Mechanical compensation value calculation method and bending machine numerical control system

Abstract

The invention discloses a mechanical compensation value calculation method and a bending machine numerical control system.The mechanical compensation value calculation method comprises the steps that firstly, a theoretical model is built based on the stress condition of the bending machine in the bending process, deflection curves of a sliding block and the end face of a workbench are solved, deflection curve expressions of the upper end face of the workbench and the sliding block are obtained, and the working condition of uneven stress of the bending machine is considered; further optimizing the theoretical model by analyzing the trend of the deflection curve and the position change of the symmetry axis of the deflection curve, abstracting a deflection curve expression, that is, each item in the expression has an unknown coefficient, performing search optimization by using a genetic algorithm, and determining each item of coefficient; according to the calculation method and system, the deflection change conditions of the bending machine at different bending positions are considered and analyzed, the calculation of the overall deflection deformation and the compensation value of the bending machine is more accurate according to an actual deflection curve optimization theoretical calculation formula, previous bending data are learned through a swarm intelligence optimization algorithm, and the bending accuracy of the bending machine is improved. And continuously optimizing mechanical compensation calculation model parameters.

Description

Technical Field

[0001] The invention relates to the technical field of bending machine control, and in particular to a mechanical compensation value calculation method and a bending machine numerical control system. Background Art

[0002] The flexural deformation of the bending machine slider and worktable caused by the force during the working process is one of the main factors affecting the bending accuracy. At the same time, mechanical compensation is a widely used method to eliminate the influence of flexural deformation. Therefore, the control accuracy of the mechanical compensation device of the bending machine control system is crucial to the bending accuracy. The main factors affecting the control accuracy of the mechanical compensation device are composed of mechanical structure and mechanical compensation value. Among them, there are many mature technologies and patents for mechanical structure, but there are few patents and papers on the calculation method of mechanical compensation value in China. At present, the research on the calculation of mechanical compensation value of bending machine in China is mainly divided into two categories: theoretical model solution and finite element analysis:

[0003] (1) The paper "Control Implementation of Deflection Compensation of Press Brake System Based on CPAC" (Xu Xiaobin, Tian Bin, Qiu Yue, et al. Modern Manufacturing Technology and Equipment, 2019, (01): 70-72. DOI: 10.16107 / j.cnki.mmte.2019.0034) simplified the machine tool structure into a simply supported beam and constructed a theoretical model of the bending deformation of the press brake. The deformation and compensation were obtained by solving the theoretical model. This method met the compensation requirements of the press brake for the deflection deformation during the processing and improved the processing accuracy of the press brake. However, the actual structure of the machine tool, the rigidity conditions of the machine tool, and the changes in related key dimensions at different customers often lead to large deviations between the calculated value of the theoretical model and the actual compensation value, increasing the debugging cost;

[0004] (2) The paper “Finite element analysis and disturbance compensation calculation of bending machine frame deflection” (Zhang Xianjin, Zhang Yifan, Wang Qi. Mechanical Research and Application, 2019, 32(05): 12-13+18. DOI: 10.106756 / j.cnki.1007-4414.2019.05.004) and the paper “Optimization of mechanical compensation device of 100t bending machine based on ANSYS Workbench” (Zhan Shaowei, Gong Junjie, Wei Yuan Source. Journal of Yangzhou University (Natural Science Edition), 2023, 26(04): 13-18+36. DOI: 10.19411 / j.1007-824x.2023.04.003) The finite element analysis method is used to analyze the deformation of the machine tool and compensate for it. However, in actual industrial applications, the deflection of the bending machine is different for different parts and different bending steps. It is unreasonable to perform finite element analysis on the force of the bending machine before each bending.

[0005] In summary, the existing calculation method (1) still has the disadvantage of weak adaptability and difficulty in applying to bending machines of different working conditions, and the calculation method (2) has the disadvantage of being difficult to actually apply in industrial production. Therefore, in view of the above shortcomings, the present invention provides a mechanical compensation value calculation method and a bending machine numerical control system, which combines a theoretical model with a swarm intelligence optimization algorithm, has strong adaptability, and the compensation value calculation accuracy is high. Summary of the invention

[0006] In order to solve the above problems, the present invention provides a mechanical compensation value calculation method and a bending machine numerical control system. By combining the theoretical model with the swarm intelligence optimization algorithm, the shortcomings of the existing calculation method, such as low flexibility and poor adaptability, are avoided, and the accuracy of the compensation value calculation can be further improved, thereby improving the bending precision. The bending machine numerical control system adjusts the coefficients of the deflection curve expression through a learning optimization module, so that the calculation result is closer and closer to the actual compensation value that meets the bending precision.

[0007] The technical solution adopted by the present invention is:

[0008] The present invention provides a method for calculating a mechanical compensation value, which is applied to a bending machine numerical control system. The method comprises the following steps:

[0009] S1: Based on the stress conditions of the bending process of the bending machine, the machine tool is simplified, a theoretical model is constructed, and the deflection curves of the slider and the end surface of the worktable are solved through basic material mechanics and elastic mechanics to obtain the deflection curve expressions of the end surface of the worktable and the slider. The specific steps are as follows:

[0010] S11: Assume that the length of the workbench is l 台 , height h 台 The upper end face is subjected to the uniform load q, and the workbench support is located at the left and right ends, and the support reaction force is ql 台 / 2, and the stress on the workbench is σ x , σ y , τ xy The stress should satisfy the following boundary conditions:

[0011]

[0012] S12: Since there is no axial force and bending moment on the top of the workbench, The following boundary conditions are also satisfied:

[0013]

[0014] S13: Assume that the stress function of the workbench is but With σ x , σ y , τxy The relationship is:

[0015]

[0016] S14: Since the workbench is subjected to uniform load, for different x, σ y The distribution is the same, that is, σ y Only related to y, at the same time, the x-axis is symmetrical about the center line of the workbench, σ x Symmetric about the y-axis, so in σ x The coefficients of the odd-order terms of x in the expression should be zero, and the stress function is Expressed as:

[0017]

[0018] At the same time, based on the relevant theory of elastic mechanics, the stress function satisfies the biharmonic equation, and for All are established, that is:

[0019]

[0020] S15: Substitute the boundary conditions of step S1 and step S2 into the solution to obtain:

[0021]

[0022] At the same time, under the theory of material mechanics and elastic mechanics, the deformation of the model also satisfies the following geometric equations and physical equations:

[0023]

[0024] Among them, E 台 is the elastic modulus of the workbench, μ 台 is the Poisson's ratio of the workbench, and the moment of inertia I is introduced 台 , we can finally obtain the theoretical model. That is, the deflection curve expression at the end surface of the workbench is:

[0025]

[0026] Among them, E 台 is the elastic modulus, the moment of inertia I 台 is the moment of inertia, and the length of the workbench is l 台 , height h 台 .

[0027] S2: Considering the working condition of uneven stress on the bending machine, the theoretical model is further optimized by analyzing the trend of the deflection curve and the position change of its symmetry axis, and the deflection curve expression further optimized based on step S1 is obtained. The specific steps are as follows:

[0028] S21: Simplify the workbench into a simply supported beam model. The fulcrums are fixed hinge supports at both ends of the beam, namely point A and point B. The distance between point A and point B is l 台 , the support reaction forces at point A and point B are F A and F B ;

[0029] S22: During bending, there is a section of the beam with a uniformly distributed load q. Let the length of this section of the beam be b. There is an equivalent effective point on the beam. This point is in the middle of the beam. The distances between this point and point A and point B are l1 and l2, respectively. Let the distance from the equivalent point to the left support be The distance from the right fulcrum is

[0030] S23: Perform stress analysis on the beam and obtain F A and F B About q, b, l 台 and the relation of l1 or l2;

[0031] S24: Substitute the boundary conditions of the shear equation for a beam of length b into the equation. The boundary conditions are: Substituting y=0, we can solve the coordinates of point P as follows: is the symmetry axis of the quartic curve, and the deflection curve expression at the end surface of the workbench in step S1 is further optimized as follows:

[0032]

[0033] In the formula, E 台 is the elastic modulus, the moment of inertia I 台 is the moment of inertia, and the length of the workbench is l 台 , height h 台 ;

[0034] The theoretical model of S1 and the simply supported beam model of this step are unified in the x-axis coordinate system to obtain According to this equation, the position where the deflection is maximum in the actual bending of the workbench can be obtained.

[0035] S3: The deflection curve expressions of the workbench and the slider after further optimization in step S2 are superimposed and abstracted, that is, each term in the formula has an unknown coefficient, and the coefficients to be optimized are determined. The force and structural performance related parameters of the bending machine currently in use are taken as known quantities to reversely solve the actual coefficients in the calculation formula, and the swarm intelligence algorithm is used for search optimization to determine the coefficients. The specific steps are as follows:

[0036] S31: The abstracted deflection curve expression is:

[0037]

[0038] Where v is the superposition of the deformation of the machine tool slide and the worktable, that is, the actual compensation value, E 台 is the elastic modulus of the workbench, the moment of inertia I 台 is the inertia moment of the workbench, q 台 is the load applied to the workbench, the length of the workbench is l 台 , the height of the workbench is h 台 , E 滑 is the elastic modulus of the slider, the moment of inertia I 滑 is the moment of inertia of the slider, q 滑 is the load applied to the slider, the slider length is l 滑 , the height of the slider is h 滑 ;

[0039] S32: When bending, v, E 台 ,I 台 ,q 台 , l 台 、h 台 , P′ 台 、E 滑 ,I 滑 , P′ 滑、 q 滑 , l 滑 and h 滑 All are known, and the coefficients K1-K in the abstracted deflection curve expression are obtained according to the genetic algorithm. 13 ,The swarm intelligence algorithm adopts genetic algorithm, and the steps of using genetic algorithm are as follows:

[0040] S321: Initialize and set basic parameters related to the genetic algorithm, including population size, number of iterations, crossover rate, and mutation rate;

[0041] S322: Design the chromosome encoding, the encoding method is: according to the coefficients K1-K in the step deflection curve expression 13 Generate a theoretical chromosome vector, and generate an initial chromosome population by making random differences in the values ​​of each element in the vector;

[0042] S323: Generate an initial population based on the theoretical model optimized in step S2, calculate the fitness value of the population, sort the fitness of individual chromosomes, and select better chromosomes to enter the next generation;

[0043] S324: performing adaptive crossover operation and adaptive mutation operation, decoding the chromosome, and calculating and evaluating the fitness of the new population to generate a new population;

[0044] Specifically, step S3241: the mutation rate is adaptively adjusted according to the formula. When the individual fitness is higher than the average fitness of the population, the mutation rate is adaptively adjusted. When the individual fitness is lower than the average fitness of the population, a fixed value with the maximum mutation probability is given. The specific formula is as follows:

[0045]

[0046] Where p m is the mutation rate, p m min is the minimum mutation rate, p m max is the maximum mutation rate, F max is the maximum fitness of the population, F b is the individual fitness of the mutated chromosome, F v is the average fitness of the population;

[0047] Step S3242: Adaptively adjust the crossover rate according to the formula. According to the concentration and dispersion degree of fitness in the population evolution process, the crossover rate of the genetic algorithm is nonlinearly adaptively adjusted, and a two-point crossover method is used to generate new individuals. When the individual fitness is higher than the average fitness of the population, the crossover rate will be adaptively adjusted. When the individual fitness is lower than the average fitness of the population, a maximum fixed value will be given to the crossover rate. The specific formula is as follows:

[0048]

[0049] Where p c is the crossover rate, p c min is the minimum crossover rate, p c max is the maximum crossover rate, F a It is the chromosome with greater fitness during the crossover operation.

[0050] S325: If the algorithm reaches the set maximum number of iterations or the optimization effect reaches the termination condition, the program ends; otherwise, repeat steps S323-S325.

[0051] Step S4: Taking the error between the theoretical compensation value and the actual compensation value as the standard for fitness evaluation, construct a fitness function:

[0052] In the formula, v i Indicates the actual compensation value that meets the bending process requirements during the bending process, v i ′ represents the superposition of the deformation values ​​of the slider and the workbench, i.e. the theoretical compensation value, with a total of n sets of data, v i With v i′ is obtained under the condition that the material, size, bending angle and bending position of the sheet remain unchanged. minF is the mean square error between the theoretical compensation value set and the actual compensation value set. The closer the mean square error is to 0, the closer the calculated value curve is to the theoretical value curve as a whole.

[0053] The actual compensation value and the theoretical compensation value are obtained in practical application as follows:

[0054] S41: Obtain theoretical compensation values ​​for the first time to generate an initial population: Substitute machine tool dimensions l1, l2, h1, h2, machine tool performance parameters I1, I2, E1, E2, and uniformly distributed loads q1, q2 into the expression after the deflection curve in step S2 is further optimized, and superimpose the deformation of the machine tool slide and the worktable;

[0055] S42: when the deflection curve expression needs to be optimized again, a new calculated value is generated according to step S3;

[0056] S43: After the machine tool is debugged, during the actual bending process, if the machine tool is subjected to large force and does not compensate for the machine tool deflection or only relies on theoretical compensation values ​​for compensation, there will be compensation errors. While ensuring that the sheet material, size, bending angle and bending position remain unchanged, gradually adjust the mechanical compensation value until the angles at both ends and the middle of the sheet are within the tolerance range required by the process. At this time, the adjusted specific compensation value is the actual compensation value.

[0057] The present invention also provides a bending machine numerical control system, comprising a numerical simulation module, and the numerical simulation module executes the above-mentioned mechanical compensation value calculation method.

[0058] Furthermore, a learning optimization module is included, which is used to save the actual compensation value and the theoretical compensation value calculated by the numerical simulation module. The learning optimization module learns and optimizes according to the error between the actual compensation value and the theoretical compensation value, and updates K1-K in step S3. 13 Coefficient, perform the next bending operation after updating the parameters. If the compensation is still not up to standard after the theoretical compensation value is used, manually adjust the mechanical compensation value until the bending angle meets the standard. The manually adjusted compensation value is the actual compensation value.

[0059] Beneficial effects of the present invention:

[0060] 1. Compared with the existing compensation value calculation method, the calculation method provided by the present invention fully considers and analyzes the change of the deflection of the bending machine at different bending positions in actual industrial applications, and optimizes the theoretical calculation formula according to the actual deflection curve, so as to calculate the overall deflection deformation and compensation value of the bending machine more accurately.

[0061] 2. The calculation method provided by the present invention combines the theoretical model with the swarm intelligence optimization algorithm, thereby avoiding the problem that the finite element analysis in the existing research scheme is difficult to integrate into the numerical control system and the problem that the calculated value is inaccurate when the theoretical model is used directly, thus affecting the bending accuracy. It is easy to implement in engineering and has obvious application value.

[0062] 3. The bending machine numerical control system provided by the present invention compares and analyzes the theoretical calculation value with the actual standard value, and learns the previous bending data through the group intelligent optimization algorithm, thereby continuously optimizing the parameters of the mechanical compensation calculation model, greatly reducing the workload of manual adjustment under the traditional scheme, and because the algorithm learning is based on the previous bending data of the current bending equipment, the present invention does not rely on the experience of the workers. Therefore, the scheme of the present invention has a strong adaptability to bending machines of different models, sizes and mechanical structures.

[0063] 4. The bending machine numerical control system provided by the present invention can effectively utilize a large amount of process data generated during the bending process, and learn the data through a swarm intelligent optimization algorithm, thereby continuously optimizing the parameters of the mechanical compensation calculation model, greatly reducing the workload of manual adjustment under the traditional scheme. As the number of workpieces processed by users increases, the system can collect more data. Based on the method proposed by the present invention, the calculation formula can be optimized multiple times to achieve the beneficial effect of "more and more accurate bending" for users. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 Schematic diagram of deformation of metal sheet according to Example 1 of the present invention.

[0065] Figure 2 Schematic diagram of bending deformation of a bending machine according to Embodiment 1 of the present invention, wherein Figure 2 (a) is the bending deformation diagram of the bending machine. Figure 2 (b) Ideal deformation diagram after adding compensation for the bending machine.

[0066] Figure 3 This is a theoretical model diagram of the force analysis of the bending machine in Example 1 of the present invention.

[0067] Figure 4 It is a schematic diagram of the force analysis of the simply supported beam model of Example 1 of the present invention.

[0068] Figure 5 This is a shear force schematic diagram of a simply supported beam model according to Example 1 of the present invention.

[0069] Figure 6 This is a flow chart of the genetic algorithm used in Example 1 of the present invention.

[0070] Figure 7 This is a schematic diagram of the working process of the bending machine numerical control system according to embodiment 2 of the present invention. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical solution and advantages of the present invention more clear, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings and a preferred implementation manner.

[0072] Example 1

[0073] refer to Figure 1 and Figure 2 When bending metal sheets, the bending machine will bend due to the mechanical properties of the metal sheet itself. After bending, the bending machine will also be deformed. The bending angles of the two ends of the sheet will be equal (α1=α3), and the bending angle in the middle will exceed the production processing standard (α1=α3≠α2). At this time, the bending deformation produced by the bending machine is as follows: Figure 2 As shown in (a), the ideal deformation after adding compensation is as follows Figure 2 (b) As shown, how to quickly obtain the compensation value close to the ideal deformation is particularly important. Therefore, this embodiment provides a method for calculating the mechanical compensation value. By combining the theoretical model with the swarm intelligence optimization algorithm, the shortcomings of the existing calculation method such as low flexibility and poor adaptability are avoided. At the same time, the accuracy of the compensation value calculation can be further improved, thereby improving the bending accuracy.

[0074] refer to Figure 3-Figure 6 , which is Embodiment 1 of the present invention, and provides a method for calculating a mechanical compensation value, which is applied to a bending machine tool, and the method steps are as follows:

[0075] S1: The slider of the bending machine cooperates with the workbench to bend the plate. The slider and the workbench are subjected to force and flexural deformation. Based on the force conditions of the bending process of the bending machine, the machine tool is simplified, a theoretical model is constructed, and the deflection curves of the slider and the end face of the workbench are solved through basic material mechanics and elastic mechanics to obtain the deflection curve expressions of the end face of the workbench and the slider.

[0076] Since the bending deformation trend of the bending machine slider and the workbench is the same, only the workbench is analyzed in the theoretical model, and the slider deformation can be solved by the same method. Figure 3 , the steps are as follows:

[0077] S11: Assume that the length of the workbench is l 台 , height h 台 The upper end face is subjected to the uniform load q, and the workbench support is located at the left and right ends, and the support reaction force is ql 台 / 2, and the stress on the workbench is σ x , σ y , τ xy The stress should satisfy the following boundary conditions:

[0078]

[0079] S12: Since there is no axial force and bending moment on the top of the workbench, The following boundary conditions are also satisfied:

[0080]

[0081] S13: Assume that the stress function of the workbench is but With σ x , σ y , τ xy The relationship is:

[0082]

[0083] S14: Since the workbench is subjected to uniform load, for different x, σ y The distribution is the same, that is, σ y Only related to y, at the same time, the x-axis is symmetrical about the center line of the workbench, σ x Symmetric about the y-axis, so in σ x The coefficients of the odd-order terms of x in the expression should be zero, and the stress function is It can be expressed as follows:

[0084]

[0085] At the same time, based on the relevant theory of elastic mechanics, the stress function satisfies the biharmonic equation, and for All are established, that is:

[0086]

[0087] S15: Substitute the boundary conditions of step S1 and step S2 into the solution to obtain:

[0088]

[0089] At the same time, under the theory of material mechanics and elastic mechanics, the deformation of the model also satisfies the following geometric equations and physical equations:

[0090]

[0091] Among them, E 台 is the elastic modulus of the workbench, μ 台 is the Poisson's ratio of the workbench, which is generally taken as 0.3, and the moment of inertia I is introduced 台 , we can finally obtain the theoretical model. That is, the deflection curve expression at the end surface of the workbench is:

[0092]

[0093] Among them, E 台 is the elastic modulus, the moment of inertia I 台 is the moment of inertia, and the length of the workbench is l 台 , height h 台 .

[0094] It can be seen from the above expression that when the load, relevant material parameters of the workbench and its external dimensions are known, the deflection curve of the upper end surface of the workbench is a quartic curve of x (in fact, through different simplified methods and theoretical models, such as the simply supported beam model or the Timoshenko beam model, the curves solved are generally quartic curves about x), which is an important reason affecting the bending accuracy of the press brake.

[0095] In the actual bending process, the following situations often occur: (1) The upper and lower dies of the bending machine are arranged in sections, and the lengths of the upper and lower dies are both shorter than the length of the machine tool; (2) The sheet metal is not placed in the middle of the machine tool during bending (asymmetric bending). Figure 3 The ideal model established in the paper bending machine is different, so the working condition of uneven stress of the bending machine needs to be considered to make the compensation value result more accurate.

[0096] S2: Considering the working condition of uneven stress on the bending machine, the theoretical model is further optimized by analyzing the trend of the deflection curve and the position change of its symmetry axis, and the deflection curve expression further optimized based on step S1 is obtained. The specific steps are as follows:

[0097] S21: From the conclusion in S2, it can be seen that under different theoretical models, the deformation trends of machine tools are similar quartic curves. Therefore, in order to facilitate analysis and calculation, the machine tool worktable is simplified to a simply supported beam model. The force analysis diagram is shown in Figure 4 As shown, the support points are fixed hinge supports at both ends of the beam, namely point A and point B, and the distance between point A and point B is l 台 , the support reaction forces at point A and point B are F A and F B ;

[0098] S22: When bending, there is a section of the beam with a uniformly distributed load q. Assuming the length of this section of the beam is b, the force on the beam is simplified and analyzed, and the following can be obtained: Figure 4 The force diagram shown in the lower part, where qb is the equivalent force of the uniformly distributed load, the point of action is At the point, the distances between the action point and point A and point B are obtained as l1 and l2 respectively. Suppose the distance between the action point and the left support after equivalent is The distance between the equivalent action point and the right support point is

[0099] S23: Perform stress analysis on the beam and obtain F A and F B About q, b, l 台 And the relationship between l1 and l2, the force balance analysis of the beam can be obtained:

[0100] S24: By consulting the internal force and deformation table of a single-span beam, its shear force diagram can be obtained as follows: Figure 5 As shown. Figure 5 It can be seen that the shear force distribution function of the beam is composed of two constant functions and one linear function. According to the calculus relationship between the shear force distribution function and the deflection deformation function, the deflection curve is a curve with the highest order of fourth order. This trend is the same as the trend obtained by the theoretical model in S2, and they can confirm each other. At this time, the maximum deflection is not in the middle of the beam, but appears at Figure 5 The shear force is zero, that is, point P.

[0101] Finding point P process: According to Figure 5 , the shear force equation of section b is a first-order equation, and the boundary conditions are: Substituting y=0, we can solve the coordinates of point P as follows: The coordinates of point P can be solved as: Among them, Figure 4 The geometric relationship in can be determined. The length of l2 must be less than l, so we get This means that point P always falls on the uniformly distributed load segment. It is also the symmetry axis of the quartic curve. The deflection curve expression at the end surface of the workbench in step S1 is further optimized as follows:

[0102]

[0103] In the formula, E 台 is the elastic modulus, the moment of inertia I 台 is the moment of inertia, and the length of the workbench is l 台 , height h 台 ;

[0104] The x-axis coordinate system of the theoretical model of S1 and the simply supported beam model of this step is unified, and it is Figure 3 and Figure 4 It can be seen that the origin of the X-axis coordinate of the theoretical model in S2 is in the middle of the machine tool, and the origin of the X-axis coordinate of the simply supported beam model in S3 is in the middle of the machine tool and on the left side of the machine tool. To unify the X-axis coordinate system to satisfy the deflection curve expression, the origin of the X-axis coordinate is uniformly moved to the middle of the machine tool. Then the X-axis coordinate of point P needs to be reduced by l / 2, that is, P′ 台 , to obtain According to this equation, the position where the deflection is maximum in the actual bending of the workbench can be obtained.

[0105] Based on the above optimized deflection curve expression, as long as the relevant material properties and external dimensions of the slider and the workbench are determined, the maximum deflection of the bending machine under theoretical conditions can be obtained, so as to adjust the actual compensation value according to the bending position and the maximum deflection. However, in actual working conditions, the shapes of the slider and the workbench of different types of bending machines are different. The maximum deflection simply obtained by the theoretical model is still not completely applicable to various working conditions, and it cannot meet the high-precision requirements under specific circumstances. At this time, it takes a lot of time to adjust the machine tool parameters in the CNC system. Therefore, based on the theoretical model, the present invention expresses the deflection curve as a quartic curve about x, and automatically searches for various coefficients through an improved genetic algorithm.

[0106] S3: From the analysis of S1 and S2, it can be determined that the deflection curve of the bending machine is a quartic curve under different working conditions, but the coefficients in the theoretical model should change accordingly with the changes in the machine tool model, machine tool performance, and machine tool structure differences. Therefore, the present invention regards the stress condition of the bending machine and the actual deflection compensation value as known quantities. The actual deflection compensation value is the compensation value that enables the bending machine to reach an ideal compensation state after bending. The coefficients in the optimized deflection curve expression are regarded as unknown quantities, and a genetic algorithm is used for search optimization to meet the actual mechanical compensation value calculation requirements. The deflection curve expression after the workbench and the slider are further optimized in step S2 is The expression is superimposed and abstracted, that is, each term in the expression has an unknown coefficient, and the coefficients to be optimized are determined. According to the force and structural performance related parameters of the bending machine currently in use, the actual coefficients in the calculation formula are reversely solved, and the swarm intelligence algorithm is used for search optimization. Swarm intelligence includes a variety of algorithms. The swarm intelligence algorithm in this embodiment uses a genetic algorithm to determine the coefficients. Since the genetic algorithm is the most widely used, the most mature, and the easiest to implement, the swarm intelligence algorithm in this embodiment uses a genetic algorithm to determine the coefficients. Other algorithms such as particle swarm algorithm, artificial bee colony algorithm, ant lion optimization algorithm, etc. can also be used. The specific steps are as follows:

[0107] S31: The abstracted deflection curve expression is:

[0108]

[0109] Where v is the superposition of the deformation of the machine tool slide and the worktable, that is, the actual compensation value, E 台 is the elastic modulus of the workbench, the moment of inertia I 台 is the inertia moment of the workbench, q 台 is the load applied to the workbench, the length of the workbench is l 台 , the height of the workbench is h 台 , E滑 is the elastic modulus of the slider, the moment of inertia I 滑 is the moment of inertia of the slider, q 滑 is the load applied to the slider, the slider length is l 滑 , the height of the slider is h 滑 ;

[0110] S32: When bending, v, E 台 ,I 台 ,q 台 , l 台 、h 台 , P′ 台 、E 滑 ,I 滑 , P′ 滑、 q 滑 , l 滑 and h 滑 All are known, and the coefficients K1-K in the abstracted deflection curve expression are obtained according to the genetic algorithm. 13 ,like Figure 6 As shown, the specific steps of using genetic algorithm are as follows:

[0111] S321: Initialize and set basic parameters related to the genetic algorithm (population size, number of iterations, crossover rate, mutation rate, etc.), the parameters include population size, number of iterations, crossover rate and mutation rate;

[0112] S322: Design the chromosome encoding, the encoding method is: according to the coefficients K1-K in the deflection curve expression obtained in step S2 13 Generate a theoretical chromosome vector, and generate an initial chromosome population by making random differences in the values ​​of each element in the vector;

[0113] S323: Based on the theoretical model of the deflection curve expression further optimized in step S24, an initial population is generated according to each coefficient in the expression, the fitness value of the population is calculated, and the fitness of the chromosome individuals is sorted, and the better chromosomes are selected to enter the next generation;

[0114] S324: performing adaptive crossover operation and adaptive mutation operation, decoding the chromosome, and calculating and evaluating the fitness of the new population to generate a new population;

[0115] The crossover rate and mutation rate of traditional genetic algorithms are both fixed values, but different crossover rates and mutation rates directly affect the accuracy and efficiency of the algorithm. Therefore, in order to enable the algorithm to better jump out of local extreme values ​​and search for global optimal solutions, based on the stability of population fitness changes, an adaptive optimization mechanism is introduced to adaptively adjust the crossover and mutation operators, which can optimize the algorithm to improve algorithm performance and quickly converge to reduce the deviation between the theoretical compensation value and the actual compensation value. The specific steps are as follows:

[0116] Step S3241: setting the adaptive adjustment mutation rate according to the formula, the specific formula is as follows:

[0117]

[0118] Where p m is the mutation rate, p m min is the minimum mutation rate, p m max is the maximum mutation rate, F max is the maximum fitness of the population, F b is the individual fitness of the mutated chromosome, F v is the average fitness of the population;

[0119] From the adaptive adjustment mutation rate formula, we can know that when the individual fitness is lower than the average fitness of the population, F b <F v The condition will adaptively adjust the mutation rate, that is, the individual mutation rate will be assigned the maximum mutation rate p mmax This is a fixed value set when the algorithm is initialized, and it is also the maximum value of the mutation rate. This is done because when the fitness of an individual is lower than the average fitness of the population, it means that the solution represented by this individual is poor, and a larger mutation probability is needed to make this solution more likely to mutate and produce a better solution.

[0120] Step S3242: Adaptively adjust the crossover rate according to the formula. According to the concentration and dispersion degree of fitness in the population evolution process, the crossover rate of the genetic algorithm is nonlinearly and adaptively adjusted, and a new individual formula is generated by a two-point crossover method, which is conducive to changing the individual structure to a greater extent. The specific formula is as follows:

[0121]

[0122] Where p c is the crossover rate, p cmin is the minimum crossover rate, p cmax is the maximum crossover rate, F a It is the chromosome with greater fitness during the crossover operation.

[0123] From the above adaptive crossover rate formula, we can see that when the individual fitness is lower than the average fitness of the population, F a <F v The crossover rate will be adjusted adaptively under the condition that the individual mutation rate will be assigned the maximum mutation rate p c max This is a fixed value set when the algorithm is initialized, and it is also the maximum value of the crossover rate. This is done because when the fitness of an individual is lower than the average fitness of the population, it means that the solution represented by this individual is poor, and a larger mutation probability is needed to make this solution more likely to mutate and produce a better solution.

[0124] S325: If the algorithm reaches the set maximum number of iterations or the optimization effect reaches the termination condition, the program ends; otherwise, S323-S325 are repeated.

[0125] Based on the above method, in order to further optimize the target and minimize the gap between the calculated deflection value and the actual value, the present invention constructs a fitness function, and uses the size of the error between the theoretical compensation value and the actual compensation value as the standard for fitness evaluation, such as step S4:

[0126] Step S4: Construct fitness function:

[0127] In the formula, v i Indicates the actual compensation value that meets the bending process requirements during the bending process, v i ′ represents the superposition of the deformation values ​​of the slider and the workbench, i.e. the theoretical compensation value, with a total of n sets of data, v i With v i ′ is one-to-one correspondence, which is obtained by ensuring that the material, size, bending angle and bending position of the sheet remain unchanged. minF is the mean square error of each item in the theoretical compensation value set and the actual compensation value set. The closer the mean square error is to 0, the greater the error is. In this embodiment, 1×10 -3 As the termination condition, the calculated value curve is closer to the theoretical value curve as a whole, so that a reasonable mechanical compensation value can be calculated when the machine tool is under different force conditions.

[0128] The actual compensation value and the theoretical compensation value are obtained in practical application as follows:

[0129] S41: Obtain theoretical compensation values ​​for the first time to generate an initial population: Substitute machine tool dimensions l1, l2, h1, h2, machine tool performance parameters I1, I2, E1, E2, and uniformly distributed loads q1, q2 into the expression after the deflection curve in step S2 is further optimized, and superimpose the deformation of the machine tool slide and the worktable;

[0130] S42: When the deflection curve expression needs to be optimized again, a new calculated value is generated according to step S3, and K1 to K2 generated by the previous optimization are used. 13 Generate a new calculated value;

[0131] S43: After the machine tool is debugged, in the actual bending process, if the machine tool is subjected to large force and does not compensate for the machine tool deflection or only relies on the theoretical compensation value for compensation, there is a compensation error, then the following may occur: Figure 1 The situation shown is that the angles at both ends of the sheet meet the actual process requirements, and the angle in the middle will become larger. Under the condition that the sheet material, size, bending angle and bending position remain unchanged, gradually adjust the mechanical compensation value until the angles at both ends and the middle of the sheet are within the tolerance range required by the process. At this time, the specific compensation value after adjustment is the actual compensation value.

[0132] Compared with the existing compensation value calculation method, this embodiment, based on step S1 and step S2, fully considers and analyzes the change of the bending machine deflection at different bending positions in actual industrial applications, and optimizes the theoretical calculation formula according to the actual deflection curve, so that the calculation of the overall bending deformation and compensation value of the bending machine is more accurate, and through step S3, the theoretical model is combined with the group intelligent optimization algorithm to avoid the problem that the finite element analysis in the existing research scheme is difficult to integrate into the numerical control system and the calculation value is inaccurate when the theoretical model is directly used, thus affecting the bending accuracy. It is easy to implement in engineering and has obvious application value; through step S4, the theoretical calculation value is compared and analyzed with the actual standard value, and the previous bending data is learned through the group intelligent optimization algorithm, so as to continuously optimize the parameters of the mechanical compensation calculation model, which greatly reduces the workload of manual adjustment under the traditional scheme, and because the algorithm learning is based on the previous bending data of the current bending equipment, the present invention does not rely on the experience of workers. Therefore, the scheme described in the present invention has a strong adaptability to bending machines of different models, sizes and mechanical structures.

[0133] Example 2

[0134] The present invention also provides a bending machine numerical control system, including a numerical simulation module and a learning optimization module, the numerical simulation module is used to execute the mechanical compensation value calculation method described in Example 1, and set relevant operation buttons on the interactive interface, the learning optimization module is used to save the actual compensation value and the theoretical compensation value calculated by the numerical simulation module, the learning optimization module learns and optimizes according to the error between the actual compensation value and the theoretical compensation value, and updates K1-K in step S3 13 Coefficient, perform the next bending operation after updating the parameters. If the compensation is still not up to standard after the theoretical compensation value is used, manually adjust the mechanical compensation value until the bending angle meets the standard. The manually adjusted compensation value is the actual compensation value.

[0135] refer to Figure 7 Before the system starts bending, it will continuously obtain bending step information such as the bending angle of the sheet metal, and obtain the machine tool structure parameters and material parameters. The workflow of this system is as follows:

[0136] Step S1: During the first bending, the numerical simulation module of the system will calculate the theoretical compensation value according to the deflection curve expression obtained in step S1 or the further optimized deflection curve expression obtained in step S2.

[0137] Step S2: Determine whether the bending angle is accurate. If accurate, store the calculated value in the system's internal theoretical calculated value database, and copy the calculated value to the CNC system's internal actual value database. If inaccurate, the user manually adjusts the deflection compensation value until the bending angle meets the standard, and stores the calculated value in the system's internal theoretical calculated value database, and also stores the actual calculated value in the system's internal actual value database.

[0138] Step S3: After collecting a certain amount of data, the user can use the relevant operation buttons to learn the saved data and optimize and update the coefficient parameters in the deflection curve expression of step S1 or S2. After updating the coefficient parameters, the next bending can be performed.

[0139] After collecting a certain amount of data, the system of the present invention can optimize the calculation formula obtained in Example 1 multiple times according to actual needs, continuously reduce the deviation between the mechanical compensation calculation value and the actual value, and improve the bending accuracy. The bending machine CNC system adjusts the various coefficients of the deflection curve expression through the learning optimization module, so that the calculation result is closer and closer to the actual compensation value that meets the bending accuracy.

[0140] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications are also within the protection scope of the present invention.

Claims

1. A method for calculating a mechanical compensation value, characterized in that: Applied in the CNC system of a bending machine, the method comprises the following steps: S1: Based on the stress conditions of the bending process of the bending machine, the machine tool is simplified, a theoretical model is constructed, and the deflection curves of the slider and the end surface of the worktable are solved through basic material mechanics and elastic mechanics to obtain the deflection curve expressions of the end surface of the worktable and the slider; S2: Considering the working condition of uneven stress on the bending machine, the theoretical model is further optimized by analyzing the trend of the deflection curve and the position change of its symmetry axis, and the deflection curve expression further optimized based on step S1 is obtained; S3: The deflection curve expressions of the workbench and slider after further optimization in step S2 are superimposed and abstracted, that is, each term in the formula has an unknown coefficient, and the coefficients to be optimized are determined. The force and structural performance related parameters of the bending machine currently in use are taken as known quantities to reversely solve the actual coefficients in the calculation formula, and the swarm intelligence algorithm is used for search optimization to determine the coefficients.

2. The method for calculating the mechanical compensation value according to claim 1, characterized in that: The step S4 is also included: taking the error between the theoretical compensation value and the actual compensation value as the standard for fitness evaluation, and constructing a fitness function: In the formula, v i Indicates the actual compensation value that meets the bending process requirements during the bending process, v i ′ represents the superposition of the deformation values ​​of the slider and the workbench, i.e. the theoretical compensation value, with a total of n sets of data, v i With v i ′ is one-to-one correspondence, which is obtained when the material, size, bending angle and bending position of the sheet material remain unchanged. min F is the mean square error between the theoretical compensation value set and the actual compensation value set. The closer the mean square error is to 0, the closer the calculated value curve is to the theoretical value curve as a whole.

3. The method for calculating the mechanical compensation value according to claim 1, characterized in that: The specific steps of step S1 are as follows: S11: Assume that the length of the workbench is l 台 , height h 台 The upper end face is subjected to the uniform load q, and the workbench support is located at the left and right ends, and the support reaction force is ql 台 / 2, and the stress on the workbench is σ x , σ y , τ xy The stress should satisfy the following boundary conditions: S12: Since there is no axial force and bending moment on the top of the workbench, The following boundary conditions are also satisfied: S13: Assume that the stress function of the workbench is but With σ x , σ y , τ xy The relationship is: S14: Since the workbench is subjected to uniform load, for different x, σ y The distribution is the same, that is, σ y Only related to y, at the same time, the x-axis is symmetrical about the center line of the workbench, σ x Symmetric about the y-axis, so in σ x The coefficients of the odd-order terms of x in the expression should be zero, and the stress function is Expressed as: At the same time, based on the relevant theory of elastic mechanics, the stress function satisfies the biharmonic equation, and for All are established, that is: S15: Substitute the boundary conditions of step S1 and step S2 into the solution to obtain: At the same time, under the theory of material mechanics and elastic mechanics, the deformation of the model also satisfies the following geometric equations and physical equations: Among them, E 台 is the elastic modulus of the workbench, μ 台 is the Poisson's ratio of the workbench, and the moment of inertia I is introduced 台 , we can finally obtain the theoretical model. That is, the deflection curve expression at the end surface of the workbench is: Among them, E 台 is the elastic modulus, the moment of inertia I 台 is the moment of inertia, and the length of the workbench is l 台 , height h 台 .

4. The method for calculating the mechanical compensation value according to claim 1, characterized in that: The specific steps of step S2 are as follows: S21: Simplify the workbench into a simply supported beam model. The fulcrums are fixed hinge supports at both ends of the beam, namely point A and point B. The distance between point A and point B is l 台 , the support reaction forces at point A and point B are F A and F B ; S22: During bending, there is a section of the beam with a uniformly distributed load q. Let the length of this section of the beam be b. There is an equivalent effective point on the beam. This point is in the middle of the beam. The distances between this point and point A and point B are l1 and l2, respectively. Let the distance from the equivalent point to the left support be The distance from the right fulcrum is S23: Perform stress analysis on the beam and obtain F A and F B About q, b, l 台 and the relation of l1 or l2; S24: Substitute the boundary conditions of the shear equation for a beam of length b into the equation. The boundary conditions are: Substituting y=0, we can solve the coordinates of point P as follows: is the symmetry axis of the quartic curve, and the deflection curve expression at the end surface of the workbench in step S15 is further optimized as follows: In the formula, E 台 is the elastic modulus, the moment of inertia I 台 is the moment of inertia, and the length of the workbench is l 台 , height h 台 ; The theoretical model of S1 and the simply supported beam model of this step are unified in the x-axis coordinate system to obtain According to this equation, the position where the deflection is maximum in the actual bending of the workbench can be obtained.

5. The method for calculating the mechanical compensation value according to claim 1, characterized in that: The specific steps of step S3 are as follows: S31: The abstracted deflection curve expression is: Where v is the superposition of the deformation of the machine tool slide and the worktable, that is, the actual compensation value, E 台 is the elastic modulus of the workbench, the moment of inertia I 台 is the inertia moment of the workbench, q 台 is the load applied to the workbench, the length of the workbench is l 台 , the height of the workbench is h 台 , E 滑 is the elastic modulus of the slider, the moment of inertia I 滑 is the moment of inertia of the slider, q 滑 is the load applied to the slider, the slider length is l 滑 , the height of the slider is h 滑 ; S32: When bending, v, E 台 ,I 台 ,q 台 , l 台 、h 台 , P′ 台 、E 滑 ,I 滑 , P′ 滑 ,q 滑 , l 滑 and h 滑 All are known, and the coefficients K1-K in the abstracted deflection curve expression are obtained according to the genetic algorithm. 13 .

6. The method for calculating the mechanical compensation value according to claim 1, characterized in that: In step S3, the swarm intelligence algorithm uses a genetic algorithm. The specific steps of using the genetic algorithm to search and optimize the coefficients in the abstracted deflection curve expression are as follows: S32 1: Initialize the basic parameters related to the genetic algorithm, including population size, number of iterations, crossover rate, and mutation rate; S322: Design the chromosome encoding, the encoding method is: according to the coefficients K1-K in the step deflection curve expression 13 Generate a theoretical chromosome vector, and generate an initial chromosome population by making random differences in the values ​​of each element in the vector; S323: Generate an initial population based on the theoretical model optimized in step S2, calculate the fitness value of the population, sort the fitness of individual chromosomes, and select better chromosomes to enter the next generation; S324: performing adaptive crossover operation and adaptive mutation operation, decoding the chromosome, and calculating and evaluating the fitness of the new population to generate a new population; S325: If the algorithm reaches the set maximum number of iterations or the optimization effect reaches the termination condition, the program ends; otherwise, S323-S325 are repeated.

7. The method for calculating the mechanical compensation value according to claim 2, characterized in that: The method for obtaining the actual compensation value and the theoretical compensation value in practical application is as follows: S41: Obtain theoretical compensation values ​​for the first time to generate an initial population: Substitute machine tool dimensions l1, l2, h1, h2, machine tool performance parameters I1, I2, E1, E2, and uniformly distributed loads q1, q2 into the expression after the deflection curve in step S2 is further optimized, and superimpose the deformation of the machine tool slide and the worktable; S42: when the deflection curve expression needs to be optimized again, a new calculated value is generated according to step S3; S43: After the machine tool is debugged, during the actual bending process, if the machine tool is subjected to large force and does not compensate for the machine tool deflection or only relies on theoretical compensation values ​​for compensation, there will be compensation errors. While ensuring that the sheet material, size, bending angle and bending position remain unchanged, gradually adjust the mechanical compensation value until the angles at both ends and the middle of the sheet are within the tolerance range required by the process. At this time, the adjusted specific compensation value is the actual compensation value.

8. The method for calculating the mechanical compensation value according to claim 6, characterized in that: In step S324, the specific steps of performing the adaptive crossover operation and the adaptive mutation operation are as follows: Step S3241: Adaptively adjust the mutation rate according to the formula. When the individual fitness is higher than the average fitness of the population, the mutation rate will be adaptively adjusted. When the individual fitness is lower than the average fitness of the population, a fixed value with the maximum mutation probability will be assigned. The specific formula is as follows: Where p m is the mutation rate, p m min is the minimum mutation rate, p m max is the maximum mutation rate, F max is the maximum fitness of the population, F b is the individual fitness of the mutated chromosome, F v is the average fitness of the population; Step S3242: Adaptively adjust the crossover rate according to the formula. According to the concentration and dispersion degree of fitness in the population evolution process, the crossover rate of the genetic algorithm is nonlinearly adaptively adjusted, and a two-point crossover method is used to generate new individuals. When the individual fitness is higher than the average fitness of the population, the crossover rate will be adaptively adjusted. When the individual fitness is lower than the average fitness of the population, a maximum fixed value will be given to the crossover rate. The specific formula is as follows: Where p c is the crossover rate, p c mm is the minimum crossover rate, p c max is the maximum crossover rate, F a It is the chromosome with greater fitness during the crossover operation.

9. A bending machine numerical control system, characterized in that: It comprises a numerical simulation module, which executes the mechanical compensation value calculation method described in any one of claims 1-8.

10. The bending machine numerical control system according to claim 9, characterized in that: The module also includes a learning optimization module, which is used to save the actual compensation value and the theoretical compensation value calculated by the numerical simulation module. The learning optimization module learns and optimizes according to the error between the actual compensation value and the theoretical compensation value, and updates K1-K in step S3. 13 Coefficient, perform the next bending operation after updating the parameters. If the compensation is still not up to standard after the theoretical compensation value is used, manually adjust the mechanical compensation value until the bending angle meets the standard. The manually adjusted compensation value is the actual compensation value.

Images