Intelligent digital quenching process optimization method and system for spring

Abstract

The invention relates to the technical field of spring quenching process optimization, in particular to an intelligent digital quenching process optimization method and system for a spring. Firstly, a plurality of springs of the same type are obtained, and a first spring detection set is constructed; secondly, screening the springs of the first spring detection set to obtain a second spring detection set; thirdly, dividing the second spring detection set into a plurality of spring detection subsets according to the number of quenching process parameter sets; then, the springs of each spring detection subset are quenched according to the corresponding quenching process parameters, the quenched springs are subjected to spring quenching performance detection, and the spring quenching performance detection value corresponding to each spring detection subset is obtained; and finally, obtaining an optimal quenching process parameter set according to the spring quenching performance detection value. According to the method, various properties of the quenched spring can be combined, so that the optimal quenching process parameter set is accurately obtained, and the quenching process optimization of the spring is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of spring quenching process optimization, and in particular to an intelligent digital quenching process optimization method and system for a spring. Background Art

[0002] With the continuous development of the manufacturing industry, especially the widespread application of precision machinery and automation equipment, springs, as a key mechanical component, play a vital role in the industrial field. The performance of springs directly affects the stability and reliability of mechanical systems. Therefore, it is necessary to optimize the spring manufacturing process to ensure the performance of the spring.

[0003] The manufacturing steps of springs include cold forming, quenching and tempering. The quenching step is to heat the spring to an appropriate temperature and then cool it quickly to increase the hardness and strength of the spring. During the quenching process, key parameters such as temperature and cooling rate need to be controlled to ensure the quenching quality, so the quenching process of the spring needs to be optimized.

[0004] The optimization of the quenching process can adjust and improve the quenching process parameters based on the quality of the spring after quenching. Although the existing detection method for the quality of the spring after quenching can realize the quality detection of the spring, the detection is not comprehensive enough, resulting in inaccurate optimization of the quenching process and failure to obtain a spring with high comprehensive quality.

[0005] Therefore, an intelligent digital quenching process optimization method and system for springs are proposed. Summary of the invention

[0006] The object of the present invention is to provide an intelligent digital quenching process optimization method and system for springs. First, multiple springs of the same type are obtained to construct a first spring detection set; secondly, the springs of the first spring detection set are screened to obtain a second spring detection set; thirdly, the second spring detection set is divided into multiple spring detection subsets according to the number of quenching process parameter sets; then, the springs of each of the spring detection subsets are quenched according to the corresponding quenching process parameters, and the spring quenching performance test is performed on the quenched springs to obtain the spring quenching performance test value corresponding to each of the spring detection subsets; finally, the optimal quenching process parameter set is obtained according to the spring quenching performance test value.

[0007] To achieve the above object, the present invention provides the following technical solutions: An intelligent digital quenching process optimization method for a spring, comprising: Acquire multiple springs of the same type and construct a first spring detection set; Screening the springs in the first spring detection set to obtain a second spring detection set; Acquire M groups of quenching process parameter sets, and divide the second spring detection set into M groups of spring detection subsets; Quenching the springs of each of the spring detection subsets according to the quenching process parameters of the corresponding quenching process parameter set; Perform spring quenching performance testing on the quenched springs, and obtain the spring quenching performance test value corresponding to each of the spring test subsets, the specific steps comprising: performing surface cleaning on the quenched springs; performing spring appearance defect testing and spring hardness testing on the cleaned springs, respectively, to obtain the spring appearance defect testing comprehensive value and spring hardness testing comprehensive value of the spring test subset; Dividing each of the spring detection subsets into a spring tension detection subset and a spring compression detection subset; Performing a spring stretching test on each spring in the spring stretching test subset to obtain a spring stretching test comprehensive value of the corresponding spring test subset; performing a spring compression test on each spring in the spring compression test subset to obtain a spring compression test comprehensive value of the corresponding spring test subset; Obtaining the spring quenching performance test value of the spring test subset according to the four comprehensive test values ​​corresponding to each of the spring test subsets; According to the spring quenching performance test value, an optimal quenching process parameter set is obtained.

[0008] Furthermore, none of the springs in the first spring test set has been quenched; each of the springs undergoes the same manufacturing steps before quenching.

[0009] Furthermore, the screening of the springs in the first spring detection set to obtain the second spring detection set includes: performing spring appearance defect detection on the springs in the first spring detection set, screening out springs that pass the spring appearance defect detection, and constructing the second spring detection set; the spring appearance defect detection process includes: Acquire a spring appearance image of the spring, and input the spring appearance image into a spring appearance defect detection model; The spring appearance defect detection model includes a spring appearance image input layer, a spring appearance feature map acquisition layer, a spring appearance feature map fusion layer and a spring external defect detection value acquisition layer; The spring appearance feature map acquisition layer is used to perform a convolution operation on the spring appearance image to obtain a plurality of the appearance feature maps; the spring appearance feature map fusion layer is used to fuse a plurality of the appearance feature maps to obtain spring appearance features; The spring appearance features include spring shape features and spring defect features; the spring shape features include the maximum diameter of the spring wire, the minimum diameter of the spring wire, the maximum outer diameter of the spring, the minimum inner diameter of the spring and the non-support ring pitch; The spring defect characteristics include spring surface crack characteristics, spring surface foreign matter characteristics and spring surface depression characteristics; Obtaining a first spring defect detection value according to the spring shape feature; obtaining a second spring defect detection value according to the spring defect feature; obtaining a first spring appearance defect detection value according to the first spring defect detection value and the second spring defect detection value; When the first spring appearance defect detection value of the spring is greater than a set defect detection threshold, the spring is determined to have failed the spring appearance defect detection; otherwise, the spring is determined to have passed the spring appearance defect detection.

[0010] Furthermore, the quenching process parameter sets include: each of the quenching process parameter sets includes N quenching process parameters; and the quenching process parameter values ​​of different quenching process parameter sets are not completely the same.

[0011] Furthermore, the process of the spring stretching detection includes: Obtaining a first spring shape feature of each of the springs before performing a spring stretching test; Apply a set tension to the spring. After the moment, cancel the tension; obtain the stretched length of the spring when canceling the tension; After the moment, the second spring shape feature and the spring defect feature of each of the springs are obtained; the spring stretching detection value of the spring is obtained according to the first spring shape feature, the second spring shape feature, the stretching length and the spring defect feature; and the spring stretching detection comprehensive value of the spring detection subset corresponding to the spring stretching detection subset is obtained according to the spring stretching detection value of each spring in the spring stretching detection subset.

[0012] Furthermore, the process of the spring compression detection includes: Obtaining a third spring shape feature of each of the springs before performing a spring compression test; Apply a set pressure to the spring, After the time, the pressure is cancelled; when the pressure is cancelled, the compressed length of the spring is obtained; when the pressure is cancelled, the compressed length of the spring is obtained; After the moment, the fourth spring shape characteristic and the spring defect characteristic of each of the springs are obtained; the spring compression detection value of the spring is obtained based on the third spring shape characteristic, the fourth spring shape characteristic, the compression length and the spring defect characteristic; and the spring compression detection comprehensive value of the spring detection subset corresponding to the spring compression detection subset is obtained based on the spring compression detection value of each spring in the spring compression detection subset.

[0013] Furthermore, obtaining the optimal quenching process parameter set according to the spring quenching performance test value includes: The quenching process parameter set corresponding to the spring detection subset with the largest spring quenching performance detection value is selected as the optimal quenching process parameter set.

[0014] An intelligent digital quenching process optimization system for a spring, comprising: A first spring detection set acquisition unit: used to acquire multiple springs of the same type and construct a first spring detection set; A second spring detection set acquisition unit: used for screening the springs of the first spring detection set to acquire a second spring detection set; A quenching spring detection subset division unit is used to obtain M groups of quenching process parameter sets and divide the second spring detection set into M groups of spring detection subsets; Quenching unit: used for quenching the springs of each of the spring detection subsets according to the quenching process parameters of the corresponding quenching process parameter set; A quenching performance detection unit is used to perform a spring quenching performance detection on the quenched spring and obtain a spring quenching performance detection value corresponding to each of the spring detection subsets; the spring quenching performance detection includes a spring appearance defect detection, a spring hardness detection, a spring tension detection and a spring compression detection; The optimal quenching process parameter set acquisition unit is used to acquire the optimal quenching process parameter set according to the spring quenching performance detection value.

[0015] Furthermore, the step of obtaining the spring quenching performance test value includes: performing surface cleaning on the quenched spring; performing spring appearance defect detection and spring hardness detection on the cleaned spring respectively, and obtaining the spring appearance defect detection comprehensive value and spring hardness detection comprehensive value of the spring detection subset; Dividing each of the spring detection subsets into a spring tension detection subset and a spring compression detection subset; Performing a spring stretching test on each spring in the spring stretching test subset to obtain a spring stretching test comprehensive value of the corresponding spring test subset; Performing spring compression testing on each spring of the spring compression testing subset to obtain a spring compression testing comprehensive value of the corresponding spring testing subset; The spring quenching performance test value of the spring test subset is obtained according to the four comprehensive test values ​​corresponding to each of the spring test subsets.

[0016] Furthermore, the process of the spring stretching detection includes: Obtaining a first spring shape feature of each of the springs before performing a spring stretching test; Apply a set tension to the spring. After the moment, cancel the tension; obtain the stretched length of the spring when canceling the tension; After the moment, the second spring shape feature and the spring defect feature of each of the springs are obtained; the spring stretching detection value of the spring is obtained according to the first spring shape feature, the second spring shape feature, the stretching length and the spring defect feature; and the spring stretching detection comprehensive value of the spring detection subset corresponding to the spring stretching detection subset is obtained according to the spring stretching detection value of each spring in the spring stretching detection subset.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. Before quenching, the present invention performs spring appearance defect detection on each spring through a spring appearance defect detection model. The spring appearance defect detection model can extract the shape characteristics and defect characteristics of the spring, and calculate the first spring appearance defect detection value; the spring is screened by the first spring appearance defect detection value. This method can detect various surface defects of the spring, and remove springs that fail the appearance defect detection, avoiding the influence of the spring's own defects on the quality of the spring after quenching, thereby improving the accuracy of quenching process optimization.

[0018] 2. The present invention sets different quenching environments and quenches the spring in different quenching environments. The spring quenching performance test value is obtained by performing spring appearance defect detection, spring hardness detection, spring tension detection and spring compression detection on the quenched spring. This method can combine various characteristics of the spring after quenching, and can accurately obtain the comprehensive performance of the spring after quenching, thereby improving the accuracy of quenching process optimization.

[0019] 3. When the present invention performs a spring stretch test on a spring, the comprehensive value of the spring stretch test is calculated by the stretched length of the spring after the tension is applied, the spring defect characteristics, and the change in the shape characteristics of the spring before and after the stretching. The comprehensive value of the spring stretch test can not only measure the tensile performance of the spring but also measure the recovery ability of the spring after stretching. This method can accurately obtain the various properties of the spring under the action of tension, can accurately obtain the comprehensive value of the spring stretch test, and then accurately obtain the comprehensive performance of the spring after quenching. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A flow chart of an intelligent digital quenching process optimization method for a spring provided in an embodiment of the present invention; Figure 2 A structural diagram of an intelligent digital quenching process optimization system for a spring provided in an embodiment of the present invention; Figure 3 A flow chart for obtaining a spring quenching performance test value provided by an embodiment of the present invention; Figure 4 A flow chart for obtaining a comprehensive value of a spring stretching test provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] With the continuous development of the manufacturing industry, especially the widespread application of precision machinery and automation equipment, springs, as a key mechanical component, play a vital role in the industrial field. The performance of springs directly affects the stability and reliability of mechanical systems. Therefore, it is necessary to optimize the spring manufacturing process to ensure the performance of the spring.

[0023] The manufacturing steps of springs include cold forming, quenching and tempering. The quenching step is to heat the spring to an appropriate temperature and then cool it quickly to increase the hardness and strength of the spring. During the quenching process, key parameters such as temperature and cooling rate need to be controlled to ensure the quenching quality, so the quenching process of the spring needs to be optimized.

[0024] The present invention provides an intelligent digital quenching process optimization method for a spring, and the method is applied to an intelligent digital quenching process optimization system for a spring. The specific method flow chart and system structure diagram refer to Figure 1 and Figure 2 .

[0025] Embodiment 1 Reference Figure 1 S10 in the figure is applied to the first spring detection set acquisition unit of an intelligent digital quenching process optimization system for a spring.

[0026] Furthermore, this embodiment collects 350 springs of the same type from factory A to construct a first spring detection set; the same type means that the material, shape, size and target mass after quenching of the springs are the same; the springs in the first spring detection set are not quenched; and each of the springs undergoes the same manufacturing steps before quenching.

[0027] In this step of the present embodiment, multiple springs produced in the same batch are obtained and subjected to the same treatment before quenching. This method can avoid the influence of spring manufacturing differences on the quenching results, thereby improving the accuracy of quenching process optimization.

[0028] Reference Figure 1S20 in the figure is applied to the second spring detection set acquisition unit of an intelligent digital quenching process optimization system for a spring.

[0029] Further, the springs in the first spring detection set are subjected to spring appearance defect detection, and springs that pass the spring appearance defect detection are screened out to construct a second spring detection set; the spring appearance defect detection process includes: Acquire a spring appearance image of the spring, and input the spring appearance image into a spring appearance defect detection model; The spring appearance defect detection model includes a spring appearance image input layer, a spring appearance feature map acquisition layer, a spring appearance feature map fusion layer and a spring external defect detection value acquisition layer; The spring appearance feature map acquisition layer is used to perform a convolution operation on the spring appearance image to obtain a plurality of the appearance feature maps; the spring appearance feature map fusion layer is used to fuse a plurality of the appearance feature maps to obtain spring appearance features; The spring appearance features include spring shape features and spring defect features; the spring shape features include the maximum diameter of the spring wire, the minimum diameter of the spring wire, the maximum outer diameter of the spring, the minimum inner diameter of the spring and the non-support ring pitch; the spring defect features include spring surface crack features, spring surface foreign matter features and spring surface depression features; the surface crack features include the number of cracks and the crack area of ​​each crack; the spring surface foreign matter features include the number of foreign matter; the foreign matter includes sand and other foreign matter generated in the production steps before the spring is quenched; the spring surface depression features include the number of depressions and the area of ​​each depression; Obtain a spring standard shape feature corresponding to the spring, and obtain a first spring defect detection value based on the spring shape feature and the spring standard shape feature; obtain a second spring defect detection value based on the spring defect feature; obtain a first spring appearance defect detection value based on the first spring defect detection value and the second spring defect detection value; The calculation formula of the first spring defect detection value is: ; in, is represented as the first spring defect detection value; It is represented by the spring shape features obtained by the spring appearance defect detection model; Represented as the standard shape characteristics of the spring; is expressed as a nonlinear function; the larger the first spring defect detection value is, the larger the difference between the spring shape characteristics of the spring and the spring standard shape characteristics is; The calculation formula of the second spring defect detection value is: ; in, Expressed as the second spring defect detection value; Expressed as the number of cracks; Expressed as the crack area of ​​the crack described in item i; Expressed as the number of foreign bodies on the surface; It is expressed as the number of depressions; Expressed as the area of ​​the jth depression; , and They are expressed as crack coefficient, foreign body coefficient and concave coefficient respectively; It is expressed as an exponential function with a natural constant as base; , and The default is , and can also be changed according to actual application; The calculation formula of the first spring appearance defect detection value is: ; in, It is represented as the appearance defect detection value of the first spring; and They are represented as the first spring defect coefficient and the second spring defect coefficient respectively; and The default value is 0.5, which can be changed according to the actual application; When the first spring appearance defect detection value of the spring is greater than the set defect detection threshold; the spring is determined to have failed the spring appearance defect detection; otherwise, the spring is determined to have passed the spring appearance defect detection; the defect detection threshold is obtained based on experience.

[0030] In this step of the present embodiment, before quenching, each spring is subjected to spring appearance defect detection through a spring appearance defect detection model. The spring appearance defect detection model can extract the shape characteristics and defect characteristics of the spring, and calculate the first spring appearance defect detection value; the springs are screened according to the first spring appearance defect detection value. This method can detect various surface defects of the spring, and remove springs that fail the appearance defect detection, thereby avoiding the influence of the spring's own defects on the quality of the spring after quenching, thereby improving the accuracy of quenching process optimization.

[0031] Reference Figure 1 S30 in the figure is applied to a quenching spring detection subset division unit of an intelligent digital quenching process optimization system for a spring.

[0032] Further, M groups of quenching process parameter sets are obtained, and the second spring detection set is divided into M groups of spring detection subsets; the quenching process parameter sets include: each of the quenching process parameter sets includes N quenching process parameters; the quenching process parameter values ​​of different quenching process parameter sets are not exactly the same; each of the quenching process parameter sets includes heating temperature, holding time, cooling medium type, cooling rate and cooling time, etc.; this embodiment collects five quenching process parameter sets, and divides the second spring detection set into five spring detection subsets. Some quenching process parameters of each quenching process parameter set refer to Table 1.

[0033] Table 1 Quenching process parameter values ​​for some examples

[0034] In this embodiment, the second spring detection set is divided into multiple spring detection subsets according to the number of quenching process parameter sets, and each spring detection subset corresponds to a quenching process parameter set. This method can verify the quenching effect of each quenching process parameter set on the spring, and then select the optimal quenching process parameter set.

[0035] Reference Figure 1 The S40 in the figure is applied to the quenching unit of an intelligent digital quenching process optimization system for springs.

[0036] Furthermore, in this embodiment, the springs of each of the spring detection subsets are quenched according to the quenching process parameters of the corresponding quenching process parameter set.

[0037] Reference Figure 1 S50 in the figure is applied to the quenching performance detection unit of an intelligent digital quenching process optimization system for a spring.

[0038] Further, the spring quenching performance test is performed on the quenched spring, and the spring quenching performance test value corresponding to each of the spring test subsets is obtained, referring to Figure 3 , the specific steps include: Performing surface cleaning on the quenched spring; performing spring appearance defect detection and spring hardness detection on the cleaned spring respectively, and obtaining a comprehensive value of the spring appearance defect detection and a comprehensive value of the spring hardness detection of the spring detection subset; Dividing each of the spring detection subsets into a spring tension detection subset and a spring compression detection subset; Performing a spring stretching test on each spring in the spring stretching test subset to obtain a spring stretching test comprehensive value of the corresponding spring test subset; performing a spring compression test on each spring in the spring compression test subset to obtain a spring compression test comprehensive value of the corresponding spring test subset; The spring quenching performance test value of the spring test subset is obtained according to the four comprehensive test values ​​corresponding to each of the spring test subsets, and the calculation formula is: ; in, It is represented as the spring quenching performance test value of the kth spring test subset; It is represented as the comprehensive value of spring appearance defect detection of the kth spring detection subset; It is represented as the comprehensive value of the spring hardness test of the kth spring test subset; It is represented as the comprehensive value of the spring stretching test of the kth spring test subset; Expressed as the comprehensive value of spring compression detection of the kth spring detection subset; , , and They are respectively expressed as spring appearance defect coefficient, spring hardness coefficient, spring stretch coefficient and spring compression coefficient; , , and The default value is 1, and can be changed according to actual application.

[0039] This step of the present embodiment sets different quenching environments, and quenches the spring in different quenching environments. The spring quenching performance test value is obtained by performing spring appearance defect detection, spring hardness detection, spring tension detection and spring compression detection on the quenched spring. This method can combine various characteristics of the spring after quenching, and can accurately obtain the comprehensive performance of the spring after quenching, thereby improving the accuracy of quenching process optimization.

[0040] Furthermore, the process of obtaining the comprehensive value of the spring appearance defect detection includes: Perform a spring appearance defect inspection on each of the quenched springs to obtain a first spring appearance defect inspection value for each of the springs; average the first spring appearance defect inspection values ​​of the springs in each of the spring inspection subsets to obtain a comprehensive spring appearance defect inspection value corresponding to the spring inspection subset.

[0041] Furthermore, the process of obtaining the comprehensive value of the spring hardness test includes: Selecting H hardness test points on each of the quenched springs; obtaining a hardness value of each hardness test point; averaging the hardness values ​​of the hardness test points on each of the springs to obtain an average hardness value of each of the springs; Whether the spring hardness test is qualified is judged based on the hardness average value, and the hardness test pass rate of each spring in the spring test subset is obtained; the hardness test pass rate is the comprehensive value of the spring hardness test corresponding to the spring test subset.

[0042] Further, refer to Figure 4 , the process of spring stretching detection includes: Obtaining a first spring shape feature of each of the springs before performing a spring stretching test; Apply a set tension to the spring. After a certain time, the pulling force is cancelled; It can be 60 seconds, 90 seconds, etc.; when the tension is released, the stretched length of the spring is obtained; when the tension is released, the stretched length of the spring is obtained. After the time, obtaining a second spring shape feature and a spring defect feature of each of the springs; It can be 30 seconds, 40 seconds, etc.; obtaining a spring stretching test value of the spring according to the first spring shape feature, the second spring shape feature, the stretching length and the spring defect feature; The calculation formula of the spring stretch test value of the spring is: ; in, It is represented as the spring stretch detection value of the p-th spring in the spring stretch detection subset corresponding to the k-th spring stretch detection subset; It is expressed as the stretched length of the pth spring; Expressed as the set standard stretch length; A spring tensile shape defect value of the spring obtained according to the first spring shape feature and the second spring shape feature; It is represented as the second spring defect detection value of the p-th spring obtained according to the spring defect feature; , and They are expressed as the tensile length factor, tensile shape factor and tensile defect factor respectively; The calculation formula of the spring tensile shape defect value is: ; in, It is represented as the first spring shape feature of the p-th spring of the spring stretch detection subset corresponding to the k-th spring stretch detection subset; denoted as the second spring shape characteristic of the p-th spring; Expressed as the same nonlinear function as in the calculation process of the first spring defect detection value; The spring stretch detection value of each spring in the spring stretch detection subset is averaged to obtain the spring stretch detection comprehensive value of the spring detection subset corresponding to the spring stretch detection subset.

[0043] In this step of the present embodiment, when performing a spring stretch test on the spring, the comprehensive value of the spring stretch test is calculated by the stretched length of the spring after the tension is applied, the spring defect characteristics, and the change in the shape characteristics of the spring before and after the stretching. The comprehensive value of the spring stretch test can not only measure the tensile performance of the spring but also measure the recovery ability of the spring after stretching. This method can accurately obtain the various properties of the spring under the action of tension, can accurately obtain the comprehensive value of the spring stretch test, and further accurately obtain the comprehensive performance of the spring after quenching.

[0044] Furthermore, the process of the spring compression detection includes: Obtaining a third spring shape feature of each of the springs before performing a spring compression test; Apply a set pressure to the spring, After a certain time, the pressure is cancelled; It can be 60 seconds, 90 seconds, etc.; when the pressure is released, the compressed length of the spring is obtained; when the pressure is released, the compressed length of the spring is obtained. After the time, obtaining a fourth spring shape feature and a spring defect feature of each of the springs; It can be 30 seconds, 40 seconds, etc.; obtaining a spring compression detection value of the spring according to the third spring shape feature, the fourth spring shape feature, the compression length and the spring defect feature; The calculation formula of the spring stretch test value of the spring is: ; in, It is represented as the spring extension test value of the qth spring in the kth spring compression test subset; It is expressed as the compressed length of the qth spring; Expressed as the set standard compression length; A spring compression shape defect value of the spring obtained according to the third spring shape feature and the fourth spring shape feature; It is represented as the second spring defect detection value of the qth spring obtained according to the spring defect feature; , and They are expressed as compression length factor, compression shape factor and compression defect factor respectively; The calculation formula of the spring compression shape defect value is: ; in, The third spring shape feature of the qth spring of the kth spring compression detection subset is represented as; The fourth spring shape feature of the qth spring of the kth spring compression detection subset is represented; Expressed as the same nonlinear function as in the calculation process of the first spring defect detection value; The spring compression detection value of each spring in the spring compression detection subset is averaged to obtain a spring compression detection comprehensive value of the spring detection subset corresponding to the spring compression detection subset.

[0045] In this step of the present embodiment, when performing spring compression testing on the spring, the spring compression test comprehensive value is calculated through the compression length of the spring after pressure is applied, the spring defect characteristics, and the change in the shape characteristics of the spring before and after compression. The spring compression test comprehensive value can not only measure the compression performance of the spring but also measure the recovery ability of the spring after compression. This method can accurately obtain the various properties of the spring under pressure, can accurately obtain the spring compression test comprehensive value, and then accurately obtain the comprehensive performance of the spring after quenching.

[0046] Reference Figure 1 S60 in the figure is applied to an optimal quenching process parameter set acquisition unit of an intelligent digital quenching process optimization system for a spring.

[0047] Furthermore, the quenching process parameter set corresponding to the spring detection subset with the largest spring quenching performance test value is selected as the optimal quenching process parameter set; in this embodiment, after calculation, it is obtained that the spring quenching performance test value corresponding to the quenching process parameter set with serial number 4 is the largest, that is, the quenching process parameter set with serial number 4 is used as the optimal quenching process parameter set.

[0048] In this embodiment, the step obtains the optimal quenching process parameter set based on the spring quenching performance test value to optimize the quenching process of the spring. This method can obtain the optimal quenching process parameters based on the sample in advance, optimize the quenching process of the same type of springs, and ensure the quality of the spring after quenching.

[0049] In order to verify the effectiveness of the spring quenching process optimization method adopted in this embodiment, 1200 springs of the same type that passed the spring appearance defect inspection were collected in this embodiment, and randomly grouped to generate five spring subsets, each of which contained 240 springs; the 240 springs were quenched according to the quenching process parameters of the five quenching process parameter sets; the quenched springs were quality inspected to obtain the corresponding spring quenching performance test value and the corresponding spring quality pass rate in each of the spring subsets, the spring quality inspection pass rate refers to the proportion of the number of springs in the spring subset that passed the spring appearance defect inspection, spring hardness inspection, spring tensile inspection and spring compression inspection to the number of springs in the spring subset; refer to Table 2 for specific results.

[0050] Table 2 Example 1 Quenching process optimization verification table

[0051] In Table 2, the spring subset serial numbers and the quenching process parameter set serial numbers correspond one to one. It can be seen from Table 2 that the quenching performance test value of the spring corresponding to the quenching process parameter set with serial number 4 is the largest, and the spring quality test pass rate is no less than the spring quality test pass rates corresponding to the other four quenching process parameter sets. Therefore, the quenching process parameter set with serial number 4 is the optimal quenching process parameter set of the five quenching process parameter sets, that is, the spring quenching process optimization method proposed in this embodiment can effectively obtain the optimal quenching process parameter set.

[0052] Embodiment 2 In Example 1, the method of the present invention is combined with the system to achieve the optimization of the spring quenching process; in Example 2 of the present application, the method of the present invention will be described again, and the specific implementation process includes: A plurality of springs of the same type are obtained to construct a first spring detection set.

[0053] Furthermore, this embodiment collects 300 springs of the same type from factory B to construct a first spring detection set; the same type means that the material, shape, size and target mass after quenching of the springs are the same; the springs in the first spring detection set are not quenched; and each of the springs undergoes the same manufacturing steps before quenching.

[0054] The springs of the first spring detection set are screened to obtain a second spring detection set.

[0055] Further, the springs in the first spring detection set are subjected to spring appearance defect detection, and springs that pass the spring appearance defect detection are screened out to construct a second spring detection set; the spring appearance defect detection process includes: Acquire a spring appearance image of the spring, and input the spring appearance image into a spring appearance defect detection model; The spring appearance defect detection model includes a spring appearance image input layer, a spring appearance feature map acquisition layer, a spring appearance feature map fusion layer and a spring external defect detection value acquisition layer; The spring appearance feature map acquisition layer is used to perform a convolution operation on the spring appearance image to obtain a plurality of the appearance feature maps; the spring appearance feature map fusion layer is used to fuse a plurality of the appearance feature maps to obtain spring appearance features; The spring appearance features include spring shape features and spring defect features; the spring shape features include the maximum diameter of the spring wire, the minimum diameter of the spring wire, the maximum outer diameter of the spring, the minimum inner diameter of the spring and the non-support ring pitch; The spring defect characteristics include spring surface crack characteristics, spring surface foreign matter characteristics and spring surface depression characteristics; the surface crack characteristics include the number of cracks and the crack area of ​​each crack; the spring surface foreign matter characteristics include the number of foreign matter; the spring surface depression characteristics include the number of depressions and the area of ​​each depression; Obtain a spring standard shape feature corresponding to the spring, and obtain a first spring defect detection value based on the spring shape feature and the spring standard shape feature; obtain a second spring defect detection value based on the spring defect feature; obtain a first spring appearance defect detection value based on the first spring defect detection value and the second spring defect detection value; The calculation formula of the first spring defect detection value is: ; in, is represented as the first spring defect detection value; It is represented by the spring shape features obtained by the spring appearance defect detection model; Represented as the standard shape characteristics of the spring; is expressed as a nonlinear function; the larger the first spring defect detection value is, the larger the difference between the spring shape characteristics of the spring and the spring standard shape characteristics is; The calculation formula of the second spring defect detection value is: ; in, Expressed as the second spring defect detection value; Expressed as the number of cracks; Expressed as the crack area of ​​the crack described in item i; Expressed as the number of foreign bodies on the surface; It is expressed as the number of depressions; Expressed as the area of ​​the jth depression; , and They are expressed as crack coefficient, foreign body coefficient and concave coefficient respectively; It is expressed as an exponential function with a natural constant as base; , and The default is , and can also be changed according to actual application; The calculation formula of the first spring appearance defect detection value is: ; in, It is represented as the appearance defect detection value of the first spring; and They are represented as the first spring defect coefficient and the second spring defect coefficient respectively; and The default value is 0.5, which can be changed according to the actual application; When the first spring appearance defect detection value of the spring is greater than the set defect detection threshold; the spring is determined to have failed the spring appearance defect detection; otherwise, the spring is determined to have passed the spring appearance defect detection; the defect detection threshold is obtained based on experience.

[0056] M groups of quenching process parameter sets are obtained, and the second spring detection set is divided into M groups of spring detection subsets.

[0057] Furthermore, the quenching process parameter set includes: each of the quenching process parameter sets includes N quenching process parameters; the quenching process parameter values ​​of different quenching process parameter sets are not exactly the same; each of the quenching process parameter sets includes heating temperature, holding time, cooling medium type, cooling rate and cooling time, etc.; this embodiment collects six quenching process parameter sets, divides the second spring detection set into six spring detection subsets, and some quenching process parameters of each quenching process parameter set refer to Table 3.

[0058] Table 3 Quenching process parameter values ​​for part of Example 2

[0059] The springs of each of the spring detection subsets are quenched according to the quenching process parameters of the corresponding quenching process parameter set; the spring quenching performance test is performed on the quenched springs, and the spring quenching performance test value corresponding to each of the spring detection subsets is obtained.

[0060] Further, the spring quenching performance is tested on the quenched spring, and the spring quenching performance test value corresponding to each spring test subset is obtained. The specific steps include: Performing surface cleaning on the quenched spring; performing spring appearance defect detection and spring hardness detection on the cleaned spring respectively, and obtaining a comprehensive value of the spring appearance defect detection and a comprehensive value of the spring hardness detection of the spring detection subset; Dividing each of the spring detection subsets into a spring tension detection subset and a spring compression detection subset; Performing a spring stretching test on each spring in the spring stretching test subset to obtain a spring stretching test comprehensive value of the corresponding spring test subset; performing a spring compression test on each spring in the spring compression test subset to obtain a spring compression test comprehensive value of the corresponding spring test subset; The spring quenching performance test value of the spring test subset is obtained according to the four comprehensive test values ​​corresponding to each of the spring test subsets, and the calculation formula is: ; in, It is represented as the spring quenching performance test value of the kth spring test subset; It is represented as the comprehensive value of spring appearance defect detection of the kth spring detection subset; It is represented as the comprehensive value of the spring hardness test of the kth spring test subset; It is represented as the comprehensive value of the spring stretching test of the kth spring test subset; Expressed as the comprehensive value of spring compression detection of the kth spring detection subset; , , and They are respectively expressed as spring appearance defect coefficient, spring hardness coefficient, spring stretch coefficient and spring compression coefficient; , , and The default value is 0.25, which can be changed according to the actual application.

[0061] Furthermore, the process of obtaining the comprehensive value of the spring appearance defect detection includes: Perform a spring appearance defect inspection on each of the quenched springs to obtain a first spring appearance defect inspection value for each of the springs; average the first spring appearance defect inspection values ​​of the springs in each of the spring inspection subsets to obtain a comprehensive spring appearance defect inspection value corresponding to the spring inspection subset.

[0062] Furthermore, the process of obtaining the comprehensive value of the spring hardness test includes: Selecting H hardness test points on each of the quenched springs; obtaining a hardness value of each hardness test point; averaging the hardness values ​​of the hardness test points on each of the springs to obtain an average hardness value of each of the springs; Whether the spring hardness test is qualified is judged based on the hardness average value, and the hardness test pass rate of each spring in the spring test subset is obtained; the hardness test pass rate is the comprehensive value of the spring hardness test corresponding to the spring test subset.

[0063] Furthermore, the process of the spring stretching detection includes: Obtaining a first spring shape feature of each of the springs before performing a spring stretching test; Apply a set tension to the spring. After a certain time, the pulling force is cancelled; It can be 60 seconds, 90 seconds, etc.; when the tension is released, the stretched length of the spring is obtained; when the tension is released, the stretched length of the spring is obtained. After the time, obtaining a second spring shape feature and a spring defect feature of each of the springs; It can be 30 seconds, 40 seconds, etc.; obtaining a spring stretching test value of the spring according to the first spring shape feature, the second spring shape feature, the stretching length and the spring defect feature; The calculation formula of the spring stretch test value of the spring is: ; in, It is represented as the spring stretch detection value of the p-th spring in the spring stretch detection subset corresponding to the k-th spring stretch detection subset; It is expressed as the stretched length of the pth spring; Expressed as the set standard stretch length; A spring tensile shape defect value of the spring obtained according to the first spring shape feature and the second spring shape feature; It is represented as the second spring defect detection value of the p-th spring obtained according to the spring defect feature; , and They are expressed as the tensile length factor, tensile shape factor and tensile defect factor respectively; The calculation formula of the spring tensile shape defect value is: ; in, It is represented as the first spring shape feature of the p-th spring of the spring stretch detection subset corresponding to the k-th spring stretch detection subset; denoted as the second spring shape characteristic of the p-th spring; Expressed as the same nonlinear function as in the calculation process of the first spring defect detection value; The spring stretch detection value of each spring in the spring stretch detection subset is averaged to obtain the spring stretch detection comprehensive value of the spring detection subset corresponding to the spring stretch detection subset.

[0064] Furthermore, the process of the spring compression detection includes: Obtaining a third spring shape feature of each of the springs before performing a spring compression test; Apply a set pressure to the spring, After a certain time, the pressure is cancelled; It can be 60 seconds, 90 seconds, etc.; obtaining the compressed length of the spring when the pressure is released; In removing the pressure After the time, obtaining a fourth spring shape feature and a spring defect feature of each of the springs; It can be 30 seconds, 40 seconds, etc.; obtaining a spring compression detection value of the spring according to the third spring shape feature, the fourth spring shape feature, the compression length and the spring defect feature; The calculation formula of the spring stretch test value of the spring is: ; in, It is represented as the spring extension test value of the qth spring in the kth spring compression test subset; It is expressed as the compressed length of the qth spring; Expressed as the set standard compression length; A spring compression shape defect value of the spring obtained according to the third spring shape feature and the fourth spring shape feature; It is represented as the second spring defect detection value of the qth spring obtained according to the spring defect feature; , and They are expressed as compression length factor, compression shape factor and compression defect factor respectively; The calculation formula of the spring compression shape defect value is: ; in, The third spring shape feature of the qth spring of the kth spring compression detection subset is represented as; The fourth spring shape feature of the qth spring of the kth spring compression detection subset is represented; Expressed as the same nonlinear function as in the calculation process of the first spring defect detection value; The spring compression detection value of each spring in the spring compression detection subset is averaged to obtain a spring compression detection comprehensive value of the spring detection subset corresponding to the spring compression detection subset.

[0065] According to the spring quenching performance test value, an optimal quenching process parameter set is obtained.

[0066] Furthermore, the quenching process parameter set corresponding to the spring detection subset with the largest spring quenching performance test value is selected as the optimal quenching process parameter set; in this embodiment, after calculation, it is obtained that the spring quenching performance test value corresponding to the quenching process parameter set with serial number 5 is the largest, that is, the quenching process parameter set with serial number 5 is used as the optimal quenching process parameter set.

[0067] In order to verify the effectiveness of the spring quenching process optimization method adopted in this embodiment, 1086 springs of the same type that passed the spring appearance defect inspection were collected in this embodiment, and randomly grouped to generate six spring subsets, each of which contained 181 springs; the 181 springs were quenched according to the quenching process parameters of the six quenching process parameter sets; the quenched springs were quality inspected, and the corresponding spring quenching performance test values ​​and the corresponding spring quality pass rates in each of the spring subsets were obtained. The specific results are shown in Table 4.

[0068] Table 4 Example 2 Quenching process optimization verification table

[0069] In Table 4, the spring subset serial numbers and the quenching process parameter set serial numbers correspond one to one. It can be seen from Table 4 that the quenching performance test value of the spring corresponding to the quenching process parameter set with serial number 5 is the largest, and the spring quality test pass rate is no less than the spring quality test pass rates corresponding to the other five quenching process parameter sets. Therefore, the quenching process parameter set with serial number 5 is the optimal quenching process parameter set of the six quenching process parameter sets, that is, the spring quenching process optimization method proposed in this embodiment can effectively obtain the optimal quenching process parameter set.

[0070] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent digital quenching process optimization method for a spring, characterized in that: include: Acquire multiple springs of the same type and construct a first spring detection set; Screening the springs in the first spring detection set to obtain a second spring detection set; Acquire M groups of quenching process parameter sets, and divide the second spring detection set into M groups of spring detection subsets; Quenching the springs of each of the spring detection subsets according to the quenching process parameters of the corresponding quenching process parameter set; Perform spring quenching performance testing on the quenched springs, and obtain the spring quenching performance test value corresponding to each of the spring test subsets, the specific steps comprising: performing surface cleaning on the quenched springs; performing spring appearance defect testing and spring hardness testing on the cleaned springs, respectively, to obtain the spring appearance defect testing comprehensive value and spring hardness testing comprehensive value of the spring test subset; Dividing each of the spring detection subsets into a spring tension detection subset and a spring compression detection subset; Performing a spring stretching test on each spring in the spring stretching test subset to obtain a spring stretching test comprehensive value of the corresponding spring test subset; performing a spring compression test on each spring in the spring compression test subset to obtain a spring compression test comprehensive value of the corresponding spring test subset; Obtaining the spring quenching performance test value of the spring test subset according to the four comprehensive test values ​​corresponding to each of the spring test subsets; According to the spring quenching performance test value, an optimal quenching process parameter set is obtained.

2. The intelligent digital quenching process optimization method for a spring according to claim 1 is characterized in that: The obtaining of a plurality of springs of the same type and constructing a first spring detection set comprises: the springs in the first spring detection set are not subjected to quenching treatment; and each of the springs undergoes the same manufacturing steps before quenching.

3. The intelligent digital quenching process optimization method for a spring according to claim 1 is characterized in that: The step of screening the springs in the first spring test set to obtain the second spring test set includes: performing spring appearance defect detection on the springs in the first spring test set, screening out springs that pass the spring appearance defect detection, and constructing the second spring test set; the spring appearance defect detection process includes: Acquire a spring appearance image of the spring, and input the spring appearance image into a spring appearance defect detection model; The spring appearance defect detection model includes a spring appearance image input layer, a spring appearance feature map acquisition layer, a spring appearance feature map fusion layer and a spring external defect detection value acquisition layer; The spring appearance feature map acquisition layer is used to perform a convolution operation on the spring appearance image to obtain a plurality of the appearance feature maps; the spring appearance feature map fusion layer is used to fuse a plurality of the appearance feature maps to obtain spring appearance features; The spring appearance features include spring shape features and spring defect features; the spring shape features include the maximum diameter of the spring wire, the minimum diameter of the spring wire, the maximum outer diameter of the spring, the minimum inner diameter of the spring and the non-support ring pitch; The spring defect characteristics include spring surface crack characteristics, spring surface foreign matter characteristics and spring surface depression characteristics; Obtaining a first spring defect detection value according to the spring shape feature; obtaining a second spring defect detection value according to the spring defect feature; obtaining a first spring appearance defect detection value according to the first spring defect detection value and the second spring defect detection value; When the first spring appearance defect detection value of the spring is greater than a set defect detection threshold, the spring is determined to have failed the spring appearance defect detection; otherwise, the spring is determined to have passed the spring appearance defect detection.

4. The intelligent digital quenching process optimization method for a spring according to claim 1 is characterized in that: The quenching process parameter sets include: each of the quenching process parameter sets includes N quenching process parameters; and the quenching process parameter values ​​of different quenching process parameter sets are not completely the same.

5. The intelligent digital quenching process optimization method for a spring according to claim 1 is characterized in that: The process of spring stretching test includes: Obtaining a first spring shape feature of each of the springs before performing a spring stretching test; Apply a set tension to the spring. After the moment, cancel the tension; obtain the stretched length of the spring when canceling the tension; After the moment, the second spring shape feature and the spring defect feature of each of the springs are obtained; the spring stretching detection value of the spring is obtained according to the first spring shape feature, the second spring shape feature, the stretching length and the spring defect feature; and the spring stretching detection comprehensive value of the spring detection subset corresponding to the spring stretching detection subset is obtained according to the spring stretching detection value of each spring in the spring stretching detection subset.

6. The intelligent digital quenching process optimization method for a spring according to claim 1 is characterized in that: The process of spring compression detection includes: Obtaining a third spring shape feature of each of the springs before performing a spring compression test; Apply a set pressure to the spring, After the time, the pressure is cancelled; when the pressure is cancelled, the compressed length of the spring is obtained; when the pressure is cancelled After the moment, the fourth spring shape characteristic and the spring defect characteristic of each of the springs are obtained; the spring compression detection value of the spring is obtained based on the third spring shape characteristic, the fourth spring shape characteristic, the compression length and the spring defect characteristic; and the spring compression detection comprehensive value of the spring detection subset corresponding to the spring compression detection subset is obtained based on the spring compression detection value of each spring in the spring compression detection subset.

7. The intelligent digital quenching process optimization method for a spring according to claim 1 is characterized in that: The obtaining of the optimal quenching process parameter set according to the spring quenching performance test value comprises: The quenching process parameter set corresponding to the spring detection subset with the largest spring quenching performance detection value is selected as the optimal quenching process parameter set.

8. An intelligent digital quenching process optimization system for springs, characterized in that: include: A first spring detection set acquisition unit: used to acquire multiple springs of the same type and construct a first spring detection set; A second spring detection set acquisition unit: used for screening the springs of the first spring detection set to acquire a second spring detection set; A quenching spring detection subset division unit is used to obtain M groups of quenching process parameter sets and divide the second spring detection set into M groups of spring detection subsets; Quenching unit: used for quenching the springs of each of the spring detection subsets according to the quenching process parameters of the corresponding quenching process parameter set; A quenching performance detection unit is used to perform a spring quenching performance detection on the quenched spring and obtain a spring quenching performance detection value corresponding to each of the spring detection subsets; the spring quenching performance detection includes a spring appearance defect detection, a spring hardness detection, a spring tension detection and a spring compression detection; The optimal quenching process parameter set acquisition unit is used to acquire the optimal quenching process parameter set according to the spring quenching performance detection value.

9. The intelligent digital quenching process optimization system for springs according to claim 8, characterized in that: The step of obtaining the spring quenching performance test value comprises: performing surface cleaning on the quenched spring; performing spring appearance defect detection and spring hardness detection on the cleaned spring respectively, and obtaining the spring appearance defect detection comprehensive value and spring hardness detection comprehensive value of the spring detection subset; Dividing each of the spring detection subsets into a spring tension detection subset and a spring compression detection subset; Performing a spring stretching test on each spring in the spring stretching test subset to obtain a spring stretching test comprehensive value of the corresponding spring test subset; Performing spring compression testing on each spring of the spring compression testing subset to obtain a spring compression testing comprehensive value of the corresponding spring testing subset; The spring quenching performance test value of the spring test subset is obtained according to the four comprehensive test values ​​corresponding to each of the spring test subsets.

10. The intelligent digital quenching process optimization system for springs according to claim 9, characterized in that: The process of spring stretching test includes: Obtaining a first spring shape feature of each of the springs before performing a spring stretching test; Apply a set tension to the spring. After the moment, cancel the tension; obtain the stretched length of the spring when canceling the tension; After the moment, the second spring shape feature and the spring defect feature of each of the springs are obtained; the spring stretching detection value of the spring is obtained according to the first spring shape feature, the second spring shape feature, the stretching length and the spring defect feature; and the spring stretching detection comprehensive value of the spring detection subset corresponding to the spring stretching detection subset is obtained according to the spring stretching detection value of each spring in the spring stretching detection subset.

Images