Cold pier steel ball quality prediction and optimization method based on big data analysis

Abstract

The invention belongs to the technical field of quality prediction, and particularly relates to a cold heading steel ball quality prediction and optimization method based on big data analysis. Firstly, cold heading pressure, steel chemical components and other data are collected; and constructing a quality evaluation system including indexes such as dimensional precision. Multi-target prediction is realized by adopting improved Pareto, a dominating relation is improved, preference coefficients are introduced to regulate and control target function importance, real number coding and specific crossover and mutation operation overcome traditional problems, and a solution space is explored. And an optimal scheme is selected by combining model solution optimization with production requirements. And comparing and analyzing actual and predicted quality data after application, adjusting the model when the error is large, re-evaluating the coefficient, optimizing algorithm parameters and performing optimization. The improved steel ball is improved in dimensional precision, enhanced in plasticity and better in comprehensive performance although the tensile strength is reduced, the defects of traditional production quality control are effectively overcome, and the product quality and the production efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of quality prediction, and in particular relates to a cold-headed steel ball quality prediction and optimization method based on big data analysis. Background Art

[0002] As a commonly used component in the field of mechanical engineering, the quality of cold-forged steel balls directly affects the performance and service life of mechanical equipment. In the production process of cold-forged steel balls, many process parameters are involved, such as cold heading pressure, ejection pressure, cold heading speed, etc. Slight changes in these parameters may have a significant impact on the quality of the steel balls. At the same time, the chemical composition of the steel itself is an intrinsic factor that determines the basic performance of the steel ball. Different chemical composition ratios will give the steel different hardness, strength, toughness and other characteristics, which will further affect the molding quality and final performance of the steel ball during the cold-forging process. Dimensional accuracy is one of the important indicators for measuring the quality of cold-forged steel balls. Failure to meet dimensional accuracy will directly affect the assembly accuracy and operating stability of the steel ball in mechanical equipment.

[0003] However, the current quality control methods for cold heading steel balls mainly focus on experience accumulation and post-testing. During the production process, operators set process parameters based on past experience, but experience often has limitations and cannot fully cope with various complex and changeable production situations. Post-testing is to conduct quality inspection after the product is produced, which has obvious lag. During the production process, it is difficult to adjust the process parameters in time because it is impossible to predict possible quality problems in real time. This not only leads to a low level of production efficiency, but also frequent parameter adjustments and product rework consume a lot of time, manpower and material resources. Moreover, due to the inability to prevent quality problems in advance, it is difficult to maintain a stable product quality, and the scrap rate remains high, which greatly increases production costs and reduces the market competitiveness of enterprises. Summary of the invention

[0004] In view of the technical problems existing in the above-mentioned background technology, the present invention proposes a cold heading steel ball quality prediction and optimization method based on big data analysis.

[0005] In order to achieve the above object, the technical solution adopted by the present invention comprises the following steps:

[0006] S1. First, collect the data of cold heading pressure, ejection pressure, cold heading speed, chemical composition of steel, dimensional accuracy, tensile strength, elongation after fracture and section shrinkage in the production process;

[0007] S2. Then, a quality evaluation index system is constructed. The quality evaluation includes dimensional accuracy, tensile strength, elongation after fracture and section shrinkage;

[0008] S3, then adopt the improved Pareto method to realize the prediction of multiple objectives; the improved Pareto method to realize the prediction of multiple objectives is implemented as follows:

[0009] S31. First, improve the Pareto dominance relationship and introduce the preference coefficient α i To adjust the importance of each objective function in the judgment of dominance relationship;

[0010] S32, then use real number coding to encode the cold heading process parameters, each individual represents a combination of a set of process parameters;

[0011] S33, using the tournament selection method based on the improved Pareto dominance relationship, randomly select multiple individuals from the population, and give priority to selecting non-dominated individuals to enter the next generation population according to the improved Pareto dominance relationship;

[0012] S34, then adopt the simulated binary crossover SBX method to realize the crossover operation;

[0013] S35. For each parameter X in individual X j , with mutation probability P m Perform mutation operation, the mutated individual X j′ for: Where ω is calculated based on the variation distribution index, The parameters X are j The upper and lower limits of the value of ;

[0014] S36, finally, the population is initialized, and the initial population is randomly generated to ensure that the parameter value of each individual is within a reasonable value range;

[0015] S4. Then solve and optimize the model, and select the optimization scheme of cold heading steel ball production process parameters from the optimal solution set finally obtained in combination with actual production needs;

[0016] S5. Finally, verify and adjust the model, check the accuracy and completeness of the data, re-evaluate the preference coefficient, further optimize the parameters and operations of the genetic algorithm, and then re-run the model for optimization.

[0017] Preferably, the improved dominance relationship in step S31 is specifically defined as: for two solutions x and y, And there exists an l such that α l (f l (x)-f l (y))<0, where α i The value of is set according to the production focus.

[0018] Preferably, the simulated binary crossover SBX method realizes the crossover operation as follows: for two parent individuals X1 and X2, generate a child individual X 1c and X 2c : Where j represents the parameter dimension and β is calculated based on the crossover probability.

[0019] Preferably, the calculation formula of β is: where u is a random number uniformly distributed in the interval [0,1], η c is the crossover distribution index, which is used to control the distribution range of offspring individuals around the parent individuals.

[0020] Preferably, in step S35, ω is calculated according to the variation distribution index, and the calculation formula is: Where r is a random number uniformly distributed in the interval [0,1], η m is the variation distribution index.

[0021] Preferably, the model solving and optimization in step S4 is implemented as follows:

[0022] S41, setting parameters of the genetic algorithm, including population size, maximum number of iterations, crossover probability, mutation probability, preference coefficient, crossover distribution index, and mutation distribution index;

[0023] S42, running the improved genetic algorithm to search for the Pareto optimal solution set through continuous iterative evolution, performing selection, crossover and mutation operations on the population in each iteration, and updating the non-dominated solution set according to the improved Pareto dominance relationship until the maximum number of iterations is reached;

[0024] S43. Finally, from the Pareto optimal solution set obtained, combined with the actual production needs, select an optimal solution as the optimization solution for the production process parameters of cold heading steel balls.

[0025] Compared with the prior art, the advantages and positive effects of the present invention are that a scientific evaluation system is constructed to clarify the standard basis. The preference coefficient is introduced to improve the Pareto dominance relationship, the importance of the objective function is dynamically adjusted, and real number coding and specific crossover and mutation operations are used to overcome traditional problems and efficiently explore the solution space. The model solution optimization is combined with production needs to determine the process parameters. After application, the model is adjusted through comparative analysis. After improvement, the dimensional accuracy of the steel ball is improved, the plasticity is enhanced, and although the tensile strength is reduced, the overall performance is better, which effectively solves many drawbacks of traditional production quality control and improves product quality and production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0027] Figure 1 A flow chart for realizing the structure of the present invention; DETAILED DESCRIPTION

[0028] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described below in conjunction with the accompanying drawings and embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0029] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways than those described herein. Therefore, the present invention is not limited to the specific embodiments of the following disclosure.

[0030] Embodiment, in order to effectively solve the many problems existing in the traditional cold heading steel ball production quality control method, such as reliance on experience leading to unreasonable process parameter settings, hysteresis in post-detection resulting in inability to adjust parameters in time, unstable product quality and high scrap rate, increased production costs, etc., the present invention adopts a cold heading steel ball quality prediction and optimization method based on big data analysis. The specific process is as follows Figure 1 shown.

[0031] First, in order to fully and accurately grasp the various influencing factors in the production process of cold heading steel balls, the present invention collects the data of cold heading pressure, ejection pressure, cold heading speed, chemical composition of steel, dimensional accuracy, tensile strength, elongation after fracture and cross-sectional shrinkage in the production process. These data cover multiple aspects such as process parameters, raw material characteristics and product performance indicators to ensure the comprehensiveness and accuracy of subsequent analysis.

[0032] Then, we built a quality assessment index system, and determined that the quality assessment includes dimensional accuracy, tensile strength, elongation after fracture, and cross-sectional shrinkage. Where ΔD j is the diameter deviation of the jth steel ball, N is the number of steel balls tested; elongation after fracture index where δ std is the standard elongation value after fracture, ψ std is the value of standard section reduction, where σ stdIt is the standard value of tensile strength. The scientifically constructed quality assessment index system provides clear standards and basis for subsequent quality prediction and optimization.

[0033] Then, considering that traditional technologies mostly use fixed modes to handle multi-objective optimization, this technology introduces a preference coefficient, based on the improved dominance relationship, to achieve dynamic regulation of the importance of each objective function, and accurately adapt to the production needs of cold-headed steel balls. Its unique real number encoding, simulated binary crossover and mutation operations are in line with the characteristics of process parameters, effectively overcoming the accuracy loss and computational complexity of traditional encoding conversion, efficiently exploring high-quality solution space, avoiding local optimality, and improving population diversity and evolution efficiency. First, improve the Pareto dominance relationship and introduce the preference coefficient α i To adjust the importance of each objective function in the dominance relationship judgment, the improved dominance relationship is specifically defined as: for two solutions x and y, And there exists an l such that α l (f l (x)-f l (y))<0, where α i The value of is set according to the production focus. i The update formula for dimensional accuracy is: in is the adjusted dimensional accuracy preference coefficient, ΔI D，max is the preset maximum value of dimensional accuracy deviation, and α1 is the current basic preference setting for dimensional accuracy. For tensile strength, elongation after fracture and section shrinkage: Then, the cold heading process parameters are encoded by real number coding, and each individual represents a combination of process parameters. The tournament selection method based on the improved Pareto dominance relationship is used to randomly select multiple individuals from the population, and non-dominated individuals are preferentially selected to enter the next generation population according to the improved Pareto dominance relationship. Then, the simulated binary crossover SBX method is used to implement the crossover operation. For two parent individuals X1 and X2, the offspring individual X 1c and X 2c : Where j represents the parameter dimension, and β is calculated based on the crossover probability: where u is a random number uniformly distributed in the interval [0,1], η c is the crossover distribution index, which is used to control the distribution range of offspring individuals around the parent individuals. For each parameter X in individual X j , with mutation probability P m Perform mutation operation, the mutated individual X j′ for: Where ω is calculated based on the variation distribution index, and the calculation formula is: Where r is a random number uniformly distributed in the interval [0,1], η m is the variation distribution index, The parameters X are j The upper and lower limits of the value are determined; finally, the population is initialized and the initial population is randomly generated to ensure that the parameter value of each individual is within a reasonable range.

[0034] The next step is to solve and optimize the model, set the parameters of the genetic algorithm, including population size, maximum number of iterations, crossover probability, mutation probability, preference coefficient, crossover distribution index and mutation distribution index; run the improved genetic algorithm, and search for the Pareto optimal solution set through continuous iterative evolution. In each iteration, perform selection, crossover and mutation operations on the population, and update the non-dominated solution set according to the improved Pareto dominance relationship until the maximum number of iterations is reached; finally, from the final Pareto optimal solution set, combined with actual production needs, select an optimal solution as the optimization solution for the production process parameters of cold heading steel balls.

[0035] Finally, the optimized process parameters are applied to the actual production of cold heading steel balls, and a certain number of steel ball samples are produced. The quality of the produced steel ball samples is tested to obtain the actual quality data. The actual quality data is compared and analyzed with the quality data predicted by the optimization model, and the error index is calculated. If the error index exceeds the acceptable range, the optimization model is adjusted. Check the accuracy and completeness of the data, re-evaluate the preference coefficient, further improve the parameters and operation of the genetic algorithm, and then re-run the model for optimization until a satisfactory quality prediction and optimization effect is achieved. The following is a comparison of the performance of cold heading steel balls before and after the improvement, as shown in Table 1.

[0036] Table 1: Comparison of cold heading steel ball performance before and after improvement

[0037] Actual value before optimization Actual value after process optimization Predicted value after process optimization Dimensional accuracy (mm) 0.09 0.05 0.047 Tensile strength(MPa) 462 419 415.732 Elongation after break (%) 34 38 40.140 Sectional shrinkage (%) 75 81 84.551

[0038] Analysis shows that after the process improvement, the dimensional accuracy of the cold-headed steel balls was reduced from 0.09mm to 0.05mm, and the accuracy was significantly improved. The tensile strength dropped from 462MPa to 419MPa. Although it has decreased, it may be to achieve an optimized balance of overall performance. The elongation after fracture increased from 34% to 38%, and the toughness was enhanced, which can better cope with external force impact and reduce the risk of brittle fracture, especially in scenarios with impact loads. The advantages are obvious. The cross-sectional shrinkage rate increased from 75% to 81%, further indicating that the plasticity of the material has been significantly improved, enabling the steel balls to better adapt to various conditions during processing and use. Overall, after the process improvement, the cold-headed steel balls have made significant progress in dimensional accuracy and plasticity. Although the tensile strength has decreased, the overall comprehensive mechanical properties are better.

[0039] The above description is only a preferred embodiment of the present invention and does not limit the present invention in other forms. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A method for predicting and optimizing the quality of cold heading steel balls based on big data analysis, characterized in that: The following steps are involved: S1. First, collect the data of cold heading pressure, ejection pressure, cold heading speed, chemical composition of steel, dimensional accuracy, tensile strength, elongation after fracture and section shrinkage in the production process; S2. Then, a quality evaluation index system is constructed. The quality evaluation includes dimensional accuracy, tensile strength, elongation after fracture and section shrinkage; S3, then adopt the improved Pareto method to realize the prediction of multiple objectives; the improved Pareto method to realize the prediction of multiple objectives is realized as follows: S31. First, improve the Pareto dominance relationship and introduce the preference coefficient α i To adjust the importance of each objective function in the judgment of dominance relationship; S32, then use real number coding to encode the cold heading process parameters, each individual represents a combination of a set of process parameters; S33, using the tournament selection method based on the improved Pareto dominance relationship, randomly select multiple individuals from the population, and give priority to selecting non-dominated individuals to enter the next generation population according to the improved Pareto dominance relationship; S34, then adopt the simulated binary crossover SBX method to realize the crossover operation; S35. For each parameter X in individual X j , with mutation probability P m Perform mutation operation, the mutated individual X j′ for: Where ω is calculated based on the variation distribution index, The parameters X are j The upper and lower limits of the value of ; S36, finally, the population is initialized, and the initial population is randomly generated to ensure that the parameter value of each individual is within a reasonable value range; S4. Then solve and optimize the model, and select the optimization scheme of cold heading steel ball production process parameters from the optimal solution set finally obtained in combination with actual production needs; S5. Finally, verify and adjust the model, check the accuracy and completeness of the data, re-evaluate the preference coefficient, further optimize the parameters and operations of the genetic algorithm, and then re-run the model for optimization.

2. The method for predicting and optimizing the quality of cold-forged steel balls based on big data analysis according to claim 1, characterized in that: The improved dominance relationship in step S31 is specifically defined as: for two solutions x and y, And there exists an l such that α l (f l (x)-f l (y))<0, where α i The value of is set according to the production focus.

3. The method for predicting and optimizing the quality of cold-forged steel balls based on big data analysis according to claim 1, characterized in that: The step S34 simulates the binary crossover SBX method to implement the crossover operation as follows: for two parent individuals X1 and X2, a child individual X is generated. 1c and X 2c : Where j represents the parameter dimension and β is calculated based on the crossover probability.

4. The method for predicting and optimizing the quality of cold-forged steel balls based on big data analysis according to claim 3 is characterized in that: The calculation formula of β is: where u is a random number uniformly distributed in the interval [0,1], η c is the crossover distribution index, which is used to control the distribution range of offspring individuals around the parent individuals.

5. The method for predicting and optimizing the quality of cold-forged steel balls based on big data analysis according to claim 1 is characterized in that: In step S35, ω is calculated based on the variation distribution index, and the calculation formula is: Where r is a random number uniformly distributed in the interval [0,1], η m is the variation distribution index.

6. The method for predicting and optimizing the quality of cold-forged steel balls based on big data analysis according to claim 1, characterized in that: The implementation of the model solution and optimization in step S4 is as follows: S41, setting parameters of the genetic algorithm, including population size, maximum number of iterations, crossover probability, mutation probability, preference coefficient, crossover distribution index, and mutation distribution index; S42, running the improved genetic algorithm to search for the Pareto optimal solution set through continuous iterative evolution, performing selection, crossover and mutation operations on the population in each iteration, and updating the non-dominated solution set according to the improved Pareto dominance relationship until the maximum number of iterations is reached; S43. Finally, from the Pareto optimal solution set obtained, combined with the actual production needs, select an optimal solution as the optimization solution for the production process parameters of cold heading steel balls.

Images