A tool built-up edge adhesion identification and control method and system in an aluminum processing process

Abstract

The application discloses a tool built-up edge adhesion identification and control method and system in an aluminum processing process, and relates to the technical field of numerical control machining. The method comprises the following steps: firstly, collecting three types of machining signals of acoustic emission, contact resistance and cutting temperature in the contact area of the tool and the workpiece in the aluminum material machining of the machine tool; secondly, calculating the root mean square value of the acoustic emission, the fluctuation amplitude of the contact resistance and the cutting temperature rising rate; and thirdly, synthesizing a continuous built-up edge adhesion index for representing the adhesion degree of the built-up edge. The adhesion index is sent to a model predictive controller, the controller takes the current machining state as the initial state, relies on the predictive model containing the temperature and stress field characteristic state variables, combines the observation value and the feedback output, and predicts the subsequent evolution trend of the adhesion index. Taking the minimum of the cumulative adhesion index square sum as the target, the change rate of the index and the machining parameter constraint are considered to solve the optimal control sequence, and the first step parameter is output to control the machine tool. By means of rolling prediction and optimization, the built-up edge adhesion in the aluminum processing is precisely and ahead of time inhibited.

Description

Technical Field

[0001] This invention relates to the field of CNC machining technology, and in particular to a method and system for identifying and controlling built-up edge adhesion of cutting tools during aluminum machining. Background Technology

[0002] Built-up edge adhesion is one of the core issues affecting machining quality and tool life during aluminum processing. Especially in the machining of high-strength aluminum alloys, chips are prone to welding to the tool surface under high temperature and pressure, leading to a surge in cutting force, deterioration of surface quality, and severely restricting machining efficiency and product accuracy.

[0003] Currently, addressing built-up edge (BUE) adhesion during aluminum machining relies primarily on operator experience or offline detection, making real-time online intervention difficult. While chatter suppression methods based on cutting force, temperature, and vibration signals have emerged, employing PID controllers to adjust machining parameters by identifying chatter states, these methods still largely depend on mechanical vibration and fail to solve the complex machining problem of BUE adhesion, which has a different physical mechanism. Furthermore, the cutting force signals used are insensitive to the high-frequency elastic wave response in the early stages of BUE formation, and PID control lacks the ability to predict the evolution of machining conditions, making it difficult to achieve proactive and precise adhesion suppression. Summary of the Invention

[0004] To address the problem in existing technologies where built-up edge (BUE) adhesion during aluminum machining relies on mechanical vibration and is difficult to suppress in a timely and precise manner, this invention provides a method and system for identifying and controlling BUE adhesion during aluminum machining. By predicting future evolution trends based on an adhesion index and using rolling optimization to solve for the optimal machining parameter sequence, advanced suppression of BUE adhesion during aluminum machining is achieved. The specific technical solution is as follows: This application provides a method for identifying and controlling built-up edge adhesion in aluminum machining processes, including the following steps: The machining signals are collected from the contact area between the tool and the workpiece during the machining of aluminum by the machine tool; the machining signals include acoustic emission signals, tool-workpiece contact resistance signals, and cutting temperature signals. Based on the acoustic emission signal, contact resistance signal, and cutting temperature signal, the root mean square value of the acoustic emission signal, the fluctuation amplitude of the contact resistance signal, and the temperature rise rate of the cutting temperature signal are calculated respectively, and the built-up edge adhesion index is calculated; wherein, the built-up edge adhesion index is a continuous value used to characterize the degree of adhesion of the built-up edge. The pylomast adhesion index is input as a feedback signal to the model prediction controller, which performs the following operations sequentially in each control cycle: Using the current machining state as the initial state, the root mean square value, fluctuation amplitude, and temperature rise rate are taken as the observed values ​​of the state variables, and the built-up edge adhesion index is taken as the output feedback. The predictive control model is used to predict the evolution trend of the built-up edge adhesion index in the future prediction time domain. In the state equation of the model, the state variables include the characteristic values ​​of the temperature field and the characteristic values ​​of the stress field in the cutting zone, and the output variable is the built-up edge adhesion index. The optimal control sequence is obtained by taking the minimum sum of squares of the cumulative buildup adhesion index in the prediction time domain as the objective function, and combining the rate of change constraint of the buildup adhesion index and the range of change constraint of the processing parameters. The control variable value corresponding to the first moment in the optimal control sequence is output to the machine tool actuator to adjust the machining parameters at the next moment.

[0005] Preferably, the adhesion index of the pyometra is... The calculation formula is: in, is the root mean square value of the acoustic emission signal; The root mean square reference value of acoustic emission in the non-adhesive state; This is the reference value for contact resistance in a non-adhesive state; This represents the fluctuation amplitude of the contact resistance signal; The temperature rise rate of the cutting temperature signal; This is the baseline value for the rate of temperature rise under non-adhesive conditions. , , Preset weighting coefficients and satisfying ; Adhesion index of pyometra The values ​​are continuous, and their magnitude directly represents the degree of adhesion of the pyroplastic tumor. , , >1 indicates the nonlinear amplification index.

[0006] Preferably, before calculating the root mean square value of the acoustic emission signal, the method further includes: Multi-level wavelet packet decomposition is performed on the acoustic emission signal to decompose it into multiple frequency band components with different frequency ranges. The proportion of the energy of each frequency band component to the total energy of the acoustic emission signal is calculated. When the energy proportion of the frequency band components obtained from the decomposition exceeds a preset threshold, the contribution weight of the acoustic emission signal is adjusted. The frequency band energy ratio is dynamically adjusted based on the wavelet packet decomposition result. The adjustment formula is as follows: in, For frequency band energy; Total energy; The preset enhancement coefficient; the adjusted weights , , Renormalization to meet .

[0007] Preferably, the objective function is: in, The objective function value; To predict the length of the time domain; For the first Time information for the first Predicted value of the adhesion index of drupture at any time; For the first Time information for the first The predicted value of the control variable is controlled at all times; For the first Time information for the first The predicted value of the control variable is controlled at all times; This is the adhesion index penalty coefficient; The penalty coefficient for changes in the control quantity; It is the square of the Euclidean norm, which is the sum of the squares of the changes in each control variable.

[0008] Preferably, the step of using a model to predict the evolution trend of the adhesion index of the pyroplastic tumor in the future prediction time domain specifically includes: Multiple prediction sub-models corresponding to different pyroplastic adhesion states are pre-established. In each control cycle, the corresponding sub-model is selected from the multiple prediction sub-models for prediction based on the current value and trend of the pyroplastic adhesion index. Within the state transition interval corresponding to different sub-models, the prediction results of adjacent sub-models are weighted and fused.

[0009] Preferably, the plurality of prediction sub-models are established in the following manner: Historical data on the adhesion index of built-up edge during processing at different numerical ranges were collected, and the model parameters of each sub-model were determined by a system identification method. The system identification method includes any one of the following: least squares method, maximum likelihood method, or subspace identification method.

[0010] Preferably, when using the model to predict the evolution trend of the adhesion index of the pyroplastic tumor in the future prediction time domain, it further includes: The tooth passage frequency is calculated by the spindle speed and the number of tool teeth, and a sinusoidal periodic signal of the tooth passage frequency and its at least one harmonic is constructed as a known disturbance signal. The known disturbance signal is introduced as an additional input term into the state equation of the predictive control model.

[0011] Preferably, adjusting the processing parameters for the next moment specifically includes: The adjustment amount of the processing parameters is coordinated and transformed according to the preset proportional coefficient. The proportional coefficient is adjusted according to the current chip accumulation value, and the processing parameters at the next moment are calculated according to the adjusted proportional coefficient. The machining parameters include spindle speed or cutting line speed, feed rate or feed per revolution, and depth of cut or depth of cut.

[0012] Preferably, the root mean square value of the acoustic emission signal The calculation is performed using a sliding window recursive algorithm, and the sliding window length is... With spindle speed satisfy ,in, The sampling frequency of the acoustic emission signal. It is a preset integer.

[0013] This application also provides a tool built-up edge adhesion identification and control system for aluminum machining processes, which applies the aforementioned method and includes: The signal acquisition module is used to acquire machining signals in the contact area between the tool and the workpiece during the machining of aluminum by the machine tool; the machining signals include acoustic emission signals, tool-workpiece contact resistance signals, and cutting temperature signals. The adhesion index calculation module is used to calculate the root mean square value of the acoustic emission signal, the fluctuation amplitude of the contact resistance signal, and the temperature rise rate of the cutting temperature signal based on the acoustic emission signal, the contact resistance signal, and the cutting temperature signal, and to calculate the built-up edge adhesion index; wherein, the built-up edge adhesion index is a continuous value used to characterize the degree of adhesion of the built-up edge. The model prediction control module is used to input the tumour adhesion index as a feedback signal to the model prediction controller, and the model prediction controller performs the following operations sequentially in each control cycle: Using the current machining state as the initial state, the root mean square value, fluctuation amplitude, and temperature rise rate are taken as the observed values ​​of the state variables, and the built-up edge adhesion index is taken as the output feedback. The predictive control model is used to predict the evolution trend of the built-up edge adhesion index in the future prediction time domain. In the state equation of the model, the state variables include the characteristic values ​​of the temperature field and the characteristic values ​​of the stress field in the cutting zone, and the output variable is the built-up edge adhesion index. The optimal control sequence is obtained by taking the minimum sum of squares of the cumulative buildup adhesion index in the prediction time domain as the objective function, and combining the rate of change constraint of the buildup adhesion index and the range of change constraint of the processing parameters. The control variable value corresponding to the first moment in the optimal control sequence is output to the machine tool actuator to adjust the machining parameters at the next moment.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention presents a method for identifying and controlling built-up edge (BUE) adhesion in aluminum machining. By fusing multi-source signals from acoustic emission, contact resistance, and cutting temperature, a continuously quantified BUE adhesion index is constructed, enabling real-time and accurate identification of the adhesion state. A model predictive controller is employed to predict future evolution trends based on the adhesion index, and the optimal machining parameter sequence is solved through rolling optimization to achieve proactive inhibition of adhesion. Compared to existing technologies, this method significantly reduces the incidence of BUE, improves surface finish, extends tool life, and maintains stable machining efficiency. Attached Figure Description

[0015] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0016] Figure 1 This is a flowchart illustrating a method for identifying and controlling built-up edge adhesion in aluminum machining processes, as provided in an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram showing the installation position of the signal acquisition device on a milling machine tool according to an embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram of the curve data of the adhesion index of pyroplastic tumor provided in an embodiment of the present invention.

[0019] Figure 4 This is a schematic diagram of milling speed adjustment curve data provided in an embodiment of the present invention.

[0020] Figure 5 This is a schematic diagram of the feed rate adjustment curve data provided in an embodiment of the present invention.

[0021] Figure 6 This is a schematic diagram of milling depth adjustment curve data provided in an embodiment of the present invention.

[0022] Figure 7 The present invention provides a principle for a tool built-up edge adhesion identification and control system in aluminum processing.

[0023] Explanation of reference numerals in the attached diagram: 1-Acoustic emission sensor, 2-Conductive slip ring, 3-Thermocouple. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] It should be understood that, when used in this specification, the terms “comprising” and “including” indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0026] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0027] It should also be further understood that the term "and / or" as used in this specification refers to any combination of one or more of the associated listed items, as well as all possible combinations, and includes such combinations.

[0028] Please refer to the following examples. Figures 1 to 7 .

[0029] This application provides a method for identifying and controlling built-up edge adhesion in aluminum machining processes, including the following steps: Step S1: Collect machining signals from the contact area between the tool and the workpiece during the machining process of aluminum material by the machine tool; the machining signals include acoustic emission signals, tool-workpiece contact resistance signals, and cutting temperature signals; In the aluminum processing, an acoustic emission sensor is installed on the machine tool spindle. The sensor uses a contact mounting method, with its sensitive surface in close contact with the tool holder, and a coupling agent is used to enhance the transmission of high-frequency signals. The sampling frequency of the acoustic emission sensor is set to no less than 1MHz to capture the high-frequency elastic waves generated during the formation and shedding of built-up edge.

[0030] The contact resistance signal is acquired using a four-wire measurement method: a constant microcurrent of 1mA to 10mA is applied between the tool and the workpiece, the voltage drop between the tool and the workpiece is measured through two independent wires, and the contact resistance value is calculated according to Ohm's law. The sampling frequency is set to be no less than 10kHz.

[0031] The cutting temperature signal is acquired using the thermocouple embedding method: a microhole with a diameter of no more than 0.5 mm is machined at a distance of 0.1 mm to 0.3 mm from the cutting edge on the tool rake face. The thermocouple wire is embedded in the microhole and filled with thermally conductive insulating glue. The thermocouple output is connected to the data acquisition card, and the sampling frequency is set to no less than 100 Hz.

[0032] Installation location and requirements for the three-channel signal acquisition device (corresponding to acoustic emission sensor, conductive slip ring, and thermocouple), please refer to [reference needed]. Figure 2 The diagram shows a simplified representation of a milling machine. Based on the installed acoustic emission sensors, conductive slip rings, and thermocouples, data acquisition can be triggered synchronously, ensuring strict time alignment of all signals and forming a multi-source signal data matrix.

[0033] Step S2: Based on the acoustic emission signal, contact resistance signal, and cutting temperature signal, calculate the root mean square value of the acoustic emission signal, the fluctuation amplitude of the contact resistance signal, and the temperature rise rate of the cutting temperature signal, and calculate the built-up edge adhesion index; wherein, the built-up edge adhesion index is a continuous value used to characterize the degree of adhesion of the built-up edge. The acquired acoustic emission signal was bandpass filtered, with a passband frequency range of 50kHz to 500kHz, to remove low-frequency vibration noise and high-frequency circuit noise. The filtered signal sequence is as follows: The root mean square value is calculated using a sliding window recursive algorithm, with a window length of... Set to cover an integer number of spindle rotation cycles: in, The length of the sliding window is expressed in units of sampling points. The number of spindle rotation cycles covered by the window, a positive integer from 1 to 5; The sampling frequency of the acoustic emission signal; Main spindle frequency; The main spindle speed.

[0034] During initialization, the front The root mean square value of each sampling point is calculated using the standard formula as follows: The fluctuation amplitude of the contact resistance signal The calculation formula is: in, and These are the maximum and minimum values ​​of the contact resistance signal within the preset time window, respectively.

[0035] An adaptive differential method is used to calculate the temperature rise rate dT / dt from the cutting temperature signal. The differential step size is dynamically adjusted based on the current temperature value. A larger step size is used in the low-temperature region to ensure stability, while a smaller step size is used in the high-temperature region to improve sensitivity. The formula for calculating the temperature rise rate is: in, The current rate of temperature increase; For interval The temperature value afterwards; This represents the current temperature value.

[0036] The calculated , , Substituting the values ​​into a pre-calibrated calculation formula, the continuous numerical form of the piloerection adhesion index is obtained. This index approaches 1 in the absence of adhesion and increases monotonically with increasing adhesion; a larger value indicates more severe adhesion.

[0037] In this embodiment, the adhesion index of the tumour is... The calculation formula is: in, is the root mean square value of the acoustic emission signal; The root mean square reference value of acoustic emission in the non-adhesive state; This is the reference value for contact resistance in a non-adhesive state; This represents the fluctuation amplitude of the contact resistance signal; The temperature rise rate of the cutting temperature signal; This is the baseline value for the rate of temperature rise under non-adhesive conditions. , , Preset weighting coefficients and satisfying ; Adhesion index of pyometra The value is a continuous value, ranging from [1, +∞). The larger the value, the more severe the adhesion. , , >1 indicates a nonlinear amplification index, such as taking =1.5, =1.3, =1.8.

[0038] Step S3: The adhesion index of the drupture is input as a feedback signal to the model prediction controller. The model prediction controller performs the following operations sequentially in each control cycle: Step S31: Taking the current machining state as the initial state, the root mean square value, fluctuation amplitude, and temperature rise rate are taken as the observed values ​​of the state variables, and the built-up edge adhesion index is taken as the output feedback. The predictive control model is used to predict the evolution trend of the built-up edge adhesion index in the future prediction time domain. In the state equation of the model, the state variables include the characteristic values ​​of the temperature field and the characteristic values ​​of the stress field in the cutting zone, and the output variable is the built-up edge adhesion index. In practice, the current machining state is used as the initial state to construct a state variable vector. This vector contains temperature field characteristic values ​​reflecting the temperature distribution characteristics of the cutting zone and stress field characteristic values ​​reflecting the stress distribution characteristics of the cutting zone. The real-time calculated root mean square value of acoustic emission, contact resistance fluctuation amplitude, and temperature rise rate are used as observed values ​​of the state variables. The state variables are optimally estimated using a Kalman filter or extended Kalman filter. The estimated state variables are then substituted into the predictive control model. Using the control variables at the current moment (including spindle speed or cutting linear velocity, feed rate or feed per revolution, depth of cut or depth of cut) as input, a numerical integration method is used to recursively solve for the evolution trajectory of the state variables and built-up edge adhesion index for the next Np moments (Np ranges from 5 to 20).

[0039] Let state prediction construct a state variable vector This includes the characteristic values ​​of the temperature field and stress field in the cutting zone: in, For the first Average temperature of the cutting zone at any given time; For the first Temperature gradient in the cutting zone at all times; For the first Time-equivalent stress; For the first Constant shear stress; Observation variable vector Including features calculated in real time: State estimation is performed using an extended Kalman filter. Prediction step: in, For based on Time information Prior estimation of the state at time step; For the first The vector of control variables at any given time; The state transition Jacobian matrix; The prior estimate is the error covariance matrix; Let be the process noise covariance matrix.

[0040] Update steps: in, The Kalman gain matrix; To observe the Jacobian matrix; To observe the noise covariance matrix; For posterior state estimation; This is the posterior estimation error covariance matrix.

[0041] Substitute the estimated state into the predictive control model and recursively solve for the future. The state and output at each moment: in, To predict the time domain length, a typical value is taken as... ; For based on Time information The state at time point Step prediction; For the first Candidate control variables at time points; For the first Predictive value of the adhesion index of drupture at any given time.

[0042] Step S32: Using the minimum sum of squares of the accumulated buildup adhesion index in the prediction time domain as the objective function, and combining the constraint of the rate of change of the buildup adhesion index and the constraint of the change range of the processing parameters, the optimal control sequence is obtained. The objective function is to minimize the sum of squares of the predicted pyroplastic adhesion index at each time point in the prediction time domain. This objective function also includes a sum of squares of the changes in the control variables to suppress drastic fluctuations in the control variables.

[0043] Construct an objective function to minimize the sum of squares of the accumulated adhesion index of pyroplasia within the time domain: in, The objective function value; To predict the time domain length, the value is taken as... Positive integers; For the first Time information for the first Predicted value of the adhesion index of drupture at any time; For the first Time information for the first The predicted value of the control variable is controlled at all times; For the first Time information for the first The predicted value of the control variable is controlled at all times; The adhesion index penalty coefficient has a value of [value missing]. ; The penalty coefficient for changes in the control quantity is set to a value of [value missing]. ; It is the square of the Euclidean norm, which is the sum of the squares of the changes in each control variable.

[0044] The control variable vector is represented as: in, For milling speed; For feed rate; This represents the cutting depth.

[0045] The constraints include: (1) Constraint on the rate of change of adhesion index of pylomas: in, To be the maximum allowable rate of change, take .

[0046] (2) Constraints on the range of change of control variables: in, To determine the maximum change in milling speed, take... rpm; To determine the maximum change in feed rate, take... mm / min; To determine the maximum change in cutting depth, take... mm.

[0047] (3) Upper and lower limit constraints for control variables: The optimization problem is: in, To control the control sequence in the time domain; To control the length of the time domain, typically take ,Right now Pick .

[0048] The objective function and constraints are transformed into a standard quadratic programming form using a quadratic programming algorithm. in, The Hessian matrix; This is the vector of linear term coefficients; This is the constraint matrix; For the constraint boundary vector; These are constraints.

[0049] Step S33: Output the control variable value corresponding to the first moment in the optimal control sequence to the machine tool actuator to adjust the machining parameters for the next moment.

[0050] The control variable values ​​for the first moment are extracted from the optimal control sequence obtained from the solution and transmitted to the CNC system or servo drive unit of the machine tool via fieldbus or real-time Ethernet. The actuator adjusts the spindle speed or cutting speed, feed rate or feed per revolution, depth of cut or depth of cut in real time according to the received control commands. At the beginning of the next control cycle, the state estimation and optimization solution are re-performed using the newly acquired machining signals.

[0051] Extract the first control variable from the optimal control sequence: The first control input is extracted and transmitted to the CNC system, where it is converted into machine tool execution instructions (taking milling as an example): milling speed instruction. Feed rate command Cutting depth command ; In this embodiment, a bias compensation mechanism is also used to perform online correction of the prediction model: in, For the first The prediction error at any given time; For the first The adhesion index of the drupture was measured in real time. For the first The predicted value at any given time.

[0052] Apply an exponentially weighted moving average filter to the prediction error: in, This represents the average prediction error after filtering. This is a smoothing factor with a value of 0.2. This represents the average prediction error after filtering at the previous time step.

[0053] The filtered deviation is added to the prediction output as a feedforward compensation term: in, The compensated predicted value is used for subsequent optimization calculations.

[0054] At the start of the next control cycle, the newly acquired processing signals are used. Re-perform state estimation and optimization. Control cycle. Set to no greater than ,Right now: Corresponding control frequency This meets the real-time requirements.

[0055] It should be noted that the above embodiments primarily focus on milling in aluminum processing, and the machining parameters involved include milling speed, feed rate, and depth of cut. Actual aluminum processing encompasses various methods such as milling, turning, drilling, and grinding, with the adjusted machining parameters corresponding to different processing methods, as illustrated below: In other embodiments, the pyometra adhesion index The calculation formula can also be: This invention provides a method for identifying and controlling built-up edge (BUE) adhesion in aluminum machining. By fusing multi-source signals from acoustic emission, contact resistance, and cutting temperature, a continuously quantified BUE adhesion index is constructed, enabling real-time and accurate identification of the adhesion state. A model predictive controller is employed to predict future evolution trends based on the adhesion index, and the optimal machining parameter sequence is solved through rolling optimization to achieve proactive inhibition of adhesion. Compared to existing technologies, this method significantly reduces the incidence of BUE, improves surface finish, extends tool life, and maintains stable machining efficiency.

[0056] Specifically, in a preferred embodiment of this application, before calculating the root mean square value of the acoustic emission signal, the method further includes: Multi-level wavelet packet decomposition is performed on the acoustic emission signal to decompose it into multiple frequency band components with different frequency ranges. The proportion of the energy of each frequency band component to the total energy of the acoustic emission signal is calculated. In practice, the acoustic emission signal is decomposed into j-level wavelet packets, such as j being 3 to 5, resulting in 2^j frequency band components. The wavelet basis functions used are Daubechies wavelet (db8) or Symlets wavelet (sym8). The energy of each frequency band component is calculated as follows: in, The sampling interval; This is the reconstructed signal for the k-th frequency band.

[0057] The total energy is: The energy proportion of the high-frequency band components is: in, This is a set of high-frequency band indices, defined as frequencies higher than [a certain value]. The frequency band.

[0058] When the energy proportion of the high-frequency band components obtained from the decomposition exceeds a preset threshold At times, such as The value is 0.5, representing the contribution weight of the acoustic emission signal. The adjustment is dynamically made based on the proportion of high-frequency band energy after wavelet packet decomposition. The adjustment formula is as follows: in, In this embodiment, the energy is defined as the energy in the frequency band above 200kHz. For the total energy, ; The preset enhancement coefficient; the adjusted weights , , Renormalization to meet .

[0059] By dynamically adjusting the weight of the acoustic emission signal according to the proportion of high-frequency energy, when the proportion of high-frequency energy is high, the acoustic emission signal has fully reflected the edema event. Appropriately reducing its weight can balance the fusion with other signals and avoid the dominance of a single signal.

[0060] In some embodiments, the adhesion index penalty coefficient With the weighting coefficient , , The values ​​of these parameters are correlated to reflect the contribution of each signal to control optimization. In this implementation, a simplified correlation method can be used: Right now Take a fixed value (Typically 1.0), is proportional to the sum of the weighting coefficients (the sum of the weighting coefficients is always 1), simplifying the parameter calibration process.

[0061] Penalty coefficient for changes in control quantity It can be set to a fixed value or a state-dependent value. State dependency When the adhesion index is high, it decreases. This allows for significant adjustments.

[0062] By simplifying parameter calibration and dynamically adjusting the control intensity based on the adhesion index, rapid deployment and adaptive optimization of the control strategy are achieved.

[0063] Specifically, in a preferred embodiment of this application, the step of using a model to predict the evolution trend of the pyroplastic adhesion index in the future prediction time domain specifically includes: Multiple prediction sub-models corresponding to different pyroplastic adhesion states are pre-established. In each control cycle, the corresponding sub-model is selected from the multiple prediction sub-models for prediction based on the current value and trend of the pyroplastic adhesion index. Within the state transition interval corresponding to different sub-models, the prediction results of adjacent sub-models are weighted and fused.

[0064] The multiple prediction sub-models are established in the following way: Historical data on the adhesion index of built-up edge during processing at different numerical ranges were collected, and the model parameters of each sub-model were determined by a system identification method. The system identification method includes any one of the following: least squares method, maximum likelihood method, or subspace identification method.

[0065] In practice, multiple predictive sub-models are pre-established, each corresponding to a different range of adhesion states of pyroplastic tumors. In a typical implementation, three sub-models can be established: the first corresponds to the normal state with an adhesion index below 1.2, the second corresponds to the mild adhesion state with an adhesion index between 1.2 and 2.0, and the third corresponds to the severe adhesion state with an adhesion index above 2.0.

[0066] Each sub-model adopts a controlled autoregressive integral moving average model, with the following structure: Each sub-model adopts the form of a controlled autoregressive integral moving average model: in, The adhesion index of the plutoma; For control variables; For sub-model indexing, Corresponding to four states; For discrete-time indexing; It is a white noise sequence; This is the shift operator.

[0067] in, This represents the autoregressive part of the polynomial; These are the autoregressive coefficients. For autoregressive order; This represents the polynomial controlling the input portion; For control coefficients, To control the order; Represents the moving average partial polynomial; The moving average coefficient; The order of the moving average; The adhesion index of the plutoma; For control variables; It is a difference operator.

[0068] The parameters of the sub-models were determined by the system identification method: historical data of the built-up edge adhesion index in different ranges during the processing were collected, and the parameters were estimated by the least squares method, the maximum likelihood method or the subspace identification method to obtain the model parameters of each sub-model.

[0069] In each control cycle, a corresponding sub-model is selected for prediction based on the current magnitude and trend of the piloerection adhesion index. When the adhesion index falls within the transition interval between two sub-models, the prediction results of adjacent sub-models are weighted and fused. The weights for weighted fusion are determined based on the distance between the current adhesion index and the center points of the two state intervals; the closer the distance, the higher the weight of the corresponding sub-model. The fused prediction value changes smoothly within the transition interval, avoiding abrupt changes in the prediction results. In practice, if the current state interval is the same as the previous cycle, the corresponding sub-model is used directly for prediction. If the state changes, a weighted fusion method is used for smooth switching. in, These are the predicted values ​​from the original state sub-model. The predicted values ​​for the new state sub-model; To integrate weights, satisfy .

[0070] By establishing a multi-state sub-model and implementing weighted fusion switching, the problem that a single model cannot accurately describe the dynamic characteristics of built-up edge throughout its entire lifecycle is solved. The strategy based on this embodiment can adaptively match the optimal prediction model according to the degree of adhesion, achieving smooth switching in state transition intervals, effectively avoiding abrupt changes in prediction output, and significantly improving the prediction accuracy of the model under different processing conditions.

[0071] Specifically, in a preferred embodiment of this application, when using a model to predict the evolution trend of the adhesion index of pyroplasia in the future prediction time domain, it further includes: The tooth passage frequency is calculated by the spindle speed and the number of tool teeth, and a sinusoidal periodic signal of the tooth passage frequency and its at least one harmonic is constructed as a known disturbance signal. The known disturbance signal is introduced as an additional input term into the state equation of the predictive control model.

[0072] In this embodiment, the state space of the predictive control model can be represented as: in, The state variables at time k include the characteristic values ​​of the temperature field and stress field in the cutting zone; The control variables include milling speed, feed rate, and depth of cut; The observed variables include the root mean square value of the acoustic emission signal, the fluctuation amplitude of the contact resistance signal, and the temperature rise rate of the cutting temperature signal. The adhesion index of the plutoma; , , Nonlinear mapping functions identified from historical processing data.

[0073] The calculated disturbance sequence is substituted into the state equation as a known quantity for recursion, thus eliminating the interference of periodic fluctuations on the prediction.

[0074] Specifically, in a preferred embodiment of this application, adjusting the processing parameters for the next moment specifically includes: The adjustment amount of the processing parameters is coordinated and transformed according to the preset proportional coefficient. The proportional coefficient is adjusted according to the current chip accumulation value, and the processing parameters at the next moment are calculated according to the adjusted proportional coefficient. The material removal rate is calculated based on the processing type, taking milling as an example. The formula is as follows: Milling process: ; Turning: ; Drilling process: ; in, Material removal rate; The cutting linear velocity; This refers to the feed per tooth. Feed per revolution; This refers to the depth of cut. This refers to the milling width; The diameter of the drill bit; The main spindle speed.

[0075] The adjustment amount for processing parameters is allocated proportionally: in, The adjustment amount for the spindle speed or cutting line speed; This is an adjustment amount for the feed rate or feed per revolution; This is the adjustment amount for the depth of cut or the depth of cut; Let be the proportionality coefficient, satisfying .

[0076] The proportional coefficient is dynamically adjusted based on the buildup adhesion index. Dynamic adjustment: in, This is the default scaling factor, typically set to... , , ; The first threshold is typically taken as... ; The second threshold is typically taken as... ; To adjust the amplitude coefficient, the range of values ​​is... Typical .

[0077] The material removal rate is kept within a preset deviation range from the target value when adjusting parameters. in, To determine the target material removal rate, either the current material removal rate or a user-defined value is used. The allowable deviation coefficient is typically taken as... .

[0078] The final output parameters are calculated as follows: in, , , It is determined by the optimization results of the objective function and the proportional allocation.

[0079] By employing a feasibility-driven strategy, the boundary of the rate of change of the adhesion exponent is dynamically relaxed when constraints conflict, prioritizing the physical feasibility of adjusting control parameters and ensuring that the optimal solution remains valid. This avoids control interruptions caused by solution failures, enabling the optimization process to adapt to complex and changing processing conditions.

[0080] This application also provides a tool built-up edge adhesion identification and control system for aluminum machining processes, including: The signal acquisition module is used to acquire machining signals in the contact area between the tool and the workpiece during the machining of aluminum by the machine tool; the machining signals include acoustic emission signals, tool-workpiece contact resistance signals, and cutting temperature signals. The adhesion index calculation module is used to calculate the root mean square value of the acoustic emission signal, the fluctuation amplitude of the contact resistance signal, and the temperature rise rate of the cutting temperature signal based on the acoustic emission signal, the contact resistance signal, and the cutting temperature signal, and to calculate the built-up edge adhesion index; wherein, the built-up edge adhesion index is a continuous value used to characterize the degree of adhesion of the built-up edge. The model prediction control module is used to input the tumour adhesion index as a feedback signal to the model prediction controller, and the model prediction controller performs the following operations sequentially in each control cycle: Using the current machining state as the initial state, the root mean square value, fluctuation amplitude, and temperature rise rate are taken as the observed values ​​of the state variables, and the built-up edge adhesion index is taken as the output feedback. The predictive control model is used to predict the evolution trend of the built-up edge adhesion index in the future prediction time domain. In the state equation of the model, the state variables include the characteristic values ​​of the temperature field and the characteristic values ​​of the stress field in the cutting zone, and the output variable is the built-up edge adhesion index. The optimal control sequence is obtained by taking the minimum sum of squares of the cumulative buildup adhesion index in the prediction time domain as the objective function, and combining the rate of change constraint of the buildup adhesion index and the range of change constraint of the processing parameters. The control variable value corresponding to the first moment in the optimal control sequence is output to the machine tool actuator to adjust the machining parameters at the next moment.

[0081] The functional explanation of each unit in this embodiment is the same as that of a tool built-up edge adhesion identification and control system in an aluminum processing process, and the technical effect is the same, so it will not be repeated here.

[0082] To verify the effectiveness of the method for identifying and controlling built-up edge adhesion in aluminum machining according to an embodiment of this application, a comparative experiment was conducted by setting up a proportional-integral-derivative (PID) control scheme from the existing scheme and a model predictive control (MPC) scheme from this application for analysis.

[0083] I. Comparison of Schemes II. Experimental Conditions (a) Processing parameters: The workpiece material is 7150-T6 aluminum alloy; the cutting tool is a carbide end mill with a diameter of 10mm and 2 teeth; the spindle speed is 8000rpm; the initial milling speed is 251m / min (corresponding to a spindle speed of 8000rpm and a cutting tool diameter of 10mm); the initial feed rate is 0.12mm / z (corresponding to a feed rate of 1920mm / min); the initial depth of cut is 0.2mm; and the width of cut is 5mm.

[0084] (II) Conditions for inducing pyroplasia: At the 10th second, the feed rate was increased to 0.18 mm / z to continuously induce the formation of buildup. At the 30th second, restore the initial feed rate and observe the control effect.

[0085] (III) Control Parameters: Control period set to 50ms; prediction time domain set to 10 (corresponding to 500ms); adhesion index penalty coefficient. =1; The penalty coefficient for changes in the control quantity is 0.1.

[0086] III. Performance Comparison Conclusion according to Figure 3 The curve showing the change in adhesion index of the pyroplastic tumor, and Figures 4-6 The adjustment curves for milling speed, feed rate, and depth of cut, as shown, reveal that the peak value of the built-up edge adhesion index in this application is 1.24, a 35.4% reduction compared to the existing scheme's 1.92. This application demonstrates a significant improvement in built-up edge suppression and machining quality compared to existing PID control schemes. Furthermore, this application's scheme, based on model predictive control, can adjust parameters in advance before severe built-up edge adhesion, achieving proactive control; existing PID control only provides lag feedback based on the current error. This scheme simultaneously optimizes the three variables of milling speed, feed rate, and depth of cut, resulting in smoother parameter adjustments and effectively balancing built-up edge suppression and machining efficiency. Its overall performance is significantly superior to existing schemes, making it industrially applicable.

[0087] Those skilled in the art will recognize that the units of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the invention.

[0088] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical functional division. In actual implementation, there may be other division methods, such as multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored.

[0089] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0090] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the specification of the present invention.

Claims

1. A method for identifying and controlling built-up edge adhesion in aluminum machining processes, characterized in that, Includes the following steps: The machining signals are collected from the contact area between the tool and the workpiece during the machining of aluminum by the machine tool; the machining signals include acoustic emission signals, tool-workpiece contact resistance signals, and cutting temperature signals. Based on the acoustic emission signal, contact resistance signal, and cutting temperature signal, the root mean square value of the acoustic emission signal, the fluctuation amplitude of the contact resistance signal, and the temperature rise rate of the cutting temperature signal are calculated respectively, and the built-up edge adhesion index is calculated; wherein, the built-up edge adhesion index is a continuous value used to characterize the degree of adhesion of the built-up edge. The pylomast adhesion index is input as a feedback signal to the model prediction controller, which performs the following operations sequentially in each control cycle: Using the current machining state as the initial state, the root mean square value, fluctuation amplitude, and temperature rise rate are taken as observed values ​​of the state variables, and the built-up edge adhesion index is taken as the output feedback. The predictive control model is used to predict the evolution trend of the built-up edge adhesion index in the future prediction time domain. In the state equation of the model, the state variables include the characteristic values ​​of the temperature field and the characteristic values ​​of the stress field in the cutting zone, and the output variable is the built-up edge adhesion index. The optimal control sequence is obtained by taking the minimum sum of squares of the cumulative buildup adhesion index in the prediction time domain as the objective function, and combining the rate of change constraint of the buildup adhesion index and the range of change constraint of the processing parameters. The control variable value corresponding to the first moment in the optimal control sequence is output to the machine tool actuator to adjust the machining parameters at the next moment.

2. The method for identifying and controlling built-up edge adhesion in aluminum machining processes according to claim 1, characterized in that, The adhesion index of the plastic tumor The calculation formula is: in, is the root mean square value of the acoustic emission signal; The root mean square reference value of acoustic emission in the non-adhesive state; This is the reference value for contact resistance in a non-adhesive state; This represents the fluctuation amplitude of the contact resistance signal; The temperature rise rate of the cutting temperature signal; This is the baseline value for the rate of temperature rise under non-adhesive conditions. , , Preset weighting coefficients and satisfying ; , , >1 indicates the nonlinear amplification index.

3. The method for identifying and controlling built-up edge adhesion in aluminum machining processes according to claim 2, characterized in that, Before calculating the root mean square value of the acoustic emission signal, the following steps are also included: Multi-level wavelet packet decomposition is performed on the acoustic emission signal to decompose it into multiple frequency band components with different frequency ranges. The proportion of the energy of each frequency band component to the total energy of the acoustic emission signal is calculated. When the energy proportion of the frequency band components obtained from the decomposition exceeds a preset threshold, the contribution weight of the acoustic emission signal is adjusted. The frequency band energy ratio is dynamically adjusted based on the wavelet packet decomposition result. The adjustment formula is as follows: in, For frequency band energy; Total energy; The preset enhancement coefficient; the adjusted weights , , Renormalization to meet .

4. The method for identifying and controlling built-up edge adhesion in aluminum machining processes according to claim 1, characterized in that, The objective function is: in, The objective function value; To predict the length of the time domain; For the first Time information for the first Predicted value of the adhesion index of drupture at any time; For the first Time information for the first The predicted value of the control variable is controlled at all times; For the first Time information for the first The predicted value of the control variable is controlled at all times; This is the adhesion index penalty coefficient; The penalty coefficient for changes in the control quantity; It is the square of the Euclidean norm, which is the sum of the squares of the changes in each control variable.

5. The method for identifying and controlling built-up edge adhesion in aluminum machining processes according to claim 1, characterized in that, The specific steps of using the model to predict the evolution trend of the adhesion index of the pyroplastic tumor in the future prediction time domain include: Multiple prediction sub-models corresponding to different pyroplastic adhesion states are pre-established. In each control cycle, the corresponding sub-model is selected from the multiple prediction sub-models for prediction based on the current value and trend of the pyroplastic adhesion index. Within the state transition interval corresponding to different sub-models, the prediction results of adjacent sub-models are weighted and fused.

6. The method for identifying and controlling built-up edge adhesion in aluminum machining processes according to claim 5, characterized in that, The multiple prediction sub-models are established in the following way: Historical data on the adhesion index of built-up edge during processing at different numerical ranges were collected, and the model parameters of each sub-model were determined by a system identification method. The system identification method includes any one of the following: least squares method, maximum likelihood method, or subspace identification method.

7. The method for identifying and controlling built-up edge adhesion in aluminum machining processes according to claim 1, characterized in that, When using a model to predict the evolution trend of the adhesion index of pyroplasia in the future prediction time domain, it also includes: The tooth passage frequency is calculated by the spindle speed and the number of tool teeth, and a sinusoidal periodic signal of the tooth passage frequency and its at least one harmonic is constructed as a known disturbance signal. The known disturbance signal is introduced as an additional input term into the state equation of the predictive control model.

8. The method for identifying and controlling built-up edge adhesion in aluminum machining processes according to claim 1, characterized in that, The adjustment of the processing parameters at the next moment specifically includes: The adjustment amount of the processing parameters is coordinated and transformed according to the preset proportional coefficient. The proportional coefficient is adjusted according to the current chip accumulation value, and the processing parameters at the next moment are calculated according to the adjusted proportional coefficient. The machining parameters include spindle speed or cutting line speed, feed rate or feed per revolution, and depth of cut or depth of cut.

9. The method for identifying and controlling built-up edge adhesion in aluminum machining processes according to claim 1, characterized in that, The root mean square value of the acoustic emission signal The calculation is performed using a sliding window recursive algorithm, and the sliding window length is... With spindle speed satisfy ,in, The sampling frequency of the acoustic emission signal. It is a preset integer.

10. A tool built-up edge adhesion identification and control system in aluminum machining process, characterized in that, The method described by any one of claims 1-9 comprises: The signal acquisition module is used to acquire machining signals in the contact area between the tool and the workpiece during the machining of aluminum by the machine tool; the machining signals include acoustic emission signals, tool-workpiece contact resistance signals, and cutting temperature signals. The adhesion index calculation module is used to calculate the root mean square value of the acoustic emission signal, the fluctuation amplitude of the contact resistance signal, and the temperature rise rate of the cutting temperature signal based on the acoustic emission signal, the contact resistance signal, and the cutting temperature signal, and to calculate the built-up edge adhesion index; wherein, the built-up edge adhesion index is a continuous value used to characterize the degree of adhesion of the built-up edge. The model prediction control module is used to input the tumour adhesion index as a feedback signal to the model prediction controller, and the model prediction controller performs the following operations sequentially in each control cycle: Using the current machining state as the initial state, the root mean square value, fluctuation amplitude, and temperature rise rate are taken as the observed values ​​of the state variables, and the built-up edge adhesion index is taken as the output feedback. The predictive control model is used to predict the evolution trend of the built-up edge adhesion index in the future prediction time domain. In the state equation of the model, the state variables include the characteristic values ​​of the temperature field and the characteristic values ​​of the stress field in the cutting zone, and the output variable is the built-up edge adhesion index. The optimal control sequence is obtained by taking the minimum sum of squares of the cumulative buildup adhesion index in the prediction time domain as the objective function, and combining the rate of change constraint of the buildup adhesion index and the range of change constraint of the processing parameters. The control variable value corresponding to the first moment in the optimal control sequence is output to the machine tool actuator to adjust the machining parameters at the next moment.

Images