Flexible workpiece deformation control system

Abstract

The invention provides a flexible workpiece deformation control system. The system is composed of a state sensing module, a deformation modeling and prediction module, a control strategy generation module and a real-time control execution module. Wherein the state sensing module collects vision, force and position data of a flexible workpiece through a multi-mode sensor, carries out data fusion and outputs a synchronous state data set; the deformation modeling and prediction module predicts a future deformation state sequence of the workpiece in real time based on a material parameter identification and deformation prediction model; the control strategy generation module solves an optimal control sequence through a target function and a constraint condition; and the real-time control execution module outputs an actuator control signal according to the optimal control sequence, so that real-time deformation control of the flexible workpiece is realized. According to the method, deformation of the flexible workpiece can be accurately controlled through data flow interaction cooperation, the machining precision requirement is met, energy consumption is minimized, and high precision and energy consumption optimization of deformation control of the flexible workpiece are achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial automation control, and in particular relates to a flexible workpiece deformation control system. Background Art

[0002] Deformation control of flexible workpieces during the manufacturing process is an important issue in industrial production, especially in precision machining and complex processes. Deformation often leads to workpiece dimensional errors, reduced machining accuracy, and even affects product quality and production efficiency. Deformation monitoring and control can be performed through mechanical clamping, temperature control, laser measurement, force sensors, etc. However, these traditional methods usually face challenges such as sensor accuracy limitations, feedback control lag, and material nonlinear response, which makes it difficult to achieve ideal deformation control effects under large deformations or complex working conditions. In recent years, with the development of intelligent manufacturing technology, based on emerging technologies such as multi-sensor fusion, machine learning algorithms, and adaptive control, the deformation control accuracy and real-time performance of flexible workpieces have been significantly improved. However, the existing flexible workpiece deformation control schemes have the following deficiencies:

[0003] (1) Usually relying on a single sensor for state perception, it is difficult to fully and accurately obtain the real-time deformation state of the flexible workpiece, which affects the deformation control effect;

[0004] (2) Simplified linear elastic models are often used, which fail to fully consider the plastic deformation characteristics of the material, resulting in reduced prediction accuracy under large deformation conditions;

[0005] (3) They often focus on deformation accuracy indicators, ignoring important factors such as energy consumption and workpiece deformation constraints. They lack the ability to identify material parameters in real time, making it difficult to ensure model accuracy.

[0006] (4) Most of them adopt a single feedback control method, which has limited adaptability to system model errors and external interference, making it difficult to achieve precise control of the deformation of flexible workpieces.

[0007] Therefore, we need to develop a flexible workpiece deformation control system that can use multimodal sensors, material identification, optimization control and dual control structure to improve deformation prediction and control accuracy. Summary of the Invention

[0008] The purpose of the present invention is to provide a flexible workpiece deformation control system to solve the problems mentioned in the above background technology of the existing flexible workpiece deformation control scheme, such as single sensor, simplified model, neglect of energy consumption, and insufficient feedback control.

[0009] To achieve the above objectives, the present invention provides a flexible workpiece deformation control system, which is specifically as follows:

[0010] State perception module: This module collects the visual data, force data, and position data of the flexible workpiece in real time through multimodal sensors, performs data preprocessing and data fusion, and outputs a synchronized state data set.

[0011] Deformation modeling and prediction module: Based on the state data set obtained by the state perception module, the module obtains the real-time material parameters of the flexible workpiece, including elastic parameter estimates and plastic parameter estimates, through real-time identification. Based on the material parameters, the module predicts the deformation state sequence of the flexible workpiece in the future prediction time domain in real time through the deformation prediction model;

[0012] Control strategy generation module: Based on the deformation state sequence, construct an objective function and define constraints to solve the optimal control sequence, wherein the constraints include deformation constraints, yield constraints, and output constraints;

[0013] The goal of deformation control is to minimize the control energy consumption while meeting the deformation accuracy requirements. The objective function is constructed as follows:

[0014]

[0015] in, represents the trajectory tracking error, represents the deformation constraint term, Represents the energy constraint term, which is used to control energy consumption. 、 、 is the weight coefficient;

[0016] Real-time control execution module: Based on the optimal control sequence obtained by the control strategy generation module, the module outputs an actuator control signal through feedforward control and feedback correction to achieve real-time deformation control of the flexible workpiece.

[0017] Based on the above solution, the visual data includes: the three-dimensional coordinates of each key point on the surface of the flexible workpiece And the deformation of the corresponding position ;

[0018] The force data includes: three-dimensional force vectors at key positions on the surface of the flexible workpiece , torque vector And the stress tensor at each key position ;

[0019] The position data includes: the position vector of the origin of the flexible workpiece coordinate system in the base coordinate system , posture rotation matrix W and the spatial coordinates of the key points on the surface of the flexible workpiece.

[0020] Based on the above solution, the identification of the elastic parameter estimate includes:

[0021] If the material of the flexible workpiece is an elastic material, the Young's modulus is used. and Poisson's ratio Describe its mechanical behavior, specifically:

[0022] A certain number of displacement-force data pairs of key points are selected from the state data set as the input data set for parameter identification. Assuming that the flexible workpiece obeys the generalized Hooke's law, the following linear relationship exists: , where σ is the stress tensor, ε is the strain tensor, and E is the Young's modulus to be identified;

[0023] Based on the position data, the strain tensor at each key point is calculated using the finite element method or other numerical methods. , solve the following optimization problem by the least squares method: , and obtain the Young's modulus The estimated value of Indicates the The displacement-force data pairs of key points, is the strain tensor at the i-th key point;

[0024] The Poisson's ratio is estimated by the least squares method based on the transverse and longitudinal strain data of the flexible workpiece. , and finally obtain the elastic parameter estimate .

[0025] Based on the above scheme, the identification of the estimated value of the plasticity parameter includes:

[0026] If the material of the flexible workpiece is plastic material, in addition to the elastic parameter estimate In addition, the yield stress , hardening modulus , hardening index The plasticity parameters are:

[0027] Select a certain number of displacement-force data pairs of key points from the state data set And the corresponding strain data As the input data set for parameter identification, a certain number of key points are selected to cover the elastic and plastic deformation stages of the flexible workpiece, and the initial estimates of the hardening parameters H and n are determined based on material experience or pre-experimental data. and ;

[0028] Assuming that the flexible workpiece obeys the isotropic hardening criterion, the yield function is: ,in, Equivalent stress, is the equivalent plastic strain, is the initial yield stress, and is the hardening parameter to be identified, i.e., the initial estimate;

[0029] The initial yield stress is estimated by the least squares method based on the elastic-plastic transition point data of the flexible workpiece. , using the Levenberg-Marquardt algorithm, solve the following optimization problem: , and obtain the hardening parameters and The estimated value of According to the initial yield stress and initial estimates and , the stress value calculated by the yield function;

[0030] Repeat the calculation until the hardening parameter and The estimated value of converges, and finally the estimated value of the plastic parameter of the flexible workpiece is obtained .

[0031] Based on the above solution, the deformation prediction model is established based on the material parameters to predict the deformation of the flexible workpiece. The construction process of the deformation prediction model is specifically as follows:

[0032] According to the length of the future prediction time domain, the deformation prediction of the flexible workpiece is divided into short-term prediction and medium-term prediction;

[0033] The short-term prediction is used to predict the deformation state of the flexible workpiece in the next 0.1 seconds, and is performed using a linear elastic finite element model that uses elastic parameter estimates updated in real time. , accurately calculate the response of the workpiece in the elastic deformation stage, interpolate the predicted node state quantity to the surface of the flexible workpiece, and obtain the future Deformation state sequence of control cycles .

[0034] Based on the above scheme, the medium-term prediction is used to predict the deformation state of the flexible workpiece in the next 1 second, and the prediction is made using an elastic-plastic finite element model, which introduces the estimated value of the plastic parameter obtained by real-time identification. , in order to reflect the nonlinear characteristics of the workpiece when plastic deformation occurs, the predicted node state quantity is interpolated to the surface of the flexible workpiece to obtain the future Deformation state sequence of control cycles ;

[0035] Will and Splicing is performed to obtain the complete deformation state sequence of the next m control cycles: ,in Indicates that at the current moment Afterwards The predicted deformation state of each control cycle.

[0036] Based on the above scheme, the trajectory tracking error It is used to measure the deviation between the actual deformation state and the expected deformation state of the flexible workpiece, and is defined as: ,in for The actual deformation state at the moment, for The desired deformation state at the moment, is the tracking error weight matrix;

[0037] The desired deformation state is an ideal deformation state sequence of the flexible workpiece generated according to the processing requirements. ,in is the number of control cycles in the future prediction time domain; the actual deformation state is obtained from the deformation state sequence Extract the deformation state sequence at the corresponding moment ;

[0038] Calculate the deformation deviation vector at each moment: ;

[0039] The tracking error weight matrix Perform weighted accumulation on the deformation deviation vector to obtain the trajectory tracking error : .

[0040] Based on the above scheme, the energy constraint term is used to measure the energy consumption during the deformation control process to achieve energy-saving optimization, and is defined as: ,in is the optimal control input, is the control energy weight matrix;

[0041] Extract the optimal control input at the current moment from the optimal control sequence of the previous control cycle ;

[0042] According to the control energy weight matrix , calculate the optimal control input The energy consumption is calculated as follows: .

[0043] Based on the aforementioned solution, the feedforward control is to calculate the optimal control input based on the deformation prediction model and the current physical state of the flexible workpiece to compensate for the known dynamic characteristics of the system;

[0044] From the optimal control sequence, extract the optimal control input at the current moment ;

[0045] According to the deformation prediction model and the material parameters, the feedforward control amount in the current state is calculated: ,in, For the flexible workpiece The state variables at time , for The optimal control input at time t, is the material parameter vector, is the feedforward controller function.

[0046] Based on the above solution, the feedback correction specifically includes:

[0047] Measure the actual state variables at the current moment through sensors ;

[0048] The actual state variable and the predicted state variables Compare and calculate the deviation:

[0049]

[0050] According to the deviation size and PID control law, calculate the feedback control amount:

[0051]

[0052] in, 、 、 They are proportional, integral and differential coefficients respectively, which can be optimized through experimental tuning or adaptive algorithm.

[0053] The present invention has the following advantages and effects compared to the prior art:

[0054] (1) A state perception module is constructed using sensors in three modalities: vision, force, and position. Multi-source data is collected and integrated in real time. Through data preprocessing and multi-modal fusion, a comprehensive perception of the deformation state of the flexible workpiece is achieved.

[0055] (2) A deformation modeling and prediction method based on material parameter identification is proposed. The elastic and plastic parameters of the material are identified by real-time data, and short-term and medium-term prediction models are established respectively. The elastic and plastic deformation characteristics are comprehensively considered to improve the accuracy of deformation prediction.

[0056] (3) A multi-objective optimization control strategy generation module was constructed. Taking into account trajectory tracking error, deformation constraints, and energy consumption, an optimization model including physical constraints such as material yield limit was established, and the optimal control sequence was solved using a numerical optimization algorithm.

[0057] (4) In terms of real-time control execution, the present invention designs a dual control structure that combines feedforward control and feedback control. The feedforward control compensates for the known dynamic characteristics of the system, and the feedback control suppresses model errors and external interference, thereby improving control performance through controller parameter optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly describes the drawings required for the specific embodiments or the description of the prior art. Similar elements or parts are generally identified by similar reference numerals throughout the drawings. Elements or parts in the drawings are not necessarily drawn to scale.

[0059] Figure 1 It is a structural diagram of a flexible workpiece deformation control system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0060] In order to more clearly illustrate the purpose, technical solutions and advantages of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. The example implementation methods can be implemented in various forms and should not be understood as being limited to the examples described herein. On the contrary, these implementation methods are provided to make the present invention more comprehensive and complete, and to fully convey the concepts of the example implementation methods to those skilled in the art.

[0061] In addition, the described features, structures or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present invention. However, it will be appreciated by those skilled in the art that the technical solutions of the present invention can be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known methods, devices, implementations or operations are not shown or described in detail to avoid blurring various aspects of the present invention.

[0062] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0063] The present invention will be described in detail below with reference to specific embodiments:

[0064] As attached Figure 1 As shown, an embodiment of the present invention provides a flexible workpiece deformation control system S100, which is mainly composed of four functional modules: a state perception module S101, a deformation modeling and prediction module S102, a control strategy generation module S103, and a real-time control execution module S104. The modules interact with each other through data streams and work together to achieve precise control of the deformation of the flexible workpiece. The details are as follows:

[0065] State perception module S101: collects the visual data, force data and position data of the flexible workpiece in real time through multimodal sensors, performs data preprocessing and data fusion, and outputs a synchronized state data set.

[0066] Deformation modeling and prediction module S102: Based on the state data set, through material parameter identification and deformation prediction model, real-time prediction of the deformation state sequence of the flexible workpiece in the future prediction time domain.

[0067] Control strategy generation module S103: Based on the deformation state sequence, solve the optimal control sequence by constructing an objective function and defining constraint conditions;

[0068] Real-time control execution module S104: outputs an actuator control signal according to the optimal control sequence through feedforward control and feedback correction to achieve real-time deformation control of the flexible workpiece.

[0069] Furthermore, the workflow of the entire system is as follows: the state perception module 101 perceives the state of the flexible workpiece in real time through the multimodal sensor, and outputs the synchronized fusion state data set to the deformation modeling and prediction module 102. The deformation modeling and prediction module 102 predicts the future deformation state sequence of the flexible workpiece through the deformation prediction model based on the state data set, and outputs the prediction result to the control strategy generation module 103. The control strategy generation module 103 comprehensively considers the objective function and constraints based on the predicted deformation state sequence, solves the optimal control sequence, and outputs it to the real-time control execution module 104. The real-time control execution module 104 generates an actuator control signal based on the received optimal control sequence, and drives the actuator to perform deformation control on the flexible workpiece. The more complete specific implementation of each module is as follows:

[0070] Preferably, the state perception module S101 uses sensors of three modalities: vision, force, and position, to collect state information of the flexible workpiece in real time, including visual data, force data, and position data;

[0071] Specifically, the visual sensor uses an industrial camera to obtain deformation information of the flexible workpiece surface by taking an image of the flexible workpiece surface; the visual data includes: the three-dimensional coordinates of each key point on the flexible workpiece surface And the deformation of the corresponding position ;

[0072] Force sensors are arranged at key positions of the flexible workpiece to measure the external forces and moments on the flexible workpiece; force data include: three-dimensional force vectors at key positions on the surface of the flexible workpiece; , torque vector And the stress tensor at each key position ;

[0073] The position sensor is used to track the position and posture of the flexible workpiece in space in real time; the position data includes: the position vector of the origin of the flexible workpiece coordinate system in the base coordinate system , posture rotation matrix W and the spatial coordinates of the key points on the surface of the flexible workpiece.

[0074] Preferably, the data preprocessing includes: timestamp alignment, data interpolation, and outlier detection; the data fusion specifically includes: after the data preprocessing step, the data of different sensors are unified to the same time base, the multimodal data after the data preprocessing are integrated to generate the synchronized state data set, which can be expressed as ,in is the spatial position vector of the key point on the surface of the flexible workpiece, are the force and torque vectors acting on the key points, is the stress tensor at each key position on the surface of the flexible workpiece .

[0075] Preferably, the material parameter identification in the deformation modeling and prediction module 102 is to accurately predict the deformation behavior of the flexible workpiece and identify the constitutive parameters of the flexible workpiece material, including:

[0076] If the material of the flexible workpiece is an elastic material, the Young's modulus is used. and Poisson's ratio Describe its mechanical behavior. Specifically, select a certain number of displacement-force data pairs from the state data set as the input data set for parameter identification. Assume that the flexible workpiece obeys the generalized Hooke's law, that is, the following linear relationship exists: , where σ is the stress tensor, ε is the strain tensor, and E is the Young's modulus to be identified. Based on the position data, the strain tensor at each key point is calculated using the finite element method or other numerical methods. , using the least squares method, solve the following optimization problem: , and obtain the Young's modulus The estimated value of Indicates the The displacement-force data pairs of key points, is the strain tensor at the i-th key point; when Young's modulus is known In the case of the flexible workpiece, the Poisson's ratio is estimated by the least squares method based on the transverse and longitudinal strain data of the flexible workpiece. , and finally obtain the elastic parameter estimation value of the flexible workpiece .

[0077] If the material of the flexible workpiece is plastic material, in addition to the elastic parameter estimate In addition, the yield stress , hardening modulus , hardening index The plastic parameters are as follows: a certain number of displacement-force data pairs of key points are selected from the state data set. And the corresponding strain data As the input data set for parameter identification, a certain number of key points are selected to cover the elastic and plastic deformation stages of the flexible workpiece, and the initial estimates of the hardening parameters H and n are determined based on material experience or pre-experimental data. and ; Assume that the flexible workpiece obeys the isotropic hardening criterion and the yield function is: ,in, Equivalent stress, is the equivalent plastic strain, is the initial yield stress, and is the hardening parameter to be identified, i.e., the initial estimated value; according to the elastic-plastic transition point data of the flexible workpiece, the initial yield stress is estimated by the least squares method , using the Levenberg-Marquardt algorithm, solve the following optimization problem: , and obtain the hardening parameters and The estimated value of According to the initial yield stress and initial estimates and , the stress value calculated by the yield function; repeat the calculation until the hardening parameter and The estimated value of converges, and finally the estimated value of the plastic parameter of the flexible workpiece is obtained .

[0078] Preferably, the deformation prediction model in the deformation modeling and prediction module 102 is established based on the identified material parameters to predict the deformation of the flexible workpiece, and the material parameters include the elastic parameter estimation value and the estimated values ​​of the plasticity parameters , the construction process of the deformation prediction model is specifically as follows:

[0079] According to the length of the future prediction time domain, the deformation prediction of the flexible workpiece is divided into short-term prediction and medium-term prediction;

[0080] The short-term prediction is used to predict the deformation state of the flexible workpiece in the next 0.1 seconds. Considering that the deformation of the flexible workpiece in this time scale is mainly elastic deformation, a linear elastic finite element model is used for prediction. The linear elastic finite element model uses the elastic parameter estimation value updated in real time. , accurately calculate the response of the workpiece in the elastic deformation stage, interpolate the predicted node state quantity to the surface of the flexible workpiece, and obtain the future Deformation state sequence of control cycles ;

[0081] The medium-term prediction is used to predict the deformation state of the flexible workpiece in the next 1 second. Considering that the flexible workpiece may undergo plastic deformation within this time scale, the elastic-plastic finite element model is used for prediction. The elastic-plastic finite element model introduces the estimated value of the plastic parameter obtained by real-time identification. , in order to reflect the nonlinear characteristics of the workpiece when plastic deformation occurs, the predicted node state quantity is interpolated to the surface of the flexible workpiece to obtain the future Deformation state sequence of control cycles ;

[0082] Will and Splicing is performed to obtain the complete deformation state sequence of the next m control cycles: ,in Indicates that at the current moment Afterwards The predicted deformation state of a control cycle; in this embodiment, m can be selected as 100.

[0083] Preferably, in this embodiment, the goal of deformation control is to minimize control energy consumption while meeting deformation accuracy requirements. The objective function in the control strategy generation module S103 is constructed as follows:

[0084]

[0085] in, represents the trajectory tracking error, represents the deformation constraint term, Represents the energy constraint term, which is used to control energy consumption. 、 、 is the weight coefficient;

[0086] Specifically, the trajectory tracking error It is used to measure the deviation between the actual deformation state and the expected deformation state of the flexible workpiece, and is defined as: ,in for The actual deformation state at the moment, for The desired deformation state at the moment, is the tracking error weight matrix;

[0087] The desired deformation state is an ideal deformation state sequence of the flexible workpiece generated according to the processing requirements. ,in is the number of control cycles in the future prediction time domain; the actual deformation state is obtained from the deformation state sequence Extract the deformation state sequence at the corresponding moment ;

[0088] Calculate the deformation deviation vector at each moment: ;

[0089] The tracking error weight matrix Perform weighted accumulation on the deformation deviation vector to obtain the trajectory tracking error : .

[0090] Furthermore, the deformation constraint term is used to limit the deformation of the flexible workpiece to avoid material failure caused by excessive deformation, and is defined as: ,in For the flexible workpiece The deformation state of the moment, is the maximum allowable deformation;

[0091] Specifically, the setting of the deformation constraint threshold is to set the maximum allowable deformation of the flexible workpiece in each direction according to the material properties and processing requirements of the flexible workpiece. .

[0092] Specifically, the deformation state sequence Each predicted deformation state in Check the deformation state to see if the deformation exceeds the maximum allowable deformation. ;

[0093] For deformation exceeding the maximum allowable The predicted deformation state is calculated, and the sum of the squares of its excess limits is used as the deformation constraint term: .

[0094] Furthermore, the energy constraint term is used to measure the energy consumption during the deformation control process to achieve energy-saving optimization, and is defined as: ,in is the optimal control input, is the control energy weight matrix;

[0095] Extract the optimal control input at the current moment from the optimal control sequence of the previous control cycle ;

[0096] According to the control energy weight matrix , calculate the optimal control input The energy consumption is calculated as follows: .

[0097] The selection of R needs to comprehensively consider the power limit and energy consumption level of the actuator, and can be set as a diagonal matrix inversely proportional to the rated power of the actuator.

[0098] Furthermore, to ensure that the deformation control process satisfies the various physical constraints of the deformation control of the flexible workpiece, the constraints imposed on the objective function include:

[0099] (1) Deformation constraint: It is required that during the deformation control process, the deformation of each position of the flexible workpiece does not exceed the maximum allowable deformation , the expression is:

[0100]

[0101] in, For location in The deformation displacement at the moment, is the maximum allowable deformation, is the spatial domain of the flexible workpiece, is the total deformation control time;

[0102] (2) Yield constraint: It requires that the stress level at each position of the flexible workpiece does not exceed the yield limit of its material during deformation control. The expression is:

[0103]

[0104] in, For location in The stress tensor at time , is the yield stress of the material of the flexible workpiece, is the spatial domain of the flexible workpiece, is the total deformation control time;

[0105] (3) Output constraint: It requires that the output force or displacement of the actuator does not exceed its physical limit during the deformation control process. The expression is:

[0106]

[0107] in, for The control input of the moment actuator, and are the minimum and maximum output capabilities of the actuator, Control time for the total deformation.

[0108] Furthermore, based on the objective function and the constraint conditions, the optimal deformation control problem of the flexible workpiece is solved to obtain the optimal control sequence in the entire control time domain;

[0109] For example, given the initial state and target state Under the condition of , solve the following optimal deformation control problem:

[0110]

[0111] Constraints: , , , , ;

[0112] It should be noted that the initial state refers to the physical state of the flexible workpiece at the beginning of deformation control, including position, stress, deformation, etc. This information serves as the starting condition of deformation control and helps define the goals and constraints of the deformation control process;

[0113] Specifically, the continuous-time optimal deformation control problem is discretized into a nonlinear programming problem by using a finite difference method or a pseudospectral method;

[0114] Specifically, the control input sequence is introduced , express the state variable sequence as a function of the control input, and transform the nonlinear programming problem into a parameter optimization problem as follows:

[0115]

[0116] Constraints: , , , , ;

[0117] Specifically, numerical optimization algorithms, including sequential quadratic programming and interior point method, are used to solve the parameter optimization problem and obtain the sequence of optimal control inputs. , that is, the optimal control sequence.

[0118] Preferably, the feedforward control in the real-time control execution module S104 is to calculate the optimal control input based on the deformation prediction model and the current physical state of the flexible workpiece to compensate for the known dynamic characteristics of the system;

[0119] Specifically, from the optimal control sequence, the optimal control input at the current moment is extracted ;

[0120] Specifically, the feedforward control amount in the current state is calculated according to the deformation prediction model and the material parameters:

[0121]

[0122] in, For the flexible workpiece The state variables at time , for The optimal control input at time t, is the material parameter vector, It is a feedforward controller function, which is constructed based on the inverse solution, pseudo-inverse solution or other approximation methods of the deformation prediction model, and needs to comprehensively consider factors such as model accuracy and computational efficiency.

[0123] Specifically, the feedforward control quantity Outputting the output as a feedforward control instruction for the actuator, wherein the execution frequency of the feedforward control is consistent with the update frequency of the deformation prediction model;

[0124] Furthermore, the feedback correction in the real-time control execution module S104 is to adjust the optimal control input in real time according to the deviation between the system output and the expected output, so as to eliminate the error of the deformation prediction model and the influence of external interference.

[0125] Specifically, the actual state variables at the current moment are measured by sensors ;

[0126] The actual state variable and the predicted state variables Compare and calculate the deviation:

[0127]

[0128] According to the deviation size and PID control law, calculate the feedback control amount:

[0129]

[0130] in, 、 、 They are proportional, integral and differential coefficients respectively, which can be tuned by experiment or optimized by adaptive algorithm;

[0131] Furthermore, the feedforward control quantity and the feedback control quantity Add together and get the corrected control instructions:

[0132] Specifically, the control instruction Output is performed as the actuator control signal.

[0133] Furthermore, the actuator is controlled by the actuator control signal, so that the actuator completes real-time and accurate deformation control of the flexible workpiece;

[0134] It should be noted that the execution of the actuator is the final execution link of the control instruction, and its performance directly determines the effect of deformation control on the flexible workpiece.

[0135] For example, based on the characteristics of the controlled object and the control index requirements, an actuator of appropriate type and specification, such as a motor, hydraulic cylinder, or pneumatic element, is selected. Based on the electrical characteristics of the actuator, a suitable drive circuit, such as PWM drive or linear amplification, is designed. Through experimental testing, the actuator's input-output characteristic curve is established to determine the correspondence between the control command and the actuator's action. Based on the design requirements, the actuator is installed in a suitable location and mechanically debugged to ensure the actuator's range, accuracy, and stability. Based on the actuator's input-output characteristics, the control command is converted into a corresponding drive signal, such as voltage, current, or duty cycle. The resulting drive signal is then output to the actuator, driving it to complete the corresponding mechanical action.

[0136] In this embodiment, advanced multimodal sensor technology, deformation modeling and prediction methods, and optimized control strategies are employed to precisely control the deformation process of a flexible workpiece. The multimodal perception module integrates visual, force, and position data, providing comprehensive state perception capabilities. Industrial cameras, force sensors, and position sensors are used to collect workpiece deformation information in real time. After data preprocessing and fusion, an accurate state dataset is generated, ensuring the timeliness and accuracy of the information. The deformation modeling and prediction module, based on the collected data, accurately predicts the future deformation of the flexible workpiece through material parameter identification and deformation prediction models. It can handle workpieces made of elastic and plastic materials, and utilizes optimization techniques such as the finite element method and least squares method to ensure high-precision deformation prediction. The control strategy generation module, based on the predicted deformation state sequence, constructs an objective function and introduces deformation constraints to generate an optimal control strategy. This strategy not only considers deformation accuracy but also minimizes energy consumption during the control process through energy constraints. Multimodal data fusion and precise material parameter identification provide efficient prediction and control capabilities, ensuring that the deformation of the flexible workpiece meets process requirements while optimizing energy usage. This makes it suitable for high-precision, low-energy flexible workpiece deformation control tasks in modern manufacturing.

[0137] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope and spirit of the present invention are indicated by the claims. It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and that various modifications and changes can be made without departing from its scope. The scope of the present invention is limited only by the appended claims.

Claims

1. A flexible workpiece deformation control system, characterized in that: include: State perception module: This module collects the visual data, force data, and position data of the flexible workpiece in real time through multimodal sensors, performs data preprocessing and data fusion, and outputs a synchronized state data set. Deformation modeling and prediction module: Based on the state data set obtained by the state perception module, the module obtains the real-time material parameters of the flexible workpiece, including elastic parameter estimates and plastic parameter estimates, through real-time identification. Based on the material parameters, the module predicts the deformation state sequence of the flexible workpiece in the future prediction time domain in real time through the deformation prediction model; Control strategy generation module: Based on the deformation state sequence, construct an objective function and define constraints to solve the optimal control sequence, wherein the constraints include deformation constraints, yield constraints, and output constraints; The goal of deformation control is to minimize the control energy consumption while meeting the deformation accuracy requirements. The objective function is constructed as follows: ,in, represents the trajectory tracking error, represents the deformation constraint term, Represents the energy constraint term, which is used to control energy consumption. 、 、 is the weight coefficient; Real-time control execution module: Based on the optimal control sequence obtained by the control strategy generation module, the module outputs an actuator control signal through feedforward control and feedback correction to achieve real-time deformation control of the flexible workpiece.

2. A flexible workpiece deformation control system according to claim 1, characterized in that: The visual data includes: the three-dimensional coordinates of each key point on the surface of the flexible workpiece And the deformation of the corresponding position ; The force data includes: three-dimensional force vectors at key positions on the surface of the flexible workpiece , torque vector And the stress tensor at each key position ; The position data includes: the position vector of the origin of the flexible workpiece coordinate system in the base coordinate system , posture rotation matrix W and the spatial coordinates of the key points on the surface of the flexible workpiece.

3. A flexible workpiece deformation control system according to claim 1, characterized in that: The identification of the elastic parameter estimate includes: If the material of the flexible workpiece is an elastic material, the Young's modulus is used. and Poisson's ratio Describe its mechanical behavior, specifically: A certain number of displacement-force data pairs of key points are selected from the state data set as the input data set for parameter identification. Assuming that the flexible workpiece obeys the generalized Hooke's law, the following linear relationship exists: , where σ is the stress tensor, ε is the strain tensor, and E is the Young's modulus to be identified; Based on the position data, the strain tensor at each key point is calculated using the finite element method or other numerical methods. , solve the following optimization problem by the least squares method: , and obtain the Young's modulus The estimated value of Indicates the The displacement-force data pairs of key points, is the strain tensor at the i-th key point; The Poisson's ratio is estimated by the least squares method based on the transverse and longitudinal strain data of the flexible workpiece. , and finally obtain the elastic parameter estimate .

4. A flexible workpiece deformation control system according to claim 1, characterized in that: The identification of the estimated value of the plasticity parameter includes: If the material of the flexible workpiece is plastic material, in addition to the elastic parameter estimate In addition, the yield stress , hardening modulus , hardening index The plasticity parameters are: Select a certain number of displacement-force data pairs of key points from the state data set And the corresponding strain data As the input data set for parameter identification, a certain number of key points are selected to cover the elastic and plastic deformation stages of the flexible workpiece, and the initial estimates of the hardening parameters H and n are determined based on material experience or pre-experimental data. and ; Assuming that the flexible workpiece obeys the isotropic hardening criterion, the yield function is: ,in, Equivalent stress, is the equivalent plastic strain, is the initial yield stress, and is the hardening parameter to be identified, i.e., the initial estimate; The initial yield stress is estimated by the least squares method based on the elastic-plastic transition point data of the flexible workpiece. , using the Levenberg-Marquardt algorithm, solve the following optimization problem: , and obtain the hardening parameters and The estimated value of According to the initial yield stress and initial estimates and , the stress value calculated by the yield function; Repeat the calculation until the hardening parameter and The estimated value of converges, and finally the estimated value of the plastic parameter of the flexible workpiece is obtained .

5. The flexible workpiece deformation control system according to claim 1, characterized in that: The deformation prediction model is established based on the material parameters and is used to predict the deformation of the flexible workpiece. The construction process of the deformation prediction model is specifically as follows: According to the length of the future prediction time domain, the deformation prediction of the flexible workpiece is divided into short-term prediction and medium-term prediction; The short-term prediction is used to predict the deformation state of the flexible workpiece in the next 0.1 seconds, and is performed using a linear elastic finite element model that uses elastic parameter estimates updated in real time. , accurately calculate the response of the workpiece in the elastic deformation stage, interpolate the predicted node state quantity to the surface of the flexible workpiece, and obtain the future Deformation state sequence of control cycles .

6. A flexible workpiece deformation control system according to claim 5, characterized in that: The mid-term prediction is used to predict the deformation state of the flexible workpiece in the next 1 second, and the prediction is made using an elastic-plastic finite element model, which introduces the estimated plasticity parameters obtained by real-time identification. , in order to reflect the nonlinear characteristics of the workpiece when plastic deformation occurs, the predicted node state quantity is interpolated to the surface of the flexible workpiece to obtain the future Deformation state sequence of control cycles ; Will and Splicing is performed to obtain the complete deformation state sequence of the next m control cycles: ,in Indicates that at the current moment Afterwards The predicted deformation state of each control cycle.

7. The flexible workpiece deformation control system according to claim 5, characterized in that: The trajectory tracking error It is used to measure the deviation between the actual deformation state and the expected deformation state of the flexible workpiece, and is defined as: ,in for The actual deformation state at the moment, for The desired deformation state at the moment, is the tracking error weight matrix; The desired deformation state is an ideal deformation state sequence of the flexible workpiece generated according to the processing requirements. ,in is the number of control cycles in the future prediction time domain; the actual deformation state is obtained from the deformation state sequence Extract the deformation state sequence at the corresponding moment ; Calculate the deformation deviation vector at each moment: ; The tracking error weight matrix Perform weighted accumulation on the deformation deviation vector to obtain the trajectory tracking error : .

8. The flexible workpiece deformation control system according to claim 5, characterized in that: The energy constraint term is used to measure the energy consumption during the deformation control process to achieve energy-saving optimization and is defined as: ,in is the optimal control input, is the control energy weight matrix; Extract the optimal control input at the current moment from the optimal control sequence of the previous control cycle ; According to the control energy weight matrix , calculate the optimal control input The energy consumption is calculated as follows: .

9. The flexible workpiece deformation control system according to claim 8, characterized in that: The feedforward control is to calculate the optimal control input based on the deformation prediction model and the current physical state of the flexible workpiece to compensate for the known dynamic characteristics of the system; From the optimal control sequence, extract the optimal control input at the current moment ; According to the deformation prediction model and the material parameters, the feedforward control amount in the current state is calculated: ,in, For the flexible workpiece The state variables at time , for The optimal control input at time t, is the material parameter vector, is the feedforward controller function.

10. The flexible workpiece deformation control system according to claim 1, characterized in that: The feedback correction specifically includes: Measure the actual state variables at the current moment through sensors ; The actual state variable and the predicted state variables For comparison, calculate the deviation: ; According to the deviation size and PID control law, calculate the feedback control amount: ,in, 、 、 They are proportional, integral and differential coefficients respectively, which can be optimized through experimental tuning or adaptive algorithm.

Images