A method for adaptively controlling the roller state of a roll bending machine based on brainstorming mechanism

Through the adaptive control method of roller states of rolling bending machines based on brainstorming mechanism, PID parameters are optimized, and the problem of roller states in the prior art is solved, and the accuracy and efficiency of the rolling bending machine processing profiles are improved.

CN118194478BActive Publication Date: 2024-11-12XIAN UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410610135.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-16
Publication Date
2024-11-12
Estimated Expiration
2044-05-16

AI Technical Summary

Technical Problem

In the existing rolling bending process, it is difficult to achieve ideal control effects when the engineering environment changes, resulting in the failure of the profile processing accuracy.

Method used

The adaptive control method of roller states of rolling bending machines based on brainstorming mechanism is adopted. By analyzing the structure and working principle of the rolling bending machine, the PID closed-loop model and brainstorming mechanism parameters are determined, and the PID parameters are optimized to achieve adaptive control of roller states.

Benefits of technology

It significantly improves the accuracy and production efficiency of the rolling bending machine, avoids the accuracy problems caused by unstable roller states, and improves product quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118194478B_ABST
    Figure CN118194478B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for adaptively controlling the roller state of a rolling bending machine based on a brainstorming mechanism, which belongs to the technical field of rolling bending processing, and includes the following steps: S1, analyzing the structure and working principle of the rolling bending machine, and clarifying the functions and coordination relationships between the various components of the rolling bending machine; S2, determining the roller state model of the rolling bending machine based on the PID closed loop, and the relevant parameters of the brainstorming mechanism; S3, generating N groups of PID initial parameters for the roller state of the rolling bending machine; S4, evaluating the PID parameters of the roller of the rolling bending machine and updating them in combination with the brainstorming mechanism; S5, judging the termination evolution conditions, and outputting the optimal roller state adaptive control scheme of the rolling bending machine. The present invention adopts the above-mentioned method for adaptively controlling the roller state of the rolling bending machine based on the brainstorming mechanism, which solves the problem that the roller state is difficult to control in the existing rolling bending technology, thereby improving the accuracy of the profile processed by the rolling bending machine.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of roll bending, and in particular to a method for adaptively controlling the roller state of a roll bending machine based on a brainstorming mechanism. Background Art

[0002] Roll bending is a process in which the profile to be processed is gradually bent under the action of the bending moment applied by the rollers of the roll bending machine. It is an important part of the manufacturing industry. The profile rolling forming process has been widely used in many fields such as aviation manufacturing, automobile manufacturing, high-speed rail manufacturing, and the electronics industry. Therefore, the research and development of the adaptive control method of the rollers during the rolling process has significant economic benefits.

[0003] In the current roll bending process, the state control method of the roller wheel mostly adopts the PID control method. However, the parameters of the current PID method are mostly obtained by the experience of engineers, and it is difficult to achieve the ideal control effect when the engineering environment changes. In order to solve the above technical problems, the present invention proposes a roll bending machine roller state adaptive control method based on the brainstorming mechanism. Summary of the invention

[0004] The purpose of the present invention is to provide a method for adaptively controlling the roller state of a roll bending machine based on a brainstorming mechanism, which solves the problem that the roller state is difficult to control in the existing roll bending technology, thereby improving the accuracy of profile processing by the roll bending machine.

[0005] To achieve the above object, the present invention provides a method for adaptively controlling the roller state of a roll bending machine based on a brainstorming mechanism, comprising the following steps:

[0006] S1. Analyze the structure and working principle of the roll bending machine, and clarify the functions and coordination relationships between the various components of the roll bending machine;

[0007] S2, determine the roller state model of the roll bending machine based on the PID closed loop, and the relevant parameters of the brainstorming mechanism;

[0008] S3. Generate N groups of PID initial parameters according to the roller status of the roll bending machine;

[0009] S4, evaluate the PID parameters of the roll bending machine roller and update them in combination with the brainstorming mechanism;

[0010] S5. Determine the termination evolution conditions and output the optimal roll bending machine roller state adaptive control scheme.

[0011] Preferably, the specific steps of step S2 are as follows:

[0012] S201, determine the closed-loop PID control model of the roller state of the roll bending machine:

[0013] Select the displacement variable of the roller is the control quantity of the PID controller, which is controlled by the feedback value With displacement variables The difference , so that the control variable is infinitely close to the set target variable. The specific expression is as follows:

[0014] ;

[0015] ;

[0016] in, Represents the output value of the PID controller; Indicates any moment; Represents the proportional coefficient of the PID controller; Indicates the integral coefficient of the PID controller; Represents the differential coefficient of the PID controller;

[0017] S202, determine the relevant parameters of the brainstorming algorithm, including the size of the generated population , maximum number of iterations , different selection probability parameters , , , , the initial value of the number of iterations .

[0018] Preferably, in step S3, the N groups of initial PID parameter sets are , the determination of these N groups of PID parameters is random, among which , Represents the proportional coefficient of the i-th individual in PID control; It represents the integral coefficient of the ith individual in PID control; Represents the differential coefficient of the i-th individual in PID control.

[0019] Preferably, the specific steps of step S4 are as follows:

[0020] S401, evaluating the PID parameters for controlling the roller state:

[0021] Under the premise of considering the response speed and steady-state error, the objective function of the piecewise form of suppressing overshoot is adopted:

[0022] ;

[0023] in, represents the objective function value; Indicates the input quantity; represents the rise time;

[0024] S402: Update PID parameters in combination with a brainstorming mechanism.

[0025] Preferably, the specific updating steps in step S402 are as follows:

[0026] S4021, performing clustering operation on individual PID parameters;

[0027] S4022, performing a selection operation;

[0028] S4023, implement crossover and mutation operations;

[0029] S4024. Obtain a new individual.

[0030] Preferably, step S5 specifically includes: setting the maximum number of iterations , record the current number of iterations After each iteration, check whether the number of iterations has reached the maximum number of iterations. When the condition is met, the optimal PID parameters are given; if not, steps S4-S5 are repeated.

[0031] Therefore, the present invention adopts the above-mentioned method for adaptively controlling the roller state of a rolling bending machine based on the brainstorming mechanism, and uses the brainstorming mechanism to optimize the PID control parameters of the roller, so that the roller can be adaptively controlled according to the actual processing conditions, avoiding the problem of substandard production profile accuracy due to unstable roller state, and significantly improving production efficiency and product quality.

[0032] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a flow chart of an embodiment of a method for adaptively controlling the roller state of a roll bending machine based on a brainstorming mechanism of the present invention;

[0034] Figure 2 It is a three-roller state diagram of a roll bending machine according to an embodiment of a method for adaptively controlling the roll state of a roll bending machine based on a brainstorming mechanism of the present invention;

[0035] Figure 3 It is a PID parameter setting flow chart based on the brainstorming mechanism of an embodiment of a roll bending machine roller state adaptive control method based on the brainstorming mechanism of the present invention;

[0036] Figure 4 The present invention discloses an adaptive control method for roller state of a rolling bending machine based on a brainstorming mechanism, which is a control model diagram for optimizing roller state based on a brainstorming mechanism. DETAILED DESCRIPTION

[0037] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0038] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0039] Embodiment 1

[0040] The present invention provides a method for adaptively controlling the roller state of a roll bending machine based on a brainstorming mechanism, such as Figure 1 As shown, the following steps are included:

[0041] S1. Analyze the structure and working principle of the roll bending machine, including the motion characteristics of the upper roller, left and right lower rollers, servo cylinder and other components, and clarify the functions and coordination relationships between the various components of the roll bending machine;

[0042] The mechanical structure of the three-roller roll bending machine is mainly composed of upper roll, left and right lower rolls, servo cylinder, lower center roll, left and right side rolls and frame. The three-roller state diagram of the roll bending machine is as follows: Figure 2 As shown in the figure, when the oil cylinder starts to work, the servo motor drives the hydraulic pump to suck and discharge oil, so that the left and right lower rollers of the roll bending machine move in an arc along the middle pin by the servo oil cylinder. The displacement of the roller is monitored by the displacement sensor, and the displacement signal is fed back to the controller of the electrical control module. The controller calculates the control error by comparing the difference between the feedback signal and the expected displacement value, and sends a corresponding signal to control the displacement of the roller accordingly.

[0043] From this analysis, it can be seen that the final arc accuracy of the profile processed by the roll bending machine is largely determined by the position accuracy of the left and right lower rollers. The left and right lower rollers of the roll bending machine are driven by servo cylinders, the position feedback uses magnetic displacement sensors, and the control uses high-frequency response servo valves. On this basis, the brainstorming mechanism is combined to control the processing state of the roll bending machine rollers to ensure the accuracy and consistency of profile processing.

[0044] S2. Determine the roller state model of the roll bending machine based on the PID closed loop and the relevant parameters of the brainstorming mechanism. The specific steps are as follows:

[0045] S201, determine the closed-loop PID control model of the roller state of the roll bending machine:

[0046] Select the displacement variable of the roller is the control quantity of the PID controller, where PID control refers to the comprehensive control of the proportional, integral and derivative of the deviation. With displacement variables The difference , so that the control variable is infinitely close to the set target variable. The specific expression is as follows:

[0047] ;

[0048] ;

[0049] in, Represents the output value of the PID controller; Indicates any moment; Represents the proportional coefficient of the PID controller; Indicates the integral coefficient of the PID controller; Indicates the differential coefficient of the PID controller; when the proportional factor in the PID controller As the integral coefficient increases, the controller becomes more sensitive to the error. The accumulation of cumulative error becomes more sensitive if the differential coefficient is increased. , it will change the deviation more sensitively. On this basis, by adjusting the above three parameters, the predetermined control target can be achieved. The control model diagram of the roller state optimization based on the brainstorming mechanism is as follows Figure 4 shown.

[0050] S202, determine the relevant parameters of the brainstorming algorithm, including the size of the generated population (integer value), maximum number of iterations (integer value), different selection probability parameters , , , (between 0 and 1), the initial value of the number of iterations Among them, the size of the population determines the breadth of the search space, and the maximum number of iterations affects the convergence speed and search effect of the algorithm. The selection probability parameter determines the probability of an individual being selected in the evolutionary algorithm, and the initial value of the number of iterations affects the search starting point at the beginning of the algorithm. Therefore, the reasonable setting of these parameters is crucial for the PID controller to accurately control the roller state.

[0051] S3, according to the roller state of the roll bending machine, generate N groups of PID initial parameters; in this method, The same parameters are optimized, so the encoding length is 3, and N groups of initial PID parameter sets are determined , the determination of these N groups of PID parameters is random, and each individual is a potential solution. , Represents the proportional coefficient of the i-th individual in PID control; It represents the integral coefficient of the ith individual in PID control; Represents the differential coefficient of the i-th individual in PID control.

[0052] S4. Evaluate the PID parameters of the roll bending machine roller and update them in combination with the brainstorming mechanism. The specific steps are as follows:

[0053] S401, evaluating the PID parameters for controlling the roller state:

[0054] When evaluating a control system, the response speed and stability error are usually considered, but the requirements for overshoot are not fully considered. Under the premise of considering the response speed and steady-state error, a segmented objective function that suppresses overshoot is adopted. By introducing the control of overshoot, the performance of the control system can be evaluated more comprehensively, so that the system not only has a fast response and a small steady-state error, but also can effectively suppress overshoot and improve the stability and control accuracy of the system. This method of comprehensively considering different factors helps to achieve a more comprehensive and accurate evaluation and optimization of the control system. The following is the objective function:

[0055] ;

[0056] in, represents the objective function value; Indicates the input quantity; Represents the rise time; the objective function value reflects the individual fitness level. The smaller the objective function value, the greater the individual fitness value, that is, the better the individual.

[0057] S402. Update the PID parameters in combination with the brainstorming mechanism. The brainstorming optimization algorithm updates the population through discussions within and between groups, clusters individuals into different categories, and randomly selects individuals from each category. According to the preset probability, it is determined whether the individual is updated through discussion within the group or between groups. Each new individual adopts a different generation mechanism according to the different selection probabilities. In combination with the actual conditions of the present invention, the following operations are performed on each group of parameters in the PID parameter set. The PID parameter tuning flow chart based on the brainstorming mechanism is shown as follows: Figure 3 As shown,

[0058] S4021. Clustering operation is performed on PID parameter individuals. Clustering operation is a significant feature of the brainstorming mechanism. The steps of clustering operation are: select m individuals from the PID population as cluster centers, and then evaluate the status of other individuals in the population according to the Euclidean distance calculation method. According to the evaluation results, the individuals with higher similarity are classified into one category. After multiple clustering, the individuals in the population can be divided into m independent categories, which is convenient for the entire population to improve the search efficiency in evolution.

[0059] S4022. Perform a selection operation. The selection operation is a key step in the PID parameter optimization process. In the brainstorming algorithm, there is a specific mechanism for selecting the PID parameters to be updated. Among them, one mechanism is to directly generate new PID parameter individuals within the same class. When updating individuals within the same class, new individuals can be generated by perturbing existing individuals or combining them with another individual in the class; another mechanism is to generate new PID parameter individuals between different classes. When updating individuals between different classes, individuals from two classes can be randomly selected and fused in some way to generate new individuals.

[0060] S4023, implement crossover and mutation operations; in the population update process, single-point crossover is a common individual update method, which is used to exchange part of two PID parameter encoding individuals. In the population, the encoding length of each individual is 3, and each individual consists of The single-point crossover operation randomly selects a crossover point on the gene sequences of two individuals, and then exchanges the two individuals at the selected crossover point. Through the single-point crossover operation, new individuals can be generated, thereby increasing the diversity of the population.

[0061] S4024, obtaining a new individual. After selecting the individual to be updated, a new PID parameter equation is generated:

[0062] ;

[0063] in, represents the individual before updating; represents the updated individual; random(t) represents a random function, which generally uses functions such as Cauchy, Gaussian and Levy flights. is the weight coefficient. This mutation method of generating random values ​​through random functions is similar to the mutation operation in evolutionary algorithms.

[0064] ;

[0065] in, Represents logarithm The transfer function of Indicates the maximum number of iterations; Indicates the current iteration number; Indicates change The slope of the function; Represents a random value between 0 and 1. When a new individual is generated by inter-group discussion in the algorithm, two old individuals are recombined with a certain weight, and on this basis, random value individual perturbations are performed again to generate new individuals.

[0066] ;

[0067] in, Represents old individual and old individuals The weight and; and is the individual weight coefficient;

[0068] S5. Determine the termination conditions of the evolution and output the optimal roll bending machine roller state adaptive control scheme. Set the maximum number of iterations , record the current number of iterations After each iteration, check whether the number of iterations has reached the maximum number of iterations. When the condition is met, the optimal PID parameters are given; if not, steps S4-S5 are repeated.

[0069] Therefore, the present invention adopts the above-mentioned method for adaptively controlling the roller state of a rolling bending machine based on the brainstorming mechanism, which solves the problem that the roller state is difficult to control in the existing rolling bending technology, thereby improving the accuracy of processing profiles by the rolling bending machine.

[0070] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for adaptively controlling the roller state of a roll bending machine based on a brainstorming mechanism, characterized in that: The following steps are involved: S1. Analyze the structure and working principle of the roll bending machine. The mechanical structure of the three-roller roll bending machine is mainly composed of the upper roll, left and right lower rolls, servo cylinder, lower center roll, left and right side rolls and frame components. When the cylinder starts to work, the servo motor drives the hydraulic pump to absorb and discharge oil, so that the left and right lower rolls of the roll bending machine move in an arc along the middle pin by the servo cylinder. Therefore, the role and coordination relationship between the various components of the roll bending machine are clarified: The final arc accuracy of the profile processed by the roll bending machine is determined by the position accuracy of the left and right lower rollers; the left and right lower rollers of the roll bending machine are driven by servo cylinders, the position feedback uses magnetic displacement sensors, and the control uses high-frequency response servo valves. On this basis, the brainstorming mechanism is combined to control the processing state of the roll bending machine rollers; S2. Determine the roller state model of the roll bending machine based on the PID closed loop and the relevant parameters of the brainstorming mechanism. The specific steps are as follows: S201, determine the closed-loop PID control model of the roller state of the roll bending machine: Select the displacement variable y of the roller d (t) is the control variable of the PID controller, which is controlled by the feedback value y(t) and the displacement variable y d (t) makes the control variable infinitely close to the set target variable. The specific expression is as follows: e(t)=y d (t)-y(t); Where u(t) represents the output value of the PID controller; t represents the time; k p Indicates the proportional coefficient of the PID controller; k i Indicates the integral coefficient of the PID controller; k d Represents the differential coefficient of the PID controller; S202, determine the relevant parameters of the brainstorming algorithm, including the size M of the generated population, the maximum number of iterations T max , different selection probability parameters P1, P2, P3, P4, the initial value of the number of iterations T = 0; S3. Generate N groups of PID initial parameters according to the roller state of the roll bending machine; wherein the N groups of initial PID parameter set X=[X1,X2,...,X i ,...,X N ], the determination of these N groups of PID parameters is random, among which X i =[X iP ,X iI ,X iD ], X iP represents the proportional coefficient of the i-th individual in PID control; X iI represents the integral coefficient of the i-th individual in PID control; X iD Represents the differential coefficient of the i-th individual in PID control; S4, evaluating the PID parameters of the roller of the roll bending machine and updating them in combination with the brainstorming mechanism; the specific steps of step S4 are as follows: S401. Under the premise of considering the response speed and steady-state error, the PID parameters for controlling the roller state are evaluated, and a segmented objective function for suppressing overshoot is adopted: Where J represents the objective function value; u(t) represents the input; t u represents the rise time; e(t) represents the feedback value y(t) and the displacement variable y d The difference of (t); S402, updating PID parameters in combination with a brainstorming mechanism; The brainstorming optimization algorithm updates the population through discussions within and between groups, clusters individuals into different categories, and randomly selects individuals from each category; according to the preset probability, it determines whether the individual is updated through discussion within or between groups, and each new individual adopts a different generation mechanism according to the different selection probabilities; for each group of parameters in the PID parameter set, the following operations are performed: The specific updating steps in step S402 are as follows: S4021, clustering operation is performed on PID parameter individuals; the steps of clustering operation are: m individuals are selected from the PID population as cluster centers, and then the status of other individuals in the population are evaluated according to the Euclidean distance calculation method, and individuals with higher similarity are classified into one category according to the evaluation results; after multiple clustering, the individuals in the population are divided into m independent categories; S4022, performing a selection operation; the selection operation is a key step in the PID parameter optimization process: one mechanism is to directly generate a new PID parameter individual within the same class; when updating individuals within the same class, a new individual is generated by perturbing an existing individual or combining it with another individual in the class; another mechanism is to generate a new PID parameter individual between different classes; and when updating individuals between different classes, individuals in two classes are randomly selected and merged to generate a new individual; S4023, implement crossover and mutation operations; Single-point crossover is used to exchange part of two PID parameter encoding individuals; in the population, the encoding length of each individual is 3, and each individual consists of k p , k i , k d Composition; The single-point crossover operation randomly selects a crossover point on the gene sequences of two individuals, and then exchanges the two individuals at the selected crossover point; In order to obtain a new individual, after selecting the individual to be updated, a new PID parameter equation is generated: x new =x old +ξ*random(t); Among them, x new represents the individual before updating; x old represents the updated individual; random(t) represents a random function, using functions such as Cauchy, Gaussian and Levyflights, and ξ is a weight coefficient; ξ=logsig((T max / 2-T) / K)*rand(); Wherein, logsig() represents the transfer function of logarithmic sigmoid; T max Indicates the maximum number of iterations; T indicates the current number of iterations; K indicates changing the slope of the logsig() function; rand() indicates a random value from 0 to 1; When a new individual is generated by inter-group discussion in the algorithm, two old individuals are recombined with a certain weight, and on this basis, random value individual perturbations are performed again to generate new individuals, x old =w1*x old1 +w2*x old2 ; Among them, x old Represents the old individual x old1 and old individual x old2 The weighted sum of; w1 and w2 are individual weight coefficients; S4024, obtain a new individual; S5. Determine the termination evolution condition and output the optimal roll bending machine roller state adaptive control scheme, specifically: set the maximum number of iterations T max , record the current number of iterations T c After each iteration, check whether the number of iterations has reached the maximum number of iterations T max When the condition is met, the optimal PID parameters are given; if not, steps S4-S5 are repeated.