Milling force prediction method of machine tool current data driven simulation model

Abstract

The milling force prediction technology based on machine tool current data-driven simulation models belongs to the field of high-speed milling technology. It includes the following steps: First, the machine tool's feed current signal is acquired, and the component related to the milling force is extracted from the current at varying scales. The trend curve of the change is then extracted using the least squares method. Next, the machine tool's cutting parameters are obtained through OPC UA to determine the current cutting type and generate a milling force simulation waveform. Finally, the milling force simulation waveform, the trend curve of the current characteristic signal, and the ratio coefficient of the feed current to the milling force are multiplied to calculate the instantaneous milling force signal. This invention is the first to combine the milling force simulation waveform with the current change trend, achieving accurate perception of the instantaneous milling force in variable parameter cutting, and providing a novel solution for the indirect monitoring of milling forces.

Description

Technical Field

[0001] This invention belongs to the field of high-speed milling technology and relates to the problem of real-time estimation of milling force, specifically a method for predicting milling force using a simulation model driven by machine tool current data. Background Technology

[0002] Metal cutting is a process of interaction between the workpiece and the cutting tool. Material removal is primarily achieved through rotational and feed motions. The force required for material deformation and chip formation during this process is called cutting force. In milling, the cutting force originates from three sources: the resistance generated by elastic and plastic deformation during chip formation, the frictional resistance between the tool, chips, and workpiece surface, and the impact force caused by the alternating cutting action of multiple teeth. Predicting milling force quickly and accurately is crucial. Milling force is dynamically correlated with other physical parameters and is the root cause of changes in other physical quantities, directly affecting cutting temperature, tool chatter, and workpiece deformation. Common milling force measurement methods are divided into direct and indirect methods. Direct measurement involves mounting a force gauge on the workpiece or spindle to directly obtain the load signal during milling. Common force gauges can be divided into fixed and rotary force gauges, both of which are piezoelectric force gauges. They utilize the piezoelectric effect of quartz crystals to convert force signals into voltage signals, which are then connected to a data acquisition card via a signal amplifier. The installation of force gauges is subject to many restrictions. Fixed force gauges have certain requirements on the size of the workpiece and the clamping and positioning. Many large workpieces cannot be used with force gauges. Rotary force gauges, such precision instruments, are easily damaged in complex machining environments. For example, the cutting tool is prone to falling off during cycloidal milling, which can cause the piezoelectric crystal to break. Therefore, indirect measurement during milling is required.

[0003] With the advent of digital industry, data-driven models are the future direction for milling force prediction. The main method involves predicting instantaneous milling force using vibration, machine tool current, and machine tool communication data. The interaction force between the tool and workpiece transmits excitation vibration to the tool and spindle in the form of excitation force; therefore, milling force can be characterized by measuring vibration signals. However, the mapping relationship between vibration and force signals is related to the location of the test point. When the location of the test point or the tool holder changes, the vibration transmission path and frequency response function between the two change, easily causing test errors. There is a non-linear mapping relationship between machine tool current and milling force. When the mechanical load on the motor changes, the servo controller overcomes the load change by adjusting the motor current; therefore, using machine tool current to predict cutting force is scientifically feasible.

[0004] In summary, a logical feedback control relationship exists between machine tool current and milling force, and this relationship is unaffected by external interference. Therefore, machine tool current is an important measurement method for characterizing cutting force. Furthermore, current sensors are easy to install, have low equipment costs, and are well-suited for complex machining environments compared to other sensors (e.g., patents "Indirect Prediction Method for Milling Force Considering Tool Wear Based on Power Signal" (CN119036199B) and "A Bidirectional Milling Force Prediction Method and System" (CN119077436B)). However, these two methods cannot adapt to variable depth-of-cut conditions and struggle to accurately predict instantaneous milling force under variable parameter cutting conditions. To achieve dynamic high-precision prediction of instantaneous milling force, this invention studies the influence of different cutting parameters on the waveform, considers the continuous coupling relationship between adjacent cutting teeth, and establishes a milling force simulation waveform that dynamically changes with cutting parameters. A milling force prediction method based on a current data-driven simulation model is proposed, which can dynamically adjust the amplitude of the milling force through current data and motor parameters, achieving high-precision evaluation of instantaneous milling force under variable parameter cutting conditions. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention provides a milling force prediction method based on a machine tool current data-driven simulation model. This method can eliminate the dependence on force sensors and indirectly estimate the milling force signal through the current signal in the machining of variable parametric curved surfaces. It features high prediction accuracy and strong applicability.

[0006] The technical solution of this invention is as follows: A method for predicting milling force using a machine tool current data-driven simulation model includes the following steps: Step 1: Construct simulation waveforms for different milling states: For commonly used four-tooth end mills, in variable parameter milling, they can be divided into three categories according to the number of teeth involved in cutting at the same time: Category I: single-tooth milling, Category II: two-tooth milling, and Category III: multi-tooth milling.

[0007] Category I: Single-tooth milling indicates that no more than one tooth is involved in cutting simultaneously. That is, within one tool rotation cycle, only one cutting edge is in contact with the workpiece for part of the time, and the cutting edge is not in contact with the workpiece for part of the time. The formula for determining if it belongs to Category I is: (1) in, This represents the radial cutting depth. N This refers to the number of teeth on the cutting tool. This represents the axial cutting depth. The helix angle of the cutting tool. D The diameter of the cutting tool.

[0008] The simulation signal for single-tooth milling consists of multiple signals: a periodic simulation signal, a direction signal, and a tool eccentricity signal. The periodic simulation signal simulates the waveform generated by each cutting edge and calculates the interval between adjacent milling force waveforms based on the depth of cut. The direction signal simulates the axial frequency variation characteristics of the milling force. The tool eccentricity signal is generated by combining the tool eccentricity, tool diameter, and radial depth of cut to simulate the effect of tool eccentricity on the cutting amount.

[0009] The waveform of an isoperiodic signal consists of multiple discrete local sine waves, used to simulate the cutting process of different cutting edges. Between adjacent waveforms, there is a segment of zero amplitude signal, simulating the idle angle when the cutting edge is not involved in the cutting. The isoperiodic component can be represented as: (2) in, spindle frequency n Multiple values.

[0010] The directional component is used to switch the alternating directions of the milling force. The simulation formula for the directional component can be expressed as: (3) Tool eccentricity causes uneven cutting thickness, manifested as different cutting amounts between adjacent cutting edges. Therefore, the milling force waveform exhibits an alternating pattern of high and low values. Based on this cutting mechanism, the tooth pass frequency can be used as the frequency at which the step signal alternates. Then, the tool eccentricity component can be established by combining the effect of tool eccentricity on the cutting thickness. (4) in, The offset of the cutting tool.

[0011] Will Multiplying them yields the simulation signal for single-tooth milling. The expression is: (5) Category II: When the axial or radial cutting depth increases, the previous cutting edge is still cutting when the next cutting edge enters the tool-workpiece contact area. Both cutting edges participate in cutting simultaneously, thus single-tooth milling becomes two-tooth milling. The formula for determining if it belongs to Category II is: (6) The waveforms of two-tooth cutting and single-tooth cutting are different. When both cutting edges contact the chip simultaneously, the cutting forces are coupled, and the waveform shows mutual interference of single-tooth milling forces. The simulated force signal of two-tooth milling is divided into four parts: triangular wave component, coupling difference component, direction component, and tool eccentricity component.

[0012] The triangular wave component can be expanded into a Fourier series expression: (7) The coupling differential component is positively correlated with the angle of the contact area between the tool and the workpiece, and its mathematical expression is: (8) Simulation signal of two-tooth milling force Represented as: (9) Category III: Multi-tooth milling refers to milling where the number of cutting edges involved in the milling process is no less than two at any angle of tool rotation. The formula for determining whether it belongs to Category II is: (10) When the number of teeth simultaneously involved in cutting exceeds half the total number of teeth on the cutting tool, the continuous cutting between multiple teeth greatly reduces the impact between the cutting edge and the workpiece. In this case, the tool's eccentricity becomes the main characteristic of the milling force signal. The milling force is a characteristic signal with a periodic frequency, and its expression is: (11) The milling force simulation signals under different conditions are classified according to cutting parameters. During application, the real-time cutting depth and spindle speed are read through machine tool OPC UA communication. Then, the simulation waveform of milling force is established by determining the number of teeth participating in cutting simultaneously. The mathematical formula can be expressed as: (12) OPC UA communication of machine tools can acquire parameters such as spindle speed, depth of cut, feed rate and tool tip position. In this communication method, axial depth of cut and radial depth of cut can be acquired in real time. By using a judgmental analysis, it can be determined which type of milling depth under the current cutting parameters belongs to. Then, different simulation signals are established in combination with spindle speed to provide data support for extracting milling force trends from current.

[0013] Step 2: Extract current characteristic signals The maximum value of the Pearson correlation coefficient is used as the optimization index for variational mode decomposition. The number of decompositions is automatically selected by calculating the correlation function between the simulation signal and the low-frequency component of the current under the current cutting condition. K and penalty factor α Then, the low-frequency component after variational mode decomposition is used as the current characteristic signal.

[0014] Step 3: Calculate the ratio of current to milling force. The relationship between current and milling force is a logical feedback relationship, which can be expressed as: (13) in, HThis is the current torque coefficient; Friction torque; cutting torque Calculations based on the current during cutting and the current during the idle phase: (14) in, This indicates the current during cutting. This indicates the current in the unloaded section.

[0015] The cutting force is transmitted to the feed motor through structures such as sliding screws and nuts. The mathematical relationship between the cutting force and the cutting torque can be expressed as: (15) in, Indicates transmission efficiency. Indicates cutting force. This indicates the transmission ratio of the sliding screw in the transmission system.

[0016] Combining equations (14) and (15), we can obtain a quantitative relationship between the feed current and the milling force: (16) in, Indicates the increment of the feed current; The proportionality coefficient between the feed current and the milling force is given. Combining equations (14), (15), and (16), the calculation formula can be obtained as follows: (17) in, This indicates the pitch of the sliding leadscrew in a machine tool.

[0017] Step 4: Construct milling force signal When the cutting parameters change, the simulation signal The amplitude and waveform are dynamically changing. However, changes in the amplitude of the simulated signal are meaningless; it is necessary to analyze the simulated signal... Normalization. The Hilbert transform utilizes filters in imaginary circuits to analyze signals, converting real-valued signals into complex-valued signals. Therefore, the instantaneous envelope curve of the simulated signal can be calculated using the Hilbert transform. Then, the original signal and the envelope signal are combined to complete the normalization process. The calculation steps are as follows: For the original simulation signal Perform Hilbert transform: (18) Simulated signals As the real part, the transformed signal As the imaginary part, construct the analytic signal s ( t ).

[0018] (19) Analyzing signals s ( t The modulus of the signal is used as the simulation signal. Envelope signal: (20) Normalize the simulation signal: (twenty one) Finally, combining the normalized simulation waveforms g ( t Characteristic current variation trend and the ratio of feed current to milling force The instantaneous change in milling force is predicted, and its calculation formula is as follows: (twenty two) in, It indicates the changing trend of the characteristic current signal.

[0019] The beneficial effects of this invention are as follows: The milling force waveform is categorized based on the number of teeth simultaneously participating in cutting; simulation functions related to parameters such as spindle speed, depth of cut, and tool eccentricity are established; these functions align with the time-varying characteristics of the coupling state between adjacent teeth in variable-parameter cutting; and the waveform of the interaction force between the tool and the workpiece is dynamically simulated. The theoretical relationship and nonlinear interference between the feed current and the milling force signal are considered; the feed current signal is decomposed using a variable-scale approach based on the dynamic simulated force signal; and for the first time, the changing trend curve of the milling force is accurately extracted from the current characteristic signal. Attached Figure Description

[0020] Figure 1 This is a flowchart of the milling force prediction method for a machine tool current data-driven simulation model provided by the present invention; Figure 2 This is a simulation waveform diagram for Category I; Figure 3 This is a simulation waveform diagram for Category II; Figure 4 This is a simulation waveform diagram for Category III; Figure 5 A comparison diagram of current characteristic signals and milling force test signals; Figure 6 This is a comparison chart of the measured and estimated values ​​of milling force in side milling, where (a)~(c) correspond to Category I: single-tooth milling, Category II: two-tooth milling, and Category III: multi-tooth milling, respectively; Figure 7This is a comparison chart of the test and estimated values ​​of milling force in slotting milling, where (a)~(c) correspond to Category I: single-tooth milling, Category II: two-tooth milling, and Category III: multi-tooth milling, respectively. Detailed Implementation

[0021] The specific embodiments of the present invention are described in detail below with reference to the technical solutions and accompanying drawings.

[0022] This embodiment presents a method for predicting milling force using a machine tool current data-driven simulation model. Figure 1 A flowchart illustrating the milling force prediction method includes the following steps: Step 1: Construct simulation waveforms for different milling states: For commonly used four-tooth end mills, in variable parameter milling, they can be divided into three categories according to the number of teeth involved in cutting at the same time: Category I: single-tooth milling, Category II: two-tooth milling, and Category III: multi-tooth milling.

[0023] Category I: Single-tooth milling indicates that no more than one tooth is involved in cutting simultaneously. That is, within one tool rotation cycle, only one cutting edge is in contact with the workpiece for part of the time, and the cutting edge is not in contact with the workpiece for part of the time. The judgment formula is: (twenty three) in, This represents the radial cutting depth. N This refers to the number of teeth on the cutting tool. This represents the axial cutting depth. The helix angle of the cutting tool. D The diameter of the cutting tool.

[0024] The simulation signal for single-tooth milling consists of multiple signals: a periodic simulation signal, a direction signal, and a tool eccentricity signal. The periodic simulation signal simulates the waveform generated by each cutting edge and calculates the interval between adjacent milling force waveforms based on the depth of cut. The direction signal simulates the axial frequency variation characteristics of the milling force. The tool eccentricity signal is generated by combining the tool eccentricity, tool diameter, and radial depth of cut to simulate the effect of tool eccentricity on the cutting amount.

[0025] The waveform of an isoperiodic signal consists of multiple discrete local sine waves, used to simulate the cutting process of different cutting edges. Between adjacent waveforms, there is a segment of zero amplitude signal, simulating the idle angle when the cutting edge is not involved in the cutting. The isoperiodic component can be represented as: (twenty four) in, The rotational frequency of the main spindle.

[0026] The directional component is used to switch the alternating directions of the milling force. The simulation formula for the directional component can be expressed as: (25) Tool eccentricity causes uneven cutting thickness, manifested as different cutting amounts between adjacent cutting edges. Therefore, the milling force waveform exhibits an alternating pattern of high and low values. Based on this cutting mechanism, the tooth pass frequency can be used as the frequency at which the step signal alternates. Then, the tool eccentricity component can be established by combining the effect of tool eccentricity on the cutting thickness. (26) in, The offset of the cutting tool.

[0027] Will Multiplying them yields the simulation signal for single-tooth milling. The expression is: (27) In this embodiment, the obtained single-tooth milling simulation signal is as follows: Figure 2 As shown, it can be observed that adjacent milling force waveforms do not interfere with each other, there is a certain interval, and the interval time changes dynamically with the cutting depth.

[0028] Category II: When the axial or radial cutting depth increases, if the next cutting edge enters the tool-workpiece contact area while the previous cutting edge is still cutting, both cutting edges will participate in cutting simultaneously. In this case, single-tooth milling becomes two-tooth milling. The judgment formula is: (28) The waveforms of two-tooth cutting and single-tooth cutting are different. When both cutting edges contact the chip simultaneously, the cutting forces are coupled, and the waveform shows mutual interference of single-tooth milling forces. The simulated force signal of two-tooth milling is divided into four parts: triangular wave component, coupling difference component, direction component, and tool eccentricity component.

[0029] The triangular wave component can be expanded into a Fourier series expression: (29) The coupling differential component is positively correlated with the angle of the contact area between the tool and the workpiece, and its mathematical expression is: (30) Simulation signal of two-tooth milling force Represented as: (31) In this embodiment, the simulated signal of the two-tooth milling force is as follows: Figure 3 As shown, there is partial overlap between adjacent waveforms, and the overlap time varies with the difference between the cutter's exit angle and entry angle, thereby accurately simulating the waveform of the actual milling force.

[0030] Category III: Multi-tooth milling refers to milling where, at any angle of tool rotation, the number of cutting edges involved in the milling process is no less than two. The discriminant is: (32) When the number of teeth simultaneously involved in cutting exceeds half the total number of teeth on the cutting tool, the continuous cutting between multiple teeth greatly reduces the impact between the cutting edge and the workpiece. In this case, the tool's eccentricity becomes the main characteristic of the milling force signal. The milling force is a characteristic signal with a periodic frequency, and its expression is: (33) When a large number of teeth are involved, the impact of a single cutting tooth on the overall milling force is relatively small. Tool eccentricity or spindle eccentricity also has a relatively small impact on the milling force. In this case, the milling force is a periodic signal dominated by the shaft frequency, such as... Figure 4 As shown.

[0031] The milling force simulation signals under different conditions are classified according to cutting parameters. During application, the real-time cutting depth and spindle speed are read through machine tool OPC UA communication. Then, the simulation waveform of milling force is established by determining the number of teeth participating in cutting simultaneously. The mathematical formula can be expressed as: (34) OPC UA communication of machine tools can acquire parameters such as spindle speed, depth of cut, feed rate and tool tip position. In this communication method, axial depth of cut and radial depth of cut can be acquired in real time. By using a judgmental analysis, it can be determined which type of milling depth under the current cutting parameters belongs to. Then, different simulation signals are established in combination with spindle speed to provide data support for extracting milling force trends from current.

[0032] Step 2: Extract current characteristic signals The maximum value of the Pearson correlation coefficient is used as the optimization index for variational mode decomposition. The number of decompositions is automatically selected by calculating the correlation function between the simulation signal and the low-frequency component of the current under the current cutting condition. K and penalty factor α Then, the low-frequency component after variational mode decomposition is used as the current characteristic signal. During the blade side milling stage, dynamically changing parameters... K The characteristic signal of the extracted current was extracted and compared with the acquired milling force signal. It was found that the trend of the extracted current characteristic signal and the change of the milling force amplitude were almost consistent. Figure 5 This study verified the feasibility of extracting the milling force component from the current. For further quantitative analysis, the least squares method was used to fit the trend of the current characteristic signal. The resulting trend curve provided a basis for the amplitude variation of the milling force in real-time prediction.

[0033] Step 3: Calculate the ratio of current to milling force. The relationship between current and milling force is a logical feedback relationship, which can be expressed as: (35) Among them, the current torque coefficient H The torque coefficient of the machine tool motor used in this embodiment can be obtained by consulting the machine tool parameters. H =2.875, friction torque The cutting torque can be calculated based on the current during the cutting process and the current during the idle period. (36) in, This indicates the current during cutting. This indicates the current in the unloaded section.

[0034] The cutting force is transmitted to the feed motor through structures such as sliding screws and nuts. The mathematical relationship between the cutting force and the feed motor torque can be expressed as: (37) in, Indicates transmission efficiency. Indicates cutting force. This indicates the transmission ratio of the sliding screw in the transmission system.

[0035] Combining equations (36) and (37), we can obtain a quantitative relationship between the feed current and the milling force: (38) in, Indicates the increment of the feed current; The proportionality coefficient between the feed current and the milling force is given. Combining equations (36), (37), and (38), the calculation formula can be obtained as follows: (39) in, This indicates the pitch of the sliding leadscrew in a machine tool.

[0036] Step 4: Construct milling force signal When the cutting parameters change, the simulation signal The amplitude and waveform are dynamically changing. However, changes in the amplitude of the simulated signal are meaningless; it is necessary to analyze the simulated signal... Normalization. The Hilbert transform utilizes filters in imaginary circuits to analyze signals, converting real-valued signals into complex-valued signals. Therefore, the instantaneous envelope curve of the simulated signal can be calculated using the Hilbert transform. Then, the original signal and the envelope signal are combined to complete the normalization process. The calculation steps are as follows: For the original simulation signal Perform Hilbert transform: (40) Simulated signals As the real part, the transformed signal As the imaginary part, construct the analytic signal s ( t ).

[0037] (41) Analyzing signals s ( t The modulus of the signal is used as the simulation signal. Envelope signal: (42) Normalize the simulation signal: (43) Finally, combining the normalized simulation waveforms g ( t Characteristic current variation trend and the ratio of feed current to milling force The instantaneous change in milling force is predicted, and its calculation formula is as follows: (44) in, It indicates the changing trend of the characteristic current signal.

[0038] Randomly select from single-tooth milling, double-tooth milling, and multi-tooth milling states. Figure 5 Part of the waveform, such as Figure 6 As shown, the predicted milling force waveform and the measured milling force waveform have a good fit, and the proposed method can dynamically simulate the coupling state of the milling force according to the changes in cutting parameters. For example, during the tool entry stage ( Figure 6 In (a), the radial depth of cut increases slowly, and the cutting time of each tooth also gradually increases, so the interval between adjacent troughs gradually decreases. However, there is a certain error in the waveform between the predicted milling force and the measured milling force. This is because the milling force is affected by a variety of factors, such as chip deformation, thermal stress concentration effect, and uneven cutting fluid lubrication, all of which affect the waveform of the actual cutting force.

[0039] The entry and exit processes in the grooving milling stage include single-tooth and double-tooth milling, while all other cutting processes are multi-tooth milling. The predicted and measured milling force waveforms for each stage are shown below. Figure 7 As shown, the fit of the three stages is relatively good, with multi-tooth milling showing a better fit than single-tooth and double-tooth milling.

[0040] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are only used to illustrate the technical solutions of the present invention and should not be construed as limiting the present invention. Those skilled in the art can make modifications and substitutions to the above embodiments within the scope of the present invention without departing from the principles and spirit of the present invention.

Claims

1. A method for predicting milling force using a machine tool current data-driven simulation model, characterized in that, Includes the following steps: Step 1: Construct simulation waveforms for different milling states: For four-tooth end mills, in variable parameter milling, they are divided into three categories according to the number of teeth participating in cutting at the same time: Category I: single-tooth milling, Category II: two-tooth milling, and Category III: multi-tooth milling. Category I: The simulation signal for single-tooth milling consists of multiple signals, namely, the equal-period simulation signal, the direction signal, and the tool eccentricity signal; The periodic components are represented as: (2) in, spindle frequency n Multiple values; The simulation formula for the directional components is expressed as follows: (3) The tooth pass frequency is used as the frequency at which the step signal alternates, and then the tool eccentricity component is established by combining the effect of tool eccentricity on the cutting thickness: (4) in, The offset of the cutting tool; Will Multiplying them yields the simulation signal for single-tooth milling. The expression is: (5) Category II: The simulation force signal for two-tooth milling is divided into four parts: triangular wave component, coupling difference component, direction component, and tool eccentricity component. The triangular wave component can be expanded into a Fourier series expression: (7) The coupling differential component is positively correlated with the angle of the contact area between the tool and the workpiece, and its mathematical expression is: (8) Simulation signal of two-tooth milling force Represented as: (9) Category III: Milling force is a characteristic signal with a periodicity based on rotational frequency, and its expression is: (11) The milling force simulation signals under different conditions are classified according to cutting parameters. During application, the real-time cutting depth and spindle speed are read through the machine tool OPCUA communication. Then, the simulation waveform of the milling force is established by determining the number of teeth participating in the cutting simultaneously. The mathematical formula is expressed as follows: (12) The machine tool's OPC UA communication acquires parameters such as spindle speed, depth of cut, feed rate, and tool tip position. In this communication method, the axial and radial depth of cut are acquired in real time. The milling depth under the current cutting parameters is analyzed by a judgment formula to determine which type it is. Then, different simulation signals are established in combination with the spindle speed to provide data support for extracting the milling force trend from the current. Step 2: Extract current characteristic signals The maximum value of the Pearson correlation coefficient is used as the optimization index for variational mode decomposition. The number of decompositions is automatically selected by calculating the correlation function between the simulation signal and the low-frequency component of the current under the current cutting condition. K and penalty factor α Then, the low-frequency component after variational mode decomposition is used as the current characteristic signal; Step 3: Calculate the ratio of current to milling force. The relationship between current and milling force is a logical feedback relationship, expressed as: (13) in, H This is the current torque coefficient; Friction torque; cutting torque Calculations based on the current during cutting and the current during the idle phase: (14) in, This indicates the current during cutting. Indicates the current in the unloaded section; The mathematical relationship between cutting force and cutting torque is expressed as: (15) in, Indicates transmission efficiency. Indicates cutting force. This indicates the transmission ratio of the sliding screw in the transmission system; Combining equations (14) and (15), we obtain the quantitative relationship between feed current and milling force: (16) in, Indicates the increment of the feed current; The proportionality coefficient between the feed current and the milling force is represented by equations (14), (15), and (16), resulting in the following calculation formula: (17) in, Indicates the pitch of the sliding leadscrew in a machine tool; Step 4: Construct milling force signal When the cutting parameters change, the simulation signal The amplitude and waveform are dynamically changing; the instantaneous envelope curve of the simulated signal can be calculated through Hilbert transform, and then the original signal and the envelope signal are combined to complete the normalization process.

2. The milling force prediction method for a machine tool current data-driven simulation model according to claim 1, characterized in that, Category I: Single-tooth milling indicates that no more than one tooth participates in cutting simultaneously. That is, within one tool rotation cycle, only one cutting edge is in contact with the workpiece for part of the time, and the cutting edge is not in contact with the workpiece for part of the time. The formula for determining if it belongs to Category I is: (1) in, This represents the radial cutting depth. N This refers to the number of teeth on the cutting tool. This represents the axial cutting depth. The helix angle of the cutting tool. D The diameter of the cutting tool; Category II: When the axial or radial cutting depth increases, the previous cutting edge is still cutting when the next cutting edge enters the tool-workpiece contact area, and both cutting edges participate in cutting simultaneously. In this case, single-tooth milling becomes two-tooth milling; the formula for determining whether it belongs to Category II is: (6) Category III: Multi-tooth milling refers to milling where the number of cutting edges involved in the milling process is no less than two at any angle of tool rotation; the formula for determining whether it belongs to Category II is: (10)。 3. The milling force prediction method for a machine tool current data-driven simulation model according to claim 1, characterized in that, The specific calculation steps for step 4 are as follows: For the original simulation signal Perform Hilbert transform: (18) Simulated signals As the real part, the transformed signal As the imaginary part, construct the analytic signal s ( t ); (19) Analyzing signals s ( t The modulus of the signal is used as the simulation signal. Envelope signal: (20) Normalize the simulation signal: (21) Finally, combining the normalized simulation waveforms g ( t Characteristic current variation trend and the ratio of feed current to milling force The instantaneous change in milling force is predicted, and its calculation formula is as follows: (22) in, It indicates the changing trend of the characteristic current signal.

Images