Micro spring washer high-precision manufacturing method using particle swarm optimization technology

Abstract

The invention discloses a high-precision manufacturing method of a miniature spring washer by using a particle swarm optimization technology. The method comprises the steps that S1, manufacturing target parameters are set, and a multi-target optimization function is established; s2, initializing a particle swarm and distributing initial manufacturing process parameters; s3, in an optimization iteration process, triggering a variation-driven particle swarm self-organizing reconstruction mechanism, and recombining a particle swarm topological structure to break through local optimum; s4, analyzing and optimizing a plurality of manufacturing process paths by applying a manufacturing path complementary optimization technology; s5, applying an optimization result to an actual manufacturing process, and collecting and feeding back key manufacturing data in real time; s6, dynamically adjusting the search behavior of the particle swarm according to the real-time data; s7, after batch manufacturing is completed, manufacturing data are comprehensively analyzed, and an optimization strategy is updated; and S8, recording and archiving the optimized manufacturing path and process parameters, and continuously improving. The manufacturing precision and the production efficiency of the miniature spring washer are remarkably improved.

Description

Technical Field

[0001] The invention relates to the technical field of particle swarm optimization, and in particular to a high-precision manufacturing method for a micro spring washer using the particle swarm optimization technology. Background Art

[0002] First of all, as a key component widely used in mechanical equipment, the manufacturing accuracy and performance of micro spring washers directly affect the stability and service life of the entire equipment. Traditional micro spring washer manufacturing methods mainly rely on precision machining technology, which usually requires multiple experiments and experience accumulation to determine the optimal process parameters. However, due to the small size of micro spring washers, the manufacturing process is easily affected by factors such as processing errors, material non-uniformity, and equipment stability, resulting in insufficient manufacturing accuracy and poor product consistency. In addition, traditional manufacturing methods often find it difficult to take into account the requirements of high precision and high efficiency when faced with complex shapes or special materials, which further limits the application of micro spring washers in the field of high-end manufacturing.

[0003] Secondly, the existing manufacturing process of micro spring washers usually relies on fixed process parameters for production. Although this method can ensure product consistency to a certain extent, it lacks dynamic adjustment capabilities and is difficult to cope with the influence of uncertain factors such as material characteristics of different batches and changes in equipment status. For example, when dimensional deviations or surface quality do not meet standards during the production process, traditional manufacturing systems can usually only be adjusted through manual intervention, which not only increases the production cycle, but may also lead to an increase in the defective rate. In addition, traditional process methods that lack flexible adjustment and optimization capabilities often find it difficult to improve production efficiency while ensuring accuracy when faced with micro spring washers of different materials and complex shapes, resulting in an increase in production costs.

[0004] Third, particle swarm optimization (PSO) technology, as an optimization algorithm based on swarm intelligence, has been widely used in various complex optimization problems. However, there are few cases in the prior art where PSO technology is applied to the manufacturing process of micro spring washers. The local optimal problem of the traditional PSO algorithm in the high-dimensional search space and the sensitivity of parameter selection limit its application in manufacturing process optimization. Especially when faced with multi-objective optimization requirements, the PSO algorithm is prone to fall into the local optimal solution and it is difficult to find the global optimal solution. At the same time, since the manufacturing of micro spring washers involves the coordinated optimization of multiple parameters, the existing single optimization method is difficult to take into account manufacturing accuracy, efficiency and cost at the same time, which makes how to apply PSO technology to dynamically optimize process parameters in the manufacturing process a problem that needs to be solved urgently.

[0005] In summary, the existing technology has the following major defects in the manufacturing process of micro spring washers: 1. Insufficient manufacturing accuracy and poor production consistency. Especially when faced with complex shapes and special materials, traditional manufacturing processes are difficult to balance accuracy and efficiency; 2. Lack of dynamic adjustment and optimization capabilities. When faced with uncertain factors and batch differences, traditional process methods can only be adjusted through manual intervention, which affects production efficiency and product quality; 3. The application of existing optimization algorithms in manufacturing processes is limited, especially the PSO algorithm is prone to fall into local optimality in high-dimensional optimization space, and it is difficult to effectively apply it to the manufacturing process of multi-objective optimization. Therefore, how to provide a high-precision manufacturing method for micro spring washers using particle swarm optimization technology is a problem that technicians in this field urgently need to solve. Summary of the invention

[0006] One object of the present invention is to propose a high-precision manufacturing method for micro spring washers using particle swarm optimization technology. The present invention sets manufacturing target parameters and establishes a multi-objective optimization function, combines particle swarm optimization technology to dynamically adjust process parameters, and describes in detail the optimization of manufacturing process through the self-organization reconstruction mechanism of particle swarm driven by mutation and the complementary optimization technology of manufacturing path. The method has the advantages of high manufacturing precision, good product consistency and high production efficiency.

[0007] A high-precision manufacturing method for a micro spring washer using particle swarm optimization technology according to an embodiment of the present invention comprises the following steps:

[0008] S1. Setting manufacturing target parameters of the micro spring washer, wherein the target parameters include specific dimensional tolerance, elastic modulus and surface roughness, and establishing a multi-objective optimization function based on these targets;

[0009] S2, initializing the particle swarm and assigning initial manufacturing process parameters to each particle, wherein the process parameters include precise cutting speed, feed rate, processing temperature and processing pressure, and the initial particle swarm is distributed in the search space according to a preset multi-objective optimization function;

[0010] S3. During the optimization iteration of the particle swarm, the changing trend of the global optimal solution is monitored. When it is detected that the global optimal solution has not been significantly improved in multiple iterations, the mutation-driven particle swarm self-organization reconstruction mechanism is started to impose random mutations on the speed and position parameters of some particles to reorganize the topological structure of the particle swarm and break the local optimal state.

[0011] S4. In the reconstructed particle swarm, the manufacturing path complementary optimization technology is applied to analyze and select multiple possible manufacturing process paths, optimize each path separately, and perform alternating optimization and data integration between paths;

[0012] S5. Apply the process path determined by the complementary manufacturing path optimization method and the particle swarm optimization results to the actual manufacturing process of the micro spring washer. During the process, key manufacturing data including dimensional accuracy, surface quality and elastic coefficient are collected in real time, and these data are fed back to the particle swarm for further adjustment of the optimization path and direction.

[0013] S6. Based on the real-time collected manufacturing data, dynamically adjust the search behavior of the particle swarm and the configuration of manufacturing process parameters;

[0014] S7. After completing batch manufacturing, comprehensively analyze the manufacturing data of each process path, adjust and update the initialization parameters of the particle swarm and the path optimization strategy, and generate an optimization plan for subsequent batch production;

[0015] S8. Record and archive the optimized manufacturing path and process parameters, and conduct continuous improvement and optimization by analyzing the performance of variation-driven particle swarm self-organization reconstruction and manufacturing path complementary optimization technology in actual manufacturing.

[0016] Optionally, S2 includes the following steps:

[0017] S21, determine the size N of the particle group, each particle corresponds to a manufacturing process parameter combination, the process parameter combination includes a cutting speed v c , feed rate f, processing temperature T and processing pressure P;

[0018] S22. Set the initial position for each particle. The initial position represents a specific process parameter combination x i ,These initial positions are distributed in the search space according to the aforementioned multi-objective optimization function;

[0019] S23, setting an initial speed for each particle. The initial speed determines the search dynamics of the particle in the initial stage. The speed value is distributed within a reasonable range, reflecting the initial strength of the process parameter adjustment;

[0020] S24, calculating the initial fitness value of each particle, using a multi-objective optimization function to evaluate the quality of the process parameter combination corresponding to the position of each particle, and using the fitness value as a reference for subsequent optimization;

[0021] S25. Determine the local optimal position P of each particle based on the initial fitness value best and the global optimal position G of the particle swarm best ,These optimal positions will be used as reference points in the subsequent optimization process;

[0022] S26, updating the process parameter combination of each particle, adjusting its position in the search space according to the current optimal position and the initial position of the particle, forming a new process parameter combination, and providing a new starting point for subsequent optimization iterations;

[0023] S27. Perform preliminary verification on the updated particle swarm to verify the feasibility of the new process parameter combination in the manufacturing process and provide initial conditions for the subsequent particle swarm optimization iterative process.

[0024] Optionally, S3 includes the following steps:

[0025] S31, calculate the fitness value f of each particle i (t) To evaluate the pros and cons of the current process parameter combination, record the changing trend of the global optimal solution, and provide a basis for subsequent judgment on whether the particle swarm falls into a local optimal state:

[0026]

[0027] Among them, w j is the weight of the target parameter, x ij is the jth parameter value of the ith particle, is the ideal value of the target parameter, α j To adjust the parameters;

[0028] S32, define the change rate ΔG(t) of the global optimal solution to monitor the convergence speed, and adjust the search intensity in combination with the time weighting factor γ(t) to detect whether the particle swarm falls into a local optimal state during the optimization process:

[0029]

[0030] Among them, γ(t) is the time weighting factor, which gradually decreases with the increase of the number of iterations t, and β controls the deceleration slope of the factor, T max is the maximum number of iterations;

[0031] S33, triggering the mutation-driven self-organization reconstruction mechanism, by introducing the Lévy flight mechanism for mutation operation, optimizing the jumping of particles in the search space, so as to expand the search range and break the local optimal limitation:

[0032]

[0033] Among them, δ is the step size control parameter, u and v are independent normally distributed random variables, and λ is the Lévy distribution index, which ranges from 1 to 2;

[0034] S34, apply the topological reconstruction strategy of adaptive coupling mapping to readjust the connection mode and distribution structure of the mutated particle swarm:

[0035]

[0036] Where, ∈ is the coupling strength weight, is the fitness gradient of the ith particle, H(t) is the diversity entropy of the particle swarm, H max is the initial entropy value;

[0037] S35. Recalculate the fitness value after mutation The fitness value is corrected through the mutation energy adjustment mechanism:

[0038]

[0039] Among them, κ is the adjustment parameter, δE i is the energy change of particle i during the mutation process, E max is the maximum energy reference value;

[0040] S36. Calculate the health metric Q(t) of the particle swarm to evaluate the diversity and exploration ability of the particle swarm after mutation, and provide a reference for the dynamic adjustment of the subsequent mutation intensity and topological structure:

[0041]

[0042] Among them, δE i is the variation energy change of particle i, is the center of mass of the particle group;

[0043] S37. Based on the health metric Q(t), the mutation intensity δ and the Lévy flight index λ are dynamically adjusted to gradually optimize the search behavior of the particle swarm, and finally reach the global optimal solution or meet the predetermined stop condition; the formula for dynamic adjustment is:

[0044] δ(t)=δ0·exp(-ξ·Q(t));

[0045]

[0046] Among them, δ0 and ξ are the initial step size and adjustment rate, and η is the exponential adjustment factor.

[0047] Optionally, S4 includes the following steps:

[0048] S41. In the reconstructed particle swarm, all possible manufacturing process paths of the micro spring washer are analyzed and the initial process parameter combination p of each path is defined. k , and calculate the initial fitness value f of each path k (t);

[0049] S42, capture the synergy and competition relationship between paths, establish the path complementary effect matrix Θ, whose elements θ klUsed to represent the path P k and path P l The complementary and inhibitory relationship between them:

[0050]

[0051] Among them, p k ·p l represents the inner product of path parameters, p k -p l ∥ represents the Euclidean distance between path parameters, σ is the smoothing factor;

[0052] S43, based on the result of the complementary effect matrix Θ, select the path with the highest initial fitness value as the reference path P base , optimize it, and the optimized parameter combination As a reference for other paths;

[0053] S44, for each non-reference path P k To optimize, combine the complementary effect information between paths and use the adaptive path optimization formula:

[0054]

[0055] Among them, λ k For path P k The adaptive step size, θ kl is the path complementarity effect coefficient, which is used to guide the adjustment of path parameters and consider the complementary and competitive relationships between paths;

[0056] S45, through multiple iterations of alternate optimization path combination, gradually optimize the final fitness value of each path And analyze the synergistic effect between each path, and select the final optimized path as the manufacturing process path combination;

[0057] S46. In the actual manufacturing process, the optimized manufacturing process path combination is applied, and the key parameters p are monitored in real time. j (t), when it is found that the parameters in the manufacturing process deviate from the target, the path complementary effect matrix Θ is used for dynamic adjustment;

[0058] S47. After manufacturing is completed, analyze the actual manufacturing data of each process path, compare the prediction results, and modify the complementary effect matrix Θ and the optimization step size λ. k .

[0059] Optionally, S5 includes the following steps:

[0060] S51. In the manufacturing process, the optimal process path combination p is determined by the complementary manufacturing path optimization method. opt, and collect key manufacturing data d(t) in real time, including dimensional accuracy d size (t), surface quality d surface (t) and elastic modulus d elastic (t);

[0061] S52, inputting the manufacturing data d(t) collected in real time into the feedback control system, calculating the deviation Δd(t) between each data and the target value, so as to identify the errors existing in the manufacturing process;

[0062] S53, dynamically adjusting the manufacturing process parameter p(t) according to the information of the deviation Δd(t) and the path complementary effect matrix Θ;

[0063] S54, monitor the changing trend of manufacturing data in real time through the feedback control system, and calculate the convergence stability σ of the manufacturing path stability (t) Evaluate the stability of the current process path, when σ stability (t) When it drops below a certain set threshold, the process path is considered to be stable;

[0064] S55. If a significant fluctuation or deviation from the target is detected in the process path during the manufacturing process, the parameters between the paths are transferred to make adjustments and update the path combination. The adjustment formula is:

[0065] p new (t) = p opt (t)+γ·(∑ l≠k θ kl ·p l (t)-p new (t));

[0066] Among them, γ is the adjustment coefficient, which is used to achieve parameter migration and dynamic correction between paths;

[0067] S56, feeding back the real-time adjusted process parameters and manufacturing data to the particle swarm optimization algorithm to update the search direction and fitness evaluation of the particle swarm;

[0068] S57. After manufacturing is completed, analyze all collected manufacturing data, compare the actual manufacturing results with the optimization prediction results, update the path complementary effect matrix Θ and the parameter adjustment model of the process path to optimize the accuracy and consistency of subsequent manufacturing batches.

[0069] Optionally, the S6 includes the following steps:

[0070] S61. Input the real-time collected manufacturing data d(t) into the particle swarm optimization algorithm, and dynamically modify the fitness value of the particle according to the change of the data. The fitness modification formula combines the nonlinear volatility of the manufacturing data and calculates the modification factor through the entropy weight method:

[0071]

[0072] Where H(Δd(t)) is the deviation data Δd(t) = d(t) - d target The information entropy reflects the complexity of data fluctuations, and α is the adjustment coefficient;

[0073] S62, based on the corrected fitness value, apply the adaptive mutation strategy to update the particle's position x i (t+1):

[0074] x i (t+1)=x i (t)+γ·(P best (t)-x i (t))+∑ j≠i β ij Δ p (t);

[0075] Among them, γ is the local search step size, Δ p (t) is the process parameter difference vector of particles i and j, and the position is adjusted in combination with the path complementarity effect;

[0076] S63. After the particle position is adjusted, a dynamic weight adjustment mechanism is introduced to combine diversity protection and global fitness to keep the particle swarm balanced between local and global searches; the weight adjustment is updated according to the sensitivity of process parameter changes:

[0077]

[0078] Among them, λ is the adjustment coefficient, σ and D(t) represent the smoothing factor of process parameters and population diversity measure respectively;

[0079] S64. Based on dynamic adjustment, implement multi-objective collaborative optimization and introduce nonlinear Pareto optimal solution set to achieve dynamic balance among multiple manufacturing objectives;

[0080] S65, monitor the fitness convergence of the particle swarm, quantify the convergence through the synergy between paths, and the convergence index calculation formula is:

[0081]

[0082] Among them, ξ is the adjustment coefficient, θ ij is the complementary effect weight between particle i and particle j;

[0083] S66, when the convergence index σ convergence (t) When the set convergence threshold is reached, the real-time data feedback is stopped and the current process parameters are locked as the final manufacturing parameter configuration;

[0084] S67. After manufacturing is completed, based on the data feedback from the manufacturing process and combined with the analysis results of the path complementary effect matrix Θ, the particle swarm optimization algorithm is adjusted in terms of strategy.

[0085] The beneficial effects of the present invention are:

[0086] The present invention proposes a high-precision manufacturing method for micro spring washers using particle swarm optimization technology. By setting manufacturing target parameters and establishing a multi-objective optimization function, the process parameters in the manufacturing process are dynamically adjusted in combination with particle swarm optimization technology, which significantly improves the manufacturing accuracy of micro spring washers. By introducing a variation-driven particle swarm self-organization reconstruction mechanism, the local optimal state in the particle swarm optimization process is broken, the ability of global optimization is improved, and the stability and consistency of the manufacturing process are effectively improved, especially when facing complex shapes and multi-objective optimization requirements.

[0087] (2) The present invention analyzes and optimizes multiple manufacturing process paths by applying the complementary optimization technology of manufacturing paths, and realizes the alternating optimization and data integration of process paths by utilizing the synergy and competition relationship between paths, making the process path selection in the manufacturing process more flexible and efficient. This method collects key manufacturing data in real time and feeds it back to the particle swarm, dynamically adjusts the search behavior and process parameter configuration of the particle swarm, effectively responds to the changes in material properties and equipment status fluctuations that may occur in the manufacturing process, and significantly improves the adaptability and accuracy of the manufacturing process.

[0088] (3) After batch manufacturing is completed, the present invention conducts a comprehensive analysis of the manufacturing data of each process path, adjusts and updates the initialization parameters of the particle swarm and the path optimization strategy, and generates an optimization plan for subsequent batch production, ensuring the consistency of the product and the continuous improvement of the process. By recording and archiving the optimized manufacturing path and process parameters, combining the performance of the variation-driven particle swarm self-organization reconstruction and the manufacturing path complementary optimization technology in actual manufacturing, continuous improvement and optimization are carried out, which significantly reduces the manufacturing cost and improves the production efficiency.

[0089] (4) The present invention realizes the full process optimization of the micro spring washer manufacturing process by innovatively combining particle swarm optimization technology with complementary manufacturing path optimization technology, which not only improves production accuracy and efficiency, but also has high flexibility and adaptability. Especially when dealing with different materials and complex manufacturing requirements, the present invention demonstrates its significant technical advantages and provides an effective solution for the field of high-precision manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0091] Figure 1 A flowchart of a high-precision manufacturing method for micro spring washers using particle swarm optimization technology proposed by the present invention;

[0092] Figure 2 A schematic diagram of the self-organizing reconstruction mechanism driven by mutation in the iterative process of particle swarm optimization proposed by the present invention;

[0093] Figure 3 This is a schematic diagram of the application of the manufacturing path complementary optimization technology proposed in the present invention. DETAILED DESCRIPTION

[0094] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0095] refer to Figure 1-3 , a high-precision manufacturing method for micro spring washers using particle swarm optimization technology, comprising the following steps:

[0096] S1. Setting manufacturing target parameters of the micro spring washer, wherein the target parameters include specific dimensional tolerance, elastic modulus and surface roughness, and establishing a multi-objective optimization function based on these targets;

[0097] S2, initializing the particle swarm and assigning initial manufacturing process parameters to each particle, wherein the process parameters include precise cutting speed, feed rate, processing temperature and processing pressure, and the initial particle swarm is distributed in the search space according to a preset multi-objective optimization function;

[0098] S3. During the optimization iteration of the particle swarm, the changing trend of the global optimal solution is monitored. When it is detected that the global optimal solution has not been significantly improved in multiple iterations, the mutation-driven particle swarm self-organization reconstruction mechanism is started to impose random mutations on the speed and position parameters of some particles to reorganize the topological structure of the particle swarm and break the local optimal state.

[0099] S4. In the reconstructed particle swarm, the manufacturing path complementary optimization technology is applied to analyze and select multiple possible manufacturing process paths, optimize each path separately, and perform alternating optimization and data integration between paths;

[0100] S5. Apply the process path determined by the complementary manufacturing path optimization method and the particle swarm optimization results to the actual manufacturing process of the micro spring washer. During the process, key manufacturing data including dimensional accuracy, surface quality and elastic coefficient are collected in real time, and these data are fed back to the particle swarm for further adjustment of the optimization path and direction.

[0101] S6. Based on the real-time collected manufacturing data, dynamically adjust the search behavior of the particle swarm and the configuration of manufacturing process parameters;

[0102] S7. After completing batch manufacturing, comprehensively analyze the manufacturing data of each process path, adjust and update the initialization parameters of the particle swarm and the path optimization strategy, and generate an optimization plan for subsequent batch production;

[0103] S8. Record and archive the optimized manufacturing path and process parameters, and conduct continuous improvement and optimization by analyzing the performance of variation-driven particle swarm self-organization reconstruction and manufacturing path complementary optimization technology in actual manufacturing.

[0104] In this implementation, S2 includes the following steps:

[0105] S21, determine the size N of the particle group, each particle corresponds to a manufacturing process parameter combination, the process parameter combination includes a cutting speed v c , feed rate f, processing temperature T and processing pressure P;

[0106] S22. Set the initial position for each particle. The initial position represents a specific process parameter combination x i ,These initial positions are distributed in the search space according to the aforementioned multi-objective optimization function;

[0107] S23, setting an initial speed for each particle. The initial speed determines the search dynamics of the particle in the initial stage. The speed value is distributed within a reasonable range, reflecting the initial strength of the process parameter adjustment;

[0108] S24, calculating the initial fitness value of each particle, using a multi-objective optimization function to evaluate the quality of the process parameter combination corresponding to the position of each particle, and using the fitness value as a reference for subsequent optimization;

[0109] S25. Determine the local optimal position P of each particle based on the initial fitness value best and the global optimal position G of the particle swarm best ,These optimal positions will be used as reference points in the subsequent optimization process;

[0110] S26, updating the process parameter combination of each particle, adjusting its position in the search space according to the current optimal position and the initial position of the particle, forming a new process parameter combination, and providing a new starting point for subsequent optimization iterations;

[0111] S27. Perform preliminary verification on the updated particle swarm to verify the feasibility of the new process parameter combination in the manufacturing process and provide initial conditions for the subsequent particle swarm optimization iterative process.

[0112] In this implementation, S3 includes the following steps:

[0113] S31, calculate the fitness value f of each particle i (t) To evaluate the pros and cons of the current process parameter combination, record the changing trend of the global optimal solution, and provide a basis for subsequent judgment on whether the particle swarm falls into a local optimal state:

[0114]

[0115] Among them, w j is the weight of the target parameter, x ij is the jth parameter value of the ith particle, is the ideal value of the target parameter, α j To adjust the parameters;

[0116] S32, define the change rate ΔG(t) of the global optimal solution to monitor the convergence speed, and adjust the search intensity in combination with the time weighting factor γ(t) to detect whether the particle swarm falls into a local optimal state during the optimization process:

[0117]

[0118] Among them, γ(t) is the time weighting factor, which gradually decreases with the increase of the number of iterations t, and β controls the deceleration slope of the factor, T max is the maximum number of iterations;

[0119] S33, triggering the mutation-driven self-organization reconstruction mechanism, by introducing the Lévy flight mechanism for mutation operation, optimizing the jumping of particles in the search space, so as to expand the search range and break the local optimal limitation:

[0120]

[0121] Among them, δ is the step size control parameter, u and v are independent normally distributed random variables, and λ is the Lévy distribution index, which ranges from 1 to 2;

[0122] S34, apply the topological reconstruction strategy of adaptive coupling mapping to readjust the connection mode and distribution structure of the mutated particle swarm:

[0123]

[0124] Where, ∈ is the coupling strength weight, is the fitness gradient of the ith particle, H(t) is the diversity entropy of the particle swarm, H max is the initial entropy value;

[0125] S35. Recalculate the fitness value after mutation The fitness value is corrected through the mutation energy adjustment mechanism:

[0126]

[0127] Among them, κ is the adjustment parameter, δE i is the energy change of particle i during the mutation process, E max is the maximum energy reference value;

[0128] S36. Calculate the health metric Q(t) of the particle swarm to evaluate the diversity and exploration ability of the particle swarm after mutation, and provide a reference for the dynamic adjustment of the subsequent mutation intensity and topological structure:

[0129]

[0130] Among them, δE i is the variation energy change of particle i, is the center of mass of the particle group;

[0131] S37. Based on the health metric Q(t), the mutation intensity δ and the Lévy flight index λ are dynamically adjusted to gradually optimize the search behavior of the particle swarm, and finally reach the global optimal solution or meet the predetermined stop condition; the formula for dynamic adjustment is:

[0132] δ(t)=δ0·exp(-ξ·Q(t));

[0133]

[0134] Among them, δ0 and ξ are the initial step size and adjustment rate, and η is the exponential adjustment factor.

[0135] In this implementation, S4 includes the following steps:

[0136] S41. In the reconstructed particle swarm, all possible manufacturing process paths of the micro spring washer are analyzed and the initial process parameter combination p of each path is defined. k , and calculate the initial fitness value f of each path k (t);

[0137] S42, capture the synergy and competition relationship between paths, establish the path complementary effect matrix Θ, whose elements θ kl Used to represent the path P k and path P l The complementary and inhibitory relationship between them:

[0138]

[0139] Among them, p k ·p l represents the inner product of path parameters, p k -p l∥ represents the Euclidean distance between path parameters, σ is the smoothing factor;

[0140] S43, based on the result of the complementary effect matrix Θ, select the path with the highest initial fitness value as the reference path P base , optimize it, and the optimized parameter combination As a reference for other paths;

[0141] S44, for each non-reference path P k To optimize, combine the complementary effect information between paths and use the adaptive path optimization formula:

[0142]

[0143] Among them, λ k For path P k The adaptive step size, θ kl is the path complementarity effect coefficient, which is used to guide the adjustment of path parameters and consider the complementary and competitive relationships between paths;

[0144] S45, through multiple iterations of alternate optimization path combination, gradually optimize the final fitness value of each path And analyze the synergistic effect between each path, and select the final optimized path as the manufacturing process path combination;

[0145] S46. In the actual manufacturing process, the optimized manufacturing process path combination is applied, and the key parameters p are monitored in real time. j (t), when it is found that the parameters in the manufacturing process deviate from the target, the path complementary effect matrix Θ is used for dynamic adjustment;

[0146] S47. After manufacturing is completed, analyze the actual manufacturing data of each process path, compare the prediction results, and modify the complementary effect matrix Θ and the optimization step size λ. k .

[0147] In this implementation, S5 includes the following steps:

[0148] S51. In the manufacturing process, the optimal process path combination p is determined by the complementary manufacturing path optimization method. opt , and collect key manufacturing data d(t) in real time, including dimensional accuracy d size (t), surface quality d surface (t) and elastic modulus d elastic (t);

[0149] S52, inputting the manufacturing data d(t) collected in real time into the feedback control system, calculating the deviation Δd(t) between each data and the target value, so as to identify the errors existing in the manufacturing process;

[0150] S53, dynamically adjusting the manufacturing process parameter p(t) according to the information of the deviation Δd(t) and the path complementary effect matrix Θ;

[0151] S54, monitor the changing trend of manufacturing data in real time through the feedback control system, and calculate the convergence stability σ of the manufacturing path stability (t) Evaluate the stability of the current process path, when σ stability (t) When it drops below a certain set threshold, the process path is considered to be stable;

[0152] S55. If a significant fluctuation or deviation from the target is detected in the process path during the manufacturing process, the parameters between the paths are transferred to make adjustments and update the path combination. The adjustment formula is:

[0153] p new (t) = p opt (t)+γ·(∑ l≠k θ kl ·p l (t)-p new (t));

[0154] Among them, γ is the adjustment coefficient, which is used to achieve parameter migration and dynamic correction between paths;

[0155] S56, feeding back the real-time adjusted process parameters and manufacturing data to the particle swarm optimization algorithm to update the search direction and fitness evaluation of the particle swarm;

[0156] S57. After manufacturing is completed, analyze all collected manufacturing data, compare the actual manufacturing results with the optimization prediction results, update the path complementary effect matrix Θ and the parameter adjustment model of the process path to optimize the accuracy and consistency of subsequent manufacturing batches.

[0157] In this implementation, S6 includes the following steps:

[0158] S61. Input the real-time collected manufacturing data d(t) into the particle swarm optimization algorithm, and dynamically modify the fitness value of the particle according to the change of the data. The fitness modification formula combines the nonlinear volatility of the manufacturing data and calculates the modification factor through the entropy weight method:

[0159]

[0160] Where H(Δd(t) is the deviation data Δd(t) = d(t) - d target The information entropy reflects the complexity of data fluctuations, and α is the adjustment coefficient;

[0161] S62, based on the corrected fitness value, apply the adaptive mutation strategy to update the particle's position x i (t+1):

[0162] x i (t+1)=x i (t)+γ·(P best (t)-x i (t))+∑ j≠i β ij Δ p (t);

[0163] Among them, γ is the local search step size, Δ p (t) is the process parameter difference vector of particles i and j, and the position is adjusted in combination with the path complementarity effect;

[0164] S63. After the particle position is adjusted, a dynamic weight adjustment mechanism is introduced to combine diversity protection and global fitness to keep the particle swarm balanced between local and global searches; the weight adjustment is updated according to the sensitivity of process parameter changes:

[0165]

[0166] Among them, λ is the adjustment coefficient, σ and D(t) represent the smoothing factor of process parameters and population diversity measure respectively;

[0167] S64. Based on dynamic adjustment, implement multi-objective collaborative optimization and introduce nonlinear Pareto optimal solution set to achieve dynamic balance among multiple manufacturing objectives;

[0168] S65, monitor the fitness convergence of the particle swarm, quantify the convergence through the synergy between paths, and the convergence index calculation formula is:

[0169]

[0170] Among them, ξ is the adjustment coefficient, θ ij is the complementary effect weight between particle i and particle j;

[0171] S66, when the convergence index σ convergence (t) When the set convergence threshold is reached, the real-time data feedback is stopped and the current process parameters are locked as the final manufacturing parameter configuration;

[0172] S67. After manufacturing is completed, based on the data feedback from the manufacturing process and combined with the analysis results of the path complementary effect matrix Θ, the particle swarm optimization algorithm is adjusted in terms of strategy.

[0173] Embodiment 1:

[0174] In a high-precision machinery manufacturing company, the company mainly produces micro spring washers for the aerospace field. The quality requirements of these micro spring washers are extremely high, especially in key indicators such as dimensional tolerance, elastic modulus and surface roughness, which must meet strict industry standards. Traditional manufacturing methods often lead to insufficient manufacturing accuracy and poor product consistency due to the difficulty in dynamically adjusting process parameters, which in turn affects product performance and service life. In order to solve these problems, the company introduced the high-precision manufacturing method of micro spring washers using particle swarm optimization technology proposed by us.

[0175] In the actual production process, the company first set the manufacturing target parameters of the micro spring washers according to the product specifications provided by the customer, including specific dimensional tolerances, elastic modulus, surface roughness and other indicators. In order to ensure manufacturing accuracy, the company used our method to establish a multi-objective optimization function, using these target parameters as the basis for optimization.

[0176] Next, the company initialized the particle swarm and assigned initial manufacturing process parameters to each particle, including cutting speed, feed rate, processing temperature, processing pressure, etc. By setting reasonable initial conditions, the particle swarm is widely distributed in the search space to ensure a wide range of exploration.

[0177] During the production process, as the particle swarm optimization is iterated, the company monitors the changing trend of the global optimal solution in real time. After multiple iterations, when it is detected that the global optimal solution has not been significantly improved, the system automatically starts the mutation-driven particle swarm self-organization reconstruction mechanism. By applying random mutations to the speed and position parameters of some particles, the local optimal state is broken and the topological structure of the particle swarm is reorganized, thereby expanding the search space and improving the global optimization capability.

[0178] After the particle swarm reconstruction, the system applied the manufacturing path complementary optimization technology, analyzed multiple possible manufacturing process paths, and selected the optimal manufacturing path through alternating optimization and data integration between paths. This process significantly improved the efficiency and accuracy of manufacturing, especially in the case of complex shapes and multi-objective optimization requirements.

[0179] In the actual production process, the company also collects manufacturing data in real time, including key manufacturing data such as dimensional accuracy, surface quality and elastic coefficient, and feeds this data back to the particle swarm to further optimize the path and direction. In this process, the system dynamically adjusts the search behavior of the particle swarm and the configuration of manufacturing process parameters based on real-time data, so that the entire manufacturing process can flexibly respond to changes in material properties and fluctuations in equipment status, ensuring product consistency and high precision.

[0180] After completing the manufacturing of a batch of micro spring washers, the company conducted a comprehensive analysis of the manufacturing data of each process path, adjusted and updated the initialization parameters of the particle swarm and the path optimization strategy, and generated an optimization plan for subsequent batch production. This process not only improved production efficiency but also reduced manufacturing costs. Specific data proves our beneficial effects

[0181] To verify the effectiveness of the method, the company conducted several experiments under different production conditions. The experiment was conducted in the company's precision manufacturing workshop for a month, and included the manufacture of different batches of micro spring washers, focusing on key indicators such as dimensional tolerance, elastic modulus and surface roughness.

[0182] Table 1 Data table

[0183]

[0184]

[0185] It can be seen from the above data that after applying the method of the present invention, the manufacturing accuracy and production efficiency of the micro spring washers have been significantly improved, and the defective rate has been greatly reduced. These data fully prove the superiority and practicality of the method of the present invention in high-precision manufacturing.

[0186] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A high-precision manufacturing method for micro spring washers using particle swarm optimization technology, comprising the following steps: S1. Setting manufacturing target parameters of the micro spring washer, wherein the target parameters include specific dimensional tolerance, elastic modulus and surface roughness, and establishing a multi-objective optimization function based on these targets; S2, initializing the particle swarm and assigning initial manufacturing process parameters to each particle, wherein the process parameters include precise cutting speed, feed rate, processing temperature and processing pressure, and the initial particle swarm is distributed in the search space according to a preset multi-objective optimization function; S3. During the optimization iteration of the particle swarm, the changing trend of the global optimal solution is monitored. When it is detected that the global optimal solution has not been significantly improved in multiple iterations, the mutation-driven particle swarm self-organization reconstruction mechanism is started to impose random mutations on the speed and position parameters of some particles to reorganize the topological structure of the particle swarm and break the local optimal state. S4. In the reconstructed particle swarm, the manufacturing path complementary optimization technology is applied to analyze and select multiple possible manufacturing process paths, optimize each path separately, and perform alternating optimization and data integration between paths; S5. Apply the process path determined by the complementary manufacturing path optimization method and the particle swarm optimization results to the actual manufacturing process of the micro spring washer. During the process, key manufacturing data including dimensional accuracy, surface quality and elastic coefficient are collected in real time, and these data are fed back to the particle swarm for further adjustment of the optimization path and direction. S6. Based on the real-time collected manufacturing data, dynamically adjust the search behavior of the particle swarm and the configuration of manufacturing process parameters; S7. After completing batch manufacturing, comprehensively analyze the manufacturing data of each process path, adjust and update the initialization parameters of the particle swarm and the path optimization strategy, and generate an optimization plan for subsequent batch production; S8. Record and archive the optimized manufacturing path and process parameters, and conduct continuous improvement and optimization by analyzing the performance of variation-driven particle swarm self-organization reconstruction and manufacturing path complementary optimization technology in actual manufacturing.

2. The high-precision manufacturing method of micro spring washers using particle swarm optimization technology according to claim 1 is characterized in that: The S2 comprises the following steps: S21, determine the size N of the particle group, each particle corresponds to a manufacturing process parameter combination, the process parameter combination includes a cutting speed v c , feed rate f, processing temperature T and processing pressure P; S22, set the initial position for each particle, the initial position represents a specific process parameter combination x i ,These initial positions are distributed in the search space according to the aforementioned multi-objective optimization function; S23, setting an initial speed for each particle. The initial speed determines the search dynamics of the particle in the initial stage. The speed value is distributed within a reasonable range, reflecting the initial strength of the process parameter adjustment; S24, calculating the initial fitness value of each particle, using a multi-objective optimization function to evaluate the quality of the process parameter combination corresponding to the position of each particle, and using the fitness value as a reference for subsequent optimization; S25. Determine the local optimal position P of each particle based on the initial fitness value best and the global optimal position G of the particle swarm best ,These optimal positions will be used as reference points in the subsequent optimization process; S26, updating the process parameter combination of each particle, adjusting its position in the search space according to the current optimal position and the initial position of the particle, forming a new process parameter combination, and providing a new starting point for subsequent optimization iterations; S27. Perform preliminary verification on the updated particle swarm to verify the feasibility of the new process parameter combination in the manufacturing process and provide initial conditions for the subsequent particle swarm optimization iterative process.

3. The high-precision manufacturing method of micro spring washers using particle swarm optimization technology according to claim 1 is characterized in that: The S3 comprises the following steps: S31, calculate the fitness value f of each particle i 9t) To evaluate the advantages and disadvantages of the current process parameter combination, record the changing trend of the global optimal solution, and provide a basis for subsequent judgment on whether the particle swarm falls into a local optimal state: Among them, w j is the weight of the target parameter, x ij is the jth parameter value of the ith particle, is the ideal value of the target parameter, α j To adjust the parameters; S32, define the change rate ΔG(t) of the global optimal solution to monitor the convergence speed, and adjust the search intensity in combination with the time weighting factor γ(t) to detect whether the particle swarm falls into a local optimal state during the optimization process: Among them, γ(t) is the time weighting factor, which gradually decreases with the increase of the number of iterations t, and β controls the deceleration slope of the factor, T max is the maximum number of iterations; S33, triggering the mutation-driven self-organization reconstruction mechanism, by introducing the Lévy flight mechanism for mutation operation, optimizing the jumping of particles in the search space, so as to expand the search range and break the local optimal limitation: Among them, δ is the step size control parameter, u and v are independent normally distributed random variables, and λ is the Lévy distribution index, which ranges from 1 to 2; S34, apply the topological reconstruction strategy of adaptive coupling mapping to readjust the connection mode and distribution structure of the mutated particle swarm: Where ε is the coupling strength weight, is the fitness gradient of the ith particle, H(t) is the diversity entropy of the particle swarm, H max is the initial entropy value; S35. Recalculate the fitness value after mutation The fitness value is corrected through the mutation energy adjustment mechanism: Among them, κ is the adjustment parameter, δE i is the energy change of particle i during the mutation process, E max is the maximum energy reference value; S36. Calculate the health metric Q(t) of the particle swarm to evaluate the diversity and exploration ability of the particle swarm after mutation, and provide a reference for the dynamic adjustment of the subsequent mutation intensity and topological structure: Among them, δE i is the variation energy change of particle i, is the center of mass of the particle group; S37. Based on the health metric Q(t), the mutation intensity δ and the Lévy flight index λ are dynamically adjusted to gradually optimize the search behavior of the particle swarm, and finally reach the global optimal solution or meet the predetermined stop condition; the formula for dynamic adjustment is: δ(t)=δ0·exp(-ξ·Q(t)); Among them, δ0 and ξ are the initial step size and adjustment rate, and η is the exponential adjustment factor.

4. The high-precision manufacturing method of micro spring washers using particle swarm optimization technology according to claim 1 is characterized in that: The S4 comprises the following steps: S41. In the reconstructed particle swarm, all possible manufacturing process paths of the micro spring washer are analyzed and the initial process parameter combination p of each path is defined. k , and calculate the initial fitness value f of each path k (t); S42, capture the synergy and competition relationship between paths, establish the path complementary effect matrix Θ, whose elements θ kl Used to represent the path P k and path P l The complementary and inhibitory relationship between them: Among them, p k ·p l represents the inner product of path parameters, p k -p l ∥ represents the Euclidean distance between path parameters, σ is the smoothing factor; S43, based on the result of the complementary effect matrix Θ, select the path with the highest initial fitness value as the reference path P base , optimize it, and the optimized parameter combination As a reference for other paths; S44, for each non-reference path P k To optimize, combine the complementary effect information between paths and use the adaptive path optimization formula: Among them, λ k For path P k The adaptive step size, θ kl is the path complementarity effect coefficient, which is used to guide the adjustment of path parameters and consider the complementary and competitive relationships between paths; S45, through multiple iterations of alternate optimization path combination, gradually optimize the final fitness value of each path And analyze the synergistic effect between each path, and select the final optimized path as the manufacturing process path combination; S46. In the actual manufacturing process, the optimized manufacturing process path combination is applied, and the key parameters p are monitored in real time. j (t), when it is found that the parameters in the manufacturing process deviate from the target, the path complementary effect matrix Θ is used for dynamic adjustment; S47. After manufacturing is completed, analyze the actual manufacturing data of each process path, compare the prediction results, and modify the complementary effect matrix Θ and the optimization step size λ. k .

5. The high-precision manufacturing method of micro spring washers using particle swarm optimization technology according to claim 1 is characterized in that: The S5 comprises the following steps: S51. In the manufacturing process, the optimal process path combination p is determined by the complementary manufacturing path optimization method. opt , and collect key manufacturing data d(t) in real time, including dimensional accuracy d size (t), surface quality d surface (t) and elastic modulus d elastic (t); S52, inputting the manufacturing data d(t0) collected in real time into the feedback control system, calculating the deviation Δd(t) between each data and the target value, so as to identify the errors existing in the manufacturing process; S53, dynamically adjusting the manufacturing process parameter p(t) according to the information of the deviation Δd(t) and the path complementary effect matrix Θ; S54, monitor the changing trend of manufacturing data in real time through the feedback control system, and calculate the convergence stability σ of the manufacturing path stability (t) Evaluate the stability of the current process path, when σ stability (t) When it drops below a certain set threshold, the process path is considered to be stable; S55. If a significant fluctuation or deviation from the target is detected in the process path during the manufacturing process, the parameters between the paths are transferred to make adjustments and update the path combination. The adjustment formula is: p new (t)=p opt (t)+γ·(∑ l≠k θ kl ·p l (t)-p new (t)); Among them, γ is the adjustment coefficient, which is used to achieve parameter migration and dynamic correction between paths; S56, feeding back the real-time adjusted process parameters and manufacturing data to the particle swarm optimization algorithm to update the search direction and fitness evaluation of the particle swarm; S57. After manufacturing is completed, analyze all collected manufacturing data, compare the actual manufacturing results with the optimization prediction results, update the path complementary effect matrix Θ and the parameter adjustment model of the process path to optimize the accuracy and consistency of subsequent manufacturing batches.

6. The high-precision manufacturing method of micro spring washers using particle swarm optimization technology according to claim 1 is characterized in that: The S6 comprises the following steps: S61. Input the real-time collected manufacturing data d(t) into the particle swarm optimization algorithm, and dynamically modify the fitness value of the particle according to the change of the data. The fitness modification formula combines the nonlinear volatility of the manufacturing data and calculates the modification factor through the entropy weight method: Where H(Δd(t)) is the deviation data Δd(t) = d(t) - d target The information entropy reflects the complexity of data fluctuations, and α is the adjustment coefficient; S62, based on the corrected fitness value, apply the adaptive mutation strategy to update the particle position x i (t+1): x i (t+1)=x i (t)+γ·(P best (t)-x i (t))+∑ j≠i β ij ·Δ p (t); Among them, γ is the local search step size, Δ p (t) is the process parameter difference vector of particles i and j, and the position is adjusted in combination with the path complementarity effect; S63. After the particle position is adjusted, a dynamic weight adjustment mechanism is introduced to combine diversity protection and global fitness to keep the particle swarm balanced between local and global searches; the weight adjustment is updated according to the sensitivity of process parameter changes: Among them, λ is the adjustment coefficient, σ and D(t) represent the smoothing factor of process parameters and population diversity measure respectively; S64. Based on dynamic adjustment, implement multi-objective collaborative optimization and introduce nonlinear Pareto optimal solution set to achieve dynamic balance among multiple manufacturing objectives; S65, monitor the fitness convergence of the particle swarm, quantify the convergence through the synergy between paths, and the convergence index calculation formula is: Among them, ξ is the adjustment coefficient, θ ij is the complementary effect weight between particle i and particle j; S66, when the convergence index σ convergence (t) When the set convergence threshold is reached, the real-time data feedback is stopped and the current process parameters are locked as the final manufacturing parameter configuration; S67. After manufacturing is completed, based on the data feedback from the manufacturing process and combined with the analysis results of the path complementary effect matrix Θ, the particle swarm optimization algorithm is adjusted in terms of strategy.

Images