represents the row index in the digitized feature matrix. This represents the column index in the digitized feature matrix.

[0019] Optionally, step S2 includes:

[0020] S21: Surface treatment process parameters for collection and classification:

[0021] Key process parameters in the bolt surface treatment process are collected and classified to obtain a set of surface treatment process parameters. ,in, This indicates the parameters for sandblasting, including sandblasting pressure and sandblasting time. This indicates the chemical processing parameters, including the chemical solution concentration and the chemical processing temperature; This indicates the heat treatment parameters, including heating temperature and holding time; This indicates the coating treatment parameters, including coating thickness and curing time; each surface treatment process parameter has its corresponding numerical range.

[0022] S22: Establishment of the mapping relationship between surface treatment process parameters and bolt surface condition:

[0023] By analyzing the digital feature matrix Set of elements and surface treatment process parameters To establish the correlation between them, a mapping function is established to show the influence of surface treatment process parameters on the surface condition of bolts. The mapping function The data describes the effect of different combinations of process parameters on the surface roughness distribution of bolts. and surface texture images The quantitative influence relationship was determined; the mapping coefficient was calculated using multiple regression analysis. ,in This represents the sensitivity coefficient of each surface treatment process parameter to changes in surface condition;

[0024] S23: Constructing the constraint space for surface treatment process parameters :

[0025] Determine the feasible domain boundary for each surface treatment process parameter, where the constraint range for the sandblasting pressure value is: ,in Indicates the minimum effective sandblasting pressure. This indicates the maximum safe sandblasting pressure; the sandblasting time is within a specified range. ,in Indicates the minimum sandblasting time. This indicates the maximum blasting time; the constraint range for chemical solution concentration is... ,in Indicates the minimum effective concentration. This indicates the maximum safe concentration; the constraint range for chemical treatment temperature is... ,in Indicates the minimum reaction temperature. Indicates the maximum reaction temperature; the heating temperature is within a specified range. ,in Indicates the minimum processing temperature. Indicates the highest temperature the material can withstand; the holding time is within the specified range. ,in Indicates the shortest heat preservation time. Indicates the maximum heat preservation time; the coating thickness is within a specified range. ,in Indicates the minimum effective coating thickness. This indicates the maximum permissible coating thickness; the curing time is within a specified range. ,in Indicates the shortest curing time. This indicates the longest curing time.

[0026] Optionally, step S3 includes:

[0027] S31: Constructing a surface roughness gradient guidance mechanism:

[0028] Based on digital feature matrix Roughness distribution data in Calculate the surface roughness gradient field of the bolt. Surface roughness gradient field Each spatial location Corresponding to a gradient vector ,in Indicates the gradient vector at directional components, Indicates the gradient vector at directional components, Indicates the gradient vector at Components of direction; based on gradient vector Calculate the particle guiding vector The particle guiding vector Pointing to the optimal direction for improving surface quality;

[0029] When surface roughness data is subject to noise interference or gradient calculation in the boundary region is difficult, a gradient estimation alternative based on wavelet denoising is adopted, using wavelet transform. Filtering out high-frequency noise and combining radial basis function interpolation Perform boundary gradient extrapolation, where Represents the wavelet transform coefficients. Indicates the scale parameter. Indicates the translation parameter. This represents the gradient value in the boundary region. Indicates the interpolation weight coefficients. This indicates the number of data points involved in the interpolation. Represents radial basis functions. Represents the current position vector. This represents the vector of known data point locations.

[0030] S32: Design a dynamic inertia weight adjustment strategy: using a mapping function The root mean square (RMS) value of bolt surface roughness is predicted and calculated under the current combination of surface treatment process parameters. The surface treatment effect evaluation index is then calculated by comparing the changes in the RMS value before and after surface treatment. And adjust the dynamic inertia weight according to the surface treatment effect evaluation index;

[0031] S33: Solving multi-objective collaborative optimization: Construct a multi-objective optimization function that includes surface quality objective, process efficiency objective and processing cost objective, and construct an improved particle velocity update formula based on particle guiding vector and dynamic inertia weight.

[0032] Optionally, step S32 includes:

[0033] The surface treatment effect evaluation index is calculated by comparing the changes in the root mean square roughness value of the bolts before and after surface treatment. ; Root mean square value of initial surface roughness of bolt By analyzing the digital feature matrix Medium roughness distribution data The root mean square value of bolt surface roughness was obtained through statistical analysis; under the current combination of surface treatment process parameters. Through mapping function Predicted and calculated based on the process parameter values ​​corresponding to the current particle position; surface treatment effect evaluation index. The calculation formula is:

[0034] ;

[0035] Define the current iteration number as The maximum number of iterations is Dynamic inertia weight The calculation formula is:

[0036] ;

[0037] in, Indicates the initial inertia weight. Indicates the termination of inertia weight. Indicates an adaptive adjustment index. This represents the feedback coefficient.

[0038] Optionally, step S33 includes:

[0039] Calculate three objective function values, including the surface quality objective. Process efficiency target Processing cost target Surface quality target Calculated using the standard deviation of bolt surface roughness; process efficiency target. Calculate the processing cost target by taking the reciprocal of the total processing time. Construct a multi-objective optimization function by calculating the reciprocal of the unit cost. The expression is:

[0040] ;

[0041] in, The weighting coefficients representing surface quality targets. The weighting coefficients representing the process efficiency target. Weighting coefficients representing the cost objective; multi-objective optimization function The calculation results are used as particle fitness values ​​to evaluate the merits of the current particle position and to determine the individual's historical best position. and the global historical best position The update provides a basis for judgment;

[0042] Based on the particle guidance vector in step S31 and the dynamic inertia weights in step S32 An improved particle velocity update formula is constructed, which is as follows:

[0043] ;

[0044] in, Indicates the particle in the first... Speed ​​at the next iteration Indicates the particle in the first... Speed ​​at the next iteration Represents individual learning factors. Represents the global learning factor. Indicates the gradient guiding factor. and Represents a random number within the interval [0,1]. Indicates the particle in the first... Position at the next iteration; individual's historical best position and the global historical best position Both are based on multi-objective optimization functions The fitness value is selected and updated to ensure that the particle velocity update direction points to the solution space region with higher fitness.

[0045] Optionally, step S4 includes:

[0046] S41: Initialize the particle swarm:

[0047] Within the constraints of surface treatment process parameters An initial particle swarm is randomly generated internally, and the position of each particle is... This represents a combination of surface treatment process parameters, including sandblasting parameters, chemical treatment parameters, heat treatment parameters, and coating treatment parameters; the particle swarm size is set to... Initialize the velocity of each particle. Let the zero vector be the individual's historical best position. Equal to the initial position, the global historical best position Determined by comparing the initial fitness values ​​of all particles;

[0048] S42: Iterative optimization solution:

[0049] For each iteration Perform the following operations: First, based on the current particle position Through mapping function Calculate the predicted surface roughness value of the bolts under the current combination of surface treatment process parameters, and then calculate the surface treatment effect evaluation index. ; Calculate the dynamic inertia weight according to step S3 and particle guiding vector ; Calculate the values ​​of the three objective functions , , And through multi-objective optimization function Calculate the fitness value for each particle; the fitness value is used to evaluate the particle's performance in the target space consisting of surface quality, process efficiency, and processing cost targets, and serves as the basis for subsequently updating the individual's historical best position and the global historical best position. If the current particle's fitness value is better than its historical best fitness value, then the individual's historical best position is updated. If there is a particle in the current population with a fitness value better than the global historical best fitness value, then update the global historical best position. Update the individual's historical best position. and the global historical best position The new velocity is calculated based on the improved particle velocity update formula. and update particle positions. For surfaces exceeding the constraints of the surface treatment process parameters The particle positions are processed to define the boundaries;

[0050] An anomaly monitoring mechanism is implemented during the iteration process. When more than 30% of the particles are detected to have gone out of the constraint space, the particle swarm repair strategy is initiated: the out-of-bounds particles are pulled back into the feasible region using the mirror reflection method, while the elite retention strategy is adopted to keep the current generation's best 10% of particles from being modified.

[0051] S43: Generate the Pareto optimal solution set:

[0052] When the maximum number of iterations is reached The iteration process terminates when a convergence condition is met, wherein the improvement of the global optimal solution over 10 consecutive iterations is less than 0.001; a Pareto optimal solution set is selected from the non-dominated solutions in all iterations. The non-dominated solution refers to a solution that is not completely dominated by other solutions at the three objective function values; the Pareto optimal solution set All solutions are used as combinations of process parameters to form a set of candidate process parameter combinations. Each candidate process parameter combination contains complete surface treatment process parameter settings.

[0053] This invention also discloses a bolt surface treatment process parameter optimization system in a digital twin environment, comprising:

[0054] Digitalization module: Collects multi-dimensional information about the bolt surface; Based on the multi-dimensional information about the bolt surface, constructs a high-fidelity digital twin model to obtain the digital feature matrix of the bolt surface;

[0055] Constraint Module: Based on the digital feature matrix, a mapping function is established to show the influence of surface treatment process parameters on the surface condition of bolts; and a constraint space for surface treatment process parameters is constructed.

[0056] Multi-objective optimization module: Based on the mapping function, a multi-objective adaptive particle swarm optimization algorithm that integrates bolt surface quality prediction is designed;

[0057] Iterative solution module: Based on the multi-objective adaptive particle swarm optimization algorithm, iterative solution is performed within the constraint space of surface treatment process parameters to obtain a set of candidate process parameter combinations.

[0058] Compared with the prior art, the present invention has at least the following beneficial effects:

[0059] The multi-objective optimization function designed in this invention simultaneously considers three key objectives: surface quality, process efficiency, and processing cost. A coordinated balance among these objectives is achieved through a reasonable allocation of weighting coefficients. Compared to traditional single-objective optimization methods, the algorithm of this invention avoids the problem of excessively pursuing a single objective leading to severe deterioration of other objectives. The generated Pareto optimal solution set provides diverse parameter combinations for engineering practice.

[0060] This invention innovatively designs a dynamic inertia weight adjustment strategy, which dynamically adjusts the algorithm's search behavior through real-time feedback from surface treatment effect evaluation indicators. This enables the algorithm to possess strong global search capabilities in the early stages of optimization and refined local search capabilities in the later stages. The surface roughness gradient guidance mechanism provides particles with a clear search direction, avoiding the blind search problem of traditional algorithms.

[0061] This invention constructs a high-fidelity digital twin model of the bolt surface, providing the optimization algorithm with precise surface state information and process parameter mapping relationships, enabling the algorithm to make optimization decisions based on a realistic physical model. The digital feature matrix integrates multi-dimensional data such as geometric information, surface texture, and roughness distribution, providing the algorithm with a comprehensive optimization basis and significantly improving optimization accuracy. The establishment of the mapping function allows the algorithm to accurately predict the surface treatment effect under different combinations of process parameters, achieving foresight and predictability in the optimization process. Attached Figure Description

[0062] Figure 1 This is a flowchart illustrating a method for optimizing bolt surface treatment process parameters in a digital twin environment, according to an embodiment of the present invention.

[0063] Figure 2 Comparison of surface roughness before and after optimization of bolt surface treatment process parameters in a digital twin environment. Detailed Implementation

[0064] The present invention will be further described below with reference to the accompanying drawings, but this is not intended to limit the present invention in any way. Any modifications or substitutions made based on the teachings of the present invention shall fall within the protection scope of the present invention.

[0065] Example 1: A method for optimizing bolt surface treatment process parameters in a digital twin environment, such as... Figure 1 As shown, it includes the following steps:

[0066] S1: Collect multi-dimensional information about the bolt surface; based on the multi-dimensional information about the bolt surface, construct a high-fidelity digital twin model to obtain the digital feature matrix of the bolt surface:

[0067] S11: Collect multi-dimensional information about the bolt surface:

[0068] Multi-dimensional information of the bolt surface was acquired using a laser scanner, a high-resolution industrial camera, and a surface roughness measuring instrument. This multi-dimensional information included geometric point cloud data of the bolt surface. Surface texture images and roughness distribution data Among them, geometric point cloud data Record the three-dimensional coordinates of each spatial position on the bolt surface , The horizontal coordinate representing spatial location. The vertical coordinate represents the spatial location. The height coordinates represent spatial locations;

[0069] S12: Generate a high-fidelity digital twin model:

[0070] The multi-dimensional information of the bolt surface collected in step S11 is spatially registered and fused to construct a high-fidelity digital twin model; the construction process includes: firstly, using the iterative nearest point algorithm to process the geometric point cloud data. Surface texture images and roughness distribution data Spatial alignment is performed, and then a digital feature matrix of the bolt surface is constructed. Each element of this digitized feature matrix It includes the three-dimensional coordinates of each spatial location on the bolt surface, the pixel values ​​of the three-dimensional coordinates of the surface texture image, and the data values ​​of the three-dimensional coordinates of the roughness distribution data. This represents the row index in the digitized feature matrix. This represents the column index in the digitized feature matrix.

[0071] The advantage of this step lies in establishing a high-precision data foundation for optimizing bolt surface treatment process parameters. Through multi-sensor synchronous acquisition technology, comprehensive information about the bolt surface can be obtained, including geometric information, surface texture, roughness distribution, and other data from multiple dimensions. Compared to traditional single-sensor acquisition methods, this significantly improves the completeness and accuracy of the data.

[0072] This step achieves high-precision spatial registration of data from different sensors through an iterative nearest-point algorithm, solving the coordinate system unification problem in multi-source data fusion and ensuring the accuracy of data fusion. The constructed digital feature matrix organically integrates geometric information, surface texture, and roughness distribution into a unified data structure, forming a complete digital description of the bolt surface.

[0073] S2: Based on the digital feature matrix, establish a mapping function for the influence of surface treatment process parameters on the surface condition of bolts; and construct a constraint space for surface treatment process parameters;

[0074] S21: Surface treatment process parameters for collection and classification:

[0075] Key process parameters in the bolt surface treatment process are collected and classified to obtain a set of process parameters. ,in, This indicates the parameters for sandblasting, including sandblasting pressure and sandblasting time. This indicates the chemical processing parameters, including the chemical solution concentration and the chemical processing temperature; This indicates the heat treatment parameters, including heating temperature and holding time; This indicates the coating processing parameters, including coating thickness and curing time; each process parameter has its corresponding numerical range.

[0076] S22: Establishment of the mapping relationship between surface treatment process parameters and bolt surface condition:

[0077] By analyzing the digital feature matrix Set of elements and surface treatment process parameters To establish the correlation between them, a mapping function is established to show the influence of surface treatment process parameters on the surface condition of bolts. The mapping function The data describes the effect of different combinations of process parameters on the surface roughness distribution of bolts. and surface texture images The quantitative influence relationship was determined; the mapping coefficient was calculated using multiple regression analysis. ,in This represents the sensitivity coefficient of each surface treatment process parameter to changes in surface condition;

[0078] S23: Constructing the constraint space for surface treatment process parameters :

[0079] Determine the feasible domain boundary for each surface treatment process parameter, where the constraint range for the sandblasting pressure value is: ,in Indicates the minimum effective sandblasting pressure. This indicates the maximum safe sandblasting pressure; the sandblasting time is within a specified range. ,in Indicates the minimum sandblasting time. This indicates the maximum blasting time; the constraint range for chemical solution concentration is... ,in Indicates the minimum effective concentration. This indicates the maximum safe concentration; the constraint range for chemical treatment temperature is... ,in Indicates the minimum reaction temperature. Indicates the maximum reaction temperature; the heating temperature is within a specified range. ,in Indicates the minimum processing temperature. Indicates the highest temperature the material can withstand; the holding time is within the specified range. ,in Indicates the shortest heat preservation time. Indicates the maximum heat preservation time; the coating thickness is within a specified range. ,in Indicates the minimum effective coating thickness. This indicates the maximum permissible coating thickness; the curing time is within a specified range. ,in Indicates the shortest curing time. This indicates the longest curing time.

[0080] The advantage of this step lies in establishing a precise quantitative relationship model between bolt surface treatment process parameters and surface condition. By systematically collecting and classifying key process parameters in the surface treatment process, this step constructs a parameter set covering the entire process flow, including sandblasting, chemical treatment, heat treatment, and coating treatment. This achieves standardized management and unified description of process parameters, providing a complete parameter space foundation for subsequent optimization.

[0081] This step employs a multiple regression analysis method to establish a mapping function that accurately describes the quantitative impact of process parameter combinations on bolt surface condition, overcoming the problems of vague and difficult-to-quantify parameter influence relationships in traditional empirical methods. By calculating the sensitivity coefficients of each process parameter to changes in surface condition, this step achieves a quantitative assessment of the importance of process parameters, providing a scientific basis for parameter weight allocation in the optimization process.

[0082] S3: Based on the mapping function, design a multi-objective adaptive particle swarm optimization algorithm that integrates bolt surface quality prediction:

[0083] S31: Constructing a surface roughness gradient guidance mechanism:

[0084] Based on digital feature matrix Roughness distribution data in Calculate the surface roughness gradient field of the bolt. Surface roughness gradient field Each spatial location Corresponding to a gradient vector ,in Indicates the gradient vector at directional components, Indicates the gradient vector at directional components, Indicates the gradient vector at Components of direction; based on gradient vector Calculate the particle guiding vector The particle guiding vector The optimal direction for improving surface quality is indicated by the following formula:

[0085] ;

[0086] in, This represents the basic guiding strength coefficient, which is 0.3 in this embodiment; Represents the gradient vector The modulus length; when surface roughness data is subject to noise interference or gradient calculation in the boundary region is difficult, a gradient estimation alternative based on wavelet denoising is adopted, using wavelet transform. Filtering out high-frequency noise and combining radial basis function interpolation Perform boundary gradient extrapolation, where Represents the wavelet transform coefficients. Indicates the scale parameter. Indicates the translation parameter. This represents the gradient value in the boundary region. Indicates the interpolation weight coefficients. This indicates the number of data points involved in the interpolation. Represents radial basis functions. Represents the current position vector. This represents the location vector of known data points.

[0087] S32: Design a dynamic inertia weight adjustment strategy:

[0088] The surface treatment effect evaluation index is calculated by comparing the roughness changes of bolts before and after surface treatment. Root mean square value of initial surface roughness of bolt By analyzing the digital feature matrix Medium roughness distribution data The statistical analysis yielded the following calculation formula: ,in Indicates the first Roughness value of each sampling point This represents the mean roughness of all sampled points. Indicates the total number of sampling points; the root mean square value of bolt surface roughness under the current combination of surface treatment process parameters. Through mapping function Predicted and calculated based on the process parameter values ​​corresponding to the current particle position; surface treatment effect evaluation index. The calculation formula is:

[0089] ;

[0090] Define the current iteration number as The maximum number of iterations is In this embodiment, the value is 500, representing the dynamic inertia weight. The calculation formula is:

[0091] ;

[0092] in, This represents the initial inertia weight, which is 0.9 in this embodiment. This indicates the termination inertia weight, which is 0.4 in this embodiment. This indicates an adaptive adjustment index, which is 2.0 in this embodiment. This represents the feedback coefficient, which is 0.1 in this embodiment;

[0093] S33: Solving multi-objective collaborative optimization:

[0094] Using the particle guidance vector from step S31 and the dynamic inertia weights in step S32 An improved particle velocity update formula was constructed.

[0095] First, calculate the values ​​of three objective functions, including the surface quality objective. Process efficiency target Processing cost target Surface quality target The standard deviation of bolt surface roughness is calculated using the following formula:

[0096] ;

[0097] in, This represents the standard deviation of the bolt surface roughness;

[0098] Process efficiency target The formula is calculated by taking the reciprocal of the total processing time:

[0099] ;

[0100] in, This represents the total time for all surface treatment processes, including sandblasting time, chemical treatment time, heat treatment time, and coating time.

[0101] Processing cost target The formula for calculating using the reciprocal of unit cost is:

[0102] ;

[0103] in, Indicates material cost, Indicates energy cost, Represent labor costs; construct a multi-objective optimization function. The expression is:

[0104] ;

[0105] in, The weighting coefficient for the surface quality target is 0.5 in this embodiment. The weighting coefficient representing the process efficiency target is 0.3 in this embodiment. The weighting coefficient represents the processing cost objective; in this embodiment, it is 0.2. Multi-objective optimization function. The calculation results are used as particle fitness values ​​to evaluate the merits of the current particle position and to determine the individual's historical best position. and the global historical best position The update provides a basis for judgment;

[0106] The improved particle velocity update formula is as follows:

[0107] ;

[0108] in, Indicates the particle in the first... Speed ​​at the next iteration Indicates the particle in the first... Speed ​​at the next iteration This represents the individual learning factor, which is 2.0 in this embodiment. This represents the global learning factor, which is 2.0 in this embodiment. This represents the gradient guiding factor, which is 0.5 in this embodiment. and Represents a random number within the interval [0,1]. Indicates the particle in the first... Position at the next iteration; individual's historical best position and the global historical best position Both are based on multi-objective optimization functions The fitness value is selected and updated to ensure that the particle velocity update direction points to the solution space region with higher fitness.

[0109] The advantage of this step lies in the design of an intelligent multi-objective optimization algorithm that integrates bolt surface quality prediction, significantly improving the accuracy and efficiency of process parameter optimization. By constructing a surface roughness gradient guidance mechanism, this step can provide a clear search direction for the optimization algorithm based on the actual state information of the bolt surface, avoiding the blind search problem in traditional particle swarm optimization algorithms, making the optimization process more targeted and efficient.

[0110] The dynamic inertia weight adjustment strategy designed in this step has adaptive characteristics, enabling it to dynamically adjust the algorithm's search behavior based on real-time feedback from the surface treatment effect. Compared to the traditional fixed inertia weight method, this dynamic adjustment mechanism gives the algorithm strong global search capability in the early stages of optimization and refined local search capability in the later stages, effectively balancing the algorithm's exploration and development capabilities, and improving convergence speed and solution quality.

[0111] The multi-objective collaborative optimization framework constructed in this step simultaneously considers three key objectives: surface quality, process efficiency, and processing cost, overcoming the limitations of traditional single-objective optimization methods that cannot simultaneously address multiple needs. By appropriately setting weight coefficients, this step can flexibly adjust the importance of each objective according to actual engineering requirements, achieving personalized customization of process parameter optimization and meeting diverse needs in different application scenarios.

[0112] S4: Based on a multi-objective adaptive particle swarm optimization algorithm, iterative solutions are performed within the surface treatment process parameter constraint space to obtain a set of candidate process parameter combinations.

[0113] S41: Initialize the particle swarm:

[0114] Within the constraints of surface treatment process parameters An initial particle swarm is randomly generated internally, and the position of each particle is... This represents a combination of surface treatment process parameters, including sandblasting parameters, chemical treatment parameters, heat treatment parameters, and coating treatment parameters. The particle swarm size is set to... In this embodiment, the value is 50; the velocity of each particle is initialized. Let the zero vector be the individual's historical best position. Equal to the initial position, the global historical best position Determined by comparing the initial fitness values ​​of all particles;

[0115] S42: Iterative optimization solution:

[0116] For each iteration Perform the following operations: First, based on the current particle position Through mapping function Calculate the predicted surface roughness value of the bolts under the current combination of surface treatment process parameters, and then calculate the surface treatment effect evaluation index. ; Calculate the dynamic inertia weight according to step S3 and particle guiding vector ; Calculate the values ​​of the three objective functions , , And through multi-objective optimization function Calculate the fitness value for each particle; the fitness value is used to evaluate the particle's performance in the target space consisting of surface quality, process efficiency, and processing cost targets, and serves as the basis for subsequently updating the individual's historical best position and the global historical best position. If the current particle's fitness value is better than its historical best fitness value, then the individual's historical best position is updated. If there is a particle in the current population with a fitness value better than the global historical best fitness value, then update the global historical best position. Update the individual's historical best position. and the global historical best position The new velocity is calculated based on the improved particle velocity update formula. and update particle positions. For spaces outside the constraint space The particle positions are processed to define the boundaries;

[0117] An anomaly monitoring mechanism is implemented during the iteration process. When more than 30% of the particles are detected to have gone out of the constraint space, the particle swarm repair strategy is initiated: the out-of-bounds particles are pulled back into the feasible region using the mirror reflection method, while the elite retention strategy is adopted to keep the current generation's best 10% of particles from being modified.

[0118] S43: Generate the Pareto optimal solution set:

[0119] When the maximum number of iterations is reached The iteration process terminates when a convergence condition is met, wherein the improvement of the global optimal solution over 10 consecutive iterations is less than 0.001; a Pareto optimal solution set is selected from the non-dominated solutions in all iterations. The non-dominated solution refers to a solution that is not completely dominated by other solutions on the three objective functions; the Pareto optimal solution set All solutions are used as combinations of process parameters to form a set of candidate process parameter combinations. Each candidate process parameter combination includes complete surface treatment process parameter settings, and the surface roughness before and after process parameter optimization is shown in the example. Figure 2 As shown.

[0120] The advantage of this step lies in achieving efficient execution of the multi-objective adaptive particle swarm optimization algorithm and accurate generation of the Pareto optimal solution set. The iterative optimization process designed in this step is highly systematic and complete. In each iteration, not only are the surface treatment effect evaluation index and multi-objective function values ​​calculated, but the dynamic inertia weights and particle guidance vectors are also updated in real time, forming a closed-loop feedback optimization mechanism. This real-time feedback and dynamic adjustment strategy enables the algorithm to adaptively adjust the search strategy according to the optimization progress, significantly improving optimization efficiency and solution quality.

[0121] Example 2: This invention also discloses a bolt surface treatment process parameter optimization system in a digital twin environment, comprising the following four modules:

[0122] Digitalization module: Collects multi-dimensional information about the bolt surface; Based on the multi-dimensional information about the bolt surface, constructs a high-fidelity digital twin model to obtain the digital feature matrix of the bolt surface;

[0123] Constraint Module: Based on the digital feature matrix, a mapping function is established to show the influence of surface treatment process parameters on the surface condition of bolts; and a constraint space for surface treatment process parameters is constructed.

[0124] Multi-objective optimization module: Based on the mapping function, a multi-objective adaptive particle swarm optimization algorithm that integrates bolt surface quality prediction is designed;

[0125] Iterative solution module: Based on the multi-objective adaptive particle swarm optimization algorithm, iterative solution is performed within the constraint space of surface treatment process parameters to obtain a set of candidate process parameter combinations.

[0126] It should be noted that the sequence numbers of the above embodiments of the present invention are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0127] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0128] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for optimizing bolt surface treatment process parameters in a digital twin environment, characterized in that, Includes the following steps: S1: Collect multi-dimensional information about the bolt surface; based on the multi-dimensional information about the bolt surface, construct a high-fidelity digital twin model to obtain the digital feature matrix of the bolt surface; S2: Based on the digital feature matrix, establish a mapping function for the influence of surface treatment process parameters on the surface condition of bolts; and construct a constraint space for surface treatment process parameters. S3: Based on the mapping function, design a multi-objective adaptive particle swarm optimization algorithm that integrates bolt surface quality prediction; S4: Based on the multi-objective adaptive particle swarm optimization algorithm, iterative solutions are performed within the constraint space of surface treatment process parameters to obtain a set of candidate process parameter combinations; Step S1 includes: S11: Collect multi-dimensional information about the bolt surface: Multi-dimensional information of the bolt surface was acquired using a laser scanner, a high-resolution industrial camera, and a surface roughness measuring instrument. This multi-dimensional information included geometric point cloud data of the bolt surface. Surface texture images and roughness distribution data Among them, geometric point cloud data Record the three-dimensional coordinates of each spatial position on the bolt surface , The horizontal coordinate representing spatial location. The vertical coordinate representing spatial location. The height coordinates represent spatial location; S12: Generate a high-fidelity digital twin model: The multi-dimensional information of the bolt surface collected in step S11 is spatially registered and fused to construct a high-fidelity digital twin model; the construction process includes: firstly, using the iterative nearest point algorithm to process the geometric point cloud data. Surface texture images and roughness distribution data Spatial alignment is performed, and then a digital feature matrix of the bolt surface is constructed. Each element of this digitized feature matrix It includes the three-dimensional coordinates of each spatial location on the bolt surface, the pixel values ​​of the three-dimensional coordinates of the surface texture image, and the data values ​​of the three-dimensional coordinates of the roughness distribution data. This represents the row index in the digitized feature matrix. Indicates the column index in the digitized feature matrix; Step S2 includes: S21: Surface treatment process parameters for collection and classification: Key process parameters in the bolt surface treatment process are collected and classified to obtain a set of surface treatment process parameters. ,in, This indicates the parameters for sandblasting, including sandblasting pressure and sandblasting time. This indicates the chemical processing parameters, including the chemical solution concentration and the chemical processing temperature; This indicates the heat treatment parameters, including heating temperature and holding time; This indicates the coating treatment parameters, including coating thickness and curing time; each surface treatment process parameter has its corresponding numerical range. S22: Establishment of the mapping relationship between surface treatment process parameters and bolt surface condition: By analyzing the digital feature matrix Set of elements and surface treatment process parameters To establish the correlation between them, a mapping function is established to show the influence of surface treatment process parameters on the surface condition of bolts. The mapping function The data describes the effect of different combinations of process parameters on the surface roughness distribution of bolts. and surface texture images The quantitative influence relationship was determined; the mapping coefficient was calculated using multiple regression analysis. ,in This represents the sensitivity coefficient of each surface treatment process parameter to changes in surface condition; S23: Constructing the constraint space for surface treatment process parameters : Determine the feasible domain boundary for each surface treatment process parameter, where the constraint range for the sandblasting pressure value is: ,in Indicates the minimum effective sandblasting pressure. This indicates the maximum safe sandblasting pressure; the sandblasting time is within a specified range. ,in Indicates the minimum sandblasting time. This indicates the maximum blasting time; the constraint range for chemical solution concentration is... ,in Indicates the minimum effective concentration. This indicates the maximum safe concentration; the constraint range for chemical treatment temperature is... ,in Indicates the minimum reaction temperature. Indicates the maximum reaction temperature; the heating temperature is within a specified range. ,in Indicates the minimum processing temperature. Indicates the highest temperature the material can withstand; the holding time is within the specified range. ,in Indicates the shortest heat preservation time. Indicates the maximum heat preservation time; the coating thickness is within a specified range. ,in Indicates the minimum effective coating thickness. This indicates the maximum permissible coating thickness; the curing time is within a specified range. ,in Indicates the shortest curing time. Indicates the longest curing time; Step S3 includes: S31: Constructing a surface roughness gradient guidance mechanism: Based on digital feature matrix Roughness distribution data in Calculate the surface roughness gradient field of the bolt. Surface roughness gradient field Each spatial location Corresponding to a gradient vector ,in Indicates the gradient vector at directional components, Indicates the gradient vector at directional components, Indicates the gradient vector at Components of direction; based on gradient vector Calculate the particle guiding vector The particle guiding vector Pointing to the optimal direction for improving surface quality; S32: Design a dynamic inertia weight adjustment strategy: using a mapping function The root mean square (RMS) value of bolt surface roughness is predicted and calculated under the current combination of surface treatment process parameters. The surface treatment effect evaluation index is then calculated by comparing the changes in the RMS value before and after surface treatment. And adjust the dynamic inertia weight according to the surface treatment effect evaluation index; S33: Solving multi-objective collaborative optimization: Constructing a multi-objective optimization function that includes surface quality objective, process efficiency objective and processing cost objective, and constructing an improved particle velocity update formula based on particle guiding vector and dynamic inertia weight; Step S32 includes: The surface treatment effect evaluation index is calculated by comparing the changes in the root mean square roughness value of the bolts before and after surface treatment. ; Root mean square value of initial surface roughness of bolt By analyzing the digital feature matrix Medium roughness distribution data The root mean square value of bolt surface roughness was obtained through statistical analysis; under the current combination of surface treatment process parameters. Through mapping function Predicted and calculated based on the process parameter values ​​corresponding to the current particle position; surface treatment effect evaluation index. The calculation formula is: ； Define the current iteration number as The maximum number of iterations is Dynamic inertia weight The calculation formula is: ； in, Indicates the initial inertia weight. Indicates the termination of inertia weight. Indicates an adaptive adjustment index. Indicates the effect feedback coefficient; Step S33 includes: Calculate three objective function values, including the surface quality objective. Process efficiency target Processing cost target Surface quality target Calculated using the standard deviation of bolt surface roughness; process efficiency target. Calculate the processing cost target by taking the reciprocal of the total processing time. Construct a multi-objective optimization function by calculating the reciprocal of the unit cost. The expression is: ； in, The weighting coefficients representing surface quality targets. The weighting coefficients representing the process efficiency target. Weighting coefficients representing the cost objective; multi-objective optimization function The calculation results are used as particle fitness values ​​to evaluate the merits of the current particle position and to determine the individual's historical best position. and the global historical best position The update provides a basis for judgment; Based on the particle guidance vector in step S31 and the dynamic inertia weights in step S32 An improved particle velocity update formula is constructed, which is as follows: ； in, Indicates the particle at the 1st Speed ​​at the next iteration Indicates the particle at the 1st Speed ​​at the next iteration Represents individual learning factors. Represents the global learning factor. Indicates the gradient guiding factor. and Represents a random number within the interval [0,1]. Indicates the particle at the 1st Position at the next iteration; Individual's historical best position and the global historical best position Both are based on multi-objective optimization functions The fitness value is selected and updated to ensure that the particle velocity update direction points to the solution space region with higher fitness.

2. The method for optimizing bolt surface treatment process parameters in a digital twin environment according to claim 1, characterized in that, Step S4 includes: S41: Initialize the particle swarm: Within the constraints of surface treatment process parameters An initial particle swarm is randomly generated internally, and the position of each particle is... This represents a combination of surface treatment process parameters, including sandblasting parameters, chemical treatment parameters, heat treatment parameters, and coating treatment parameters; the particle swarm size is set to... Initialize the velocity of each particle. Let the zero vector be the individual's historical best position. Equal to the initial position, the global historical best position Determined by comparing the initial fitness values ​​of all particles; S42: Iterative optimization solution: For each iteration Perform the following operations: First, based on the current particle position Through mapping function Calculate the predicted surface roughness value of the bolts under the current combination of surface treatment process parameters, and then calculate the surface treatment effect evaluation index. ; Calculate the dynamic inertia weight according to step S3 and particle guiding vector ; Calculate the values ​​of the three objective functions , , And through multi-objective optimization function Calculate the fitness value for each particle; the fitness value is used to evaluate the particle's performance in the target space consisting of surface quality, process efficiency, and processing cost targets, and serves as the basis for subsequently updating the individual's historical best position and the global historical best position. If the current particle's fitness value is better than its historical best fitness value, then the individual's historical best position is updated. If there is a particle in the current population with a fitness value better than the global historical best fitness value, then update the global historical best position. Update the individual's historical best position. and the global historical best position The new velocity is calculated based on the improved particle velocity update formula. and update particle positions. For surfaces exceeding the constraints of the surface treatment process parameters The particle positions are processed to define the boundaries; S43: Generate the Pareto optimal solution set: When the maximum number of iterations is reached The iteration process terminates when a convergence condition is met, wherein the improvement of the global optimal solution over 10 consecutive iterations is less than 0.001; a Pareto optimal solution set is selected from the non-dominated solutions in all iterations. The non-dominated solution refers to a solution that is not completely dominated by other solutions at the three objective function values; the Pareto optimal solution set All solutions are used as combinations of process parameters to form a set of candidate process parameter combinations. Each candidate process parameter combination contains complete surface treatment process parameter settings.

3. A bolt surface treatment process parameter optimization system in a digital twin environment, characterized in that, include: Digital module: Collects multi-dimensional information about the bolt surface; Based on multi-dimensional information of the bolt surface, a high-fidelity digital twin model is constructed to obtain the digital feature matrix of the bolt surface; Constraint module: Based on the digital feature matrix, establish a mapping function for the influence of surface treatment process parameters on the surface condition of bolts; And construct a constraint space for surface treatment process parameters; Multi-objective optimization module: Based on the mapping function, a multi-objective adaptive particle swarm optimization algorithm that integrates bolt surface quality prediction is designed; Iterative solution module: Based on the multi-objective adaptive particle swarm optimization algorithm, iterative solution is performed within the constraint space of surface treatment process parameters to obtain a set of candidate process parameter combinations; To achieve the method for optimizing bolt surface treatment process parameters in a digital twin environment as described in any one of claims 1-2.

Images