An error compensation method for composite component machining trajectory control

Abstract

The present application relates to the field of industrial control and numerical control processing technology, and provides an error compensation method for composite component processing track control. A three-dimensional coordinate set of a theoretical track, a local fiber direction vector, a feed speed vector, a transient active electric parameter, a track geometric feature and a three-axis original cutting force vector are obtained. A spatial anisotropic projection operator, a modulus retention parameter and a size effect resistance operator are extracted respectively. When the time gradient of the force vector module length is less than zero and the modulus retention parameter is less than a preset softening threshold, an uncertainty penalty weight is generated and an effective stress equivalent is obtained. Based on the effective stress equivalent, interference calculation is performed to generate a deformation compensation scalar. The deformation compensation scalar is introduced into a basic control Jacobian matrix as an amplitude gain factor to generate a dynamic compensation matrix, coordinate transformation is performed, and an actual processing coordinate set is output as an interpolation instruction.

Description

Technical Field

[0001] This invention relates to the field of industrial control and CNC machining technology, and in particular to an error compensation method for machining trajectory control of composite components. Background Technology

[0002] In current technologies for composite component machining trajectory control, cutting error compensation typically treats the composite material as an isotropic homogeneous body and relies on a single mechanical sensing feedback mechanism. However, in the microscopic cutting dimension, existing control logic fails to establish a mapping relationship between fiber arrangement and cutting feedback. Due to the asymmetry in material elastic yield caused by the cutting angle, a single mechanical sensing feedback mechanism cannot identify the deformation differences caused by spatial orthogonal anisotropy.

[0003] In terms of time and thermodynamic evolution, the composite matrix accumulates heat energy due to continuous work, and the cutting interface temperature crosses the glass transition temperature, causing the matrix to soften and yield. This softening and yielding leads to tool displacement at the machining interface, resulting in a decrease in cutting resistance. A single mechanical sensing feedback mechanism misinterprets this decrease in cutting resistance caused by material softening as normal unloading of the cutting load. This single mechanical sensing feedback mechanism outputs a non-objective low-stress state signal, causing thermodynamic distortion in the mechanical sensing of the error compensation module.

[0004] In the kinematics and geometry dimensions of the trajectory, there is a physical repulsion between the underlying kinematic planning of the CNC system and the microscopic cutting mechanics of the material. When the tool travels to a trajectory segment with large curvature geometry and triggers a feed rate reduction command, the cutting thickness decreases. This decrease in cutting thickness induces a dimensional effect, leading to a nonlinear surge in the unit cutting resistance required to push the material. Existing CNC systems cannot predict the physical repulsion force induced by the microscopic dimensional effect in rate reduction control, and distorted perception data can further cause physical interference and machining trajectory deviations.

[0005] The aforementioned spatial interference, thermodynamic interference, and geometric interference superimpose each other during the processing of composite components, making it impossible for existing error compensation technologies to obtain accurate and objective compensation benchmarks under conditions of overlapping multi-physics fields. Summary of the Invention

[0006] This invention provides an error compensation method for machining trajectory control of composite components, which solves the technical problem that existing error compensation mechanisms treat composite materials as isotropic homogeneous bodies and rely on a single mechanical sensing feedback. This leads to distortion of the underlying mechanical sensing when the composite matrix undergoes thermal softening, which in turn causes the distorted mechanical sensing data to trigger incorrect trajectory compensation commands under the large curvature trajectory deceleration size effect, ultimately resulting in microscopic physical interference and machining trajectory deviation.

[0007] In view of the above problems, the present invention provides an error compensation method for machining trajectory control of composite components, comprising the following steps: Obtain the three-dimensional coordinate set of the theoretical trajectory of the composite component, the local fiber orientation vector of the current control cycle, and the feed speed vector; Obtain the preset compensation amplification factor and the preset sensitivity factor; Obtain transient active electrical parameters and perform time discrete integration calculation on the transient active electrical parameters to generate a work integral scalar; Obtain the trajectory geometry features of the preset look-ahead trajectory segment, as well as the three-axis original cutting force vector of the current control cycle; Based on the local fiber orientation vector and the feed velocity vector, calculate the spatial anisotropic projection operator; Based on the work integral scalar, the corresponding modulus retention parameter is extracted through a preset mapping relationship; Based on the trajectory geometry and cutting kinematics features, a size effect resistance operator is extracted. When the time gradient of the modulus of the original triaxial cutting force vector is less than 0 and the modulus retention rate parameter is less than the preset softening threshold, the feature conflict condition is determined to be met, and the uncertainty penalty weight is generated by combining the preset sensitivity coefficient and the preset compensation amplification coefficient. The effective force equivalent is obtained by multiplying the magnitude of the original triaxial cutting force vector with the uncertainty penalty weight. Based on the effective force equivalent, the spatial anisotropic projection operator, and the size effect resistance operator, an interference solution is performed to generate a deformation compensation scalar. The deformation compensation scalar is introduced as an amplitude gain factor into a preset basic control Jacobian matrix to generate a dynamic compensation matrix. The dynamic compensation matrix is ​​used to transform the three-dimensional coordinate set of the theoretical trajectory to generate the actual machining coordinate set, and the actual machining coordinate set is output as a trajectory interpolation command.

[0008] This invention provides an error compensation system for machining trajectory control of composite components, comprising a memory and a processor. The memory stores computer instructions, and the processor executes the computer instructions to implement the aforementioned error compensation method for machining trajectory control of composite components.

[0009] The technical solution provided in this application has at least the following technical effects: By extracting the absolute value of the spatial inner product of the local fiber orientation vector and the feed velocity vector, a spatial anisotropic projection operator is extracted. This compensates for the microscopic mechanical deformation deviation caused by the forward and reverse arrangement of composite components, and improves the accuracy of physical feature extraction in the spatial dimension of the error compensation method.

[0010] By acquiring transient active electrical parameters and performing discrete integral calculations to generate a work integral scalar, and combining the time gradient of the modulus of the triaxial original cutting force vector to generate uncertainty penalty weights, the error of the decrease in cutting resistance caused by the composite matrix crossing the glass transition temperature is eliminated, avoiding the distortion risk of a single mechanical sensing feedback mechanism misjudging material softening as load unloading.

[0011] By using the current cutting thickness to extract the size effect resistance operator when the curvature derivative is greater than the preset dimensionality reduction threshold to trigger the feed deceleration state, the surge in unit cutting resistance caused by the thinning of the cutting thickness is compensated, and the tool overcut and machining physical interference at the large curvature trajectory position are eliminated.

[0012] By interferometrically solving the effective force equivalent, spatial anisotropic projection operator and size effect resistance operator to generate a deformation compensation scalar, and introducing the deformation compensation scalar into the preset basic control Jacobian matrix to generate a dynamic compensation matrix, the fusion and cancellation of spatial features, thermodynamic features and kinematic features are realized, and the actual machining coordinate set after multi-physics error correction is directly output, ensuring the accuracy of machine tool dynamic trajectory execution. Attached Figure Description

[0013] Figure 1 A flowchart of an error compensation method for machining trajectory control of composite components provided in an embodiment of the present invention; Figure 2 This is a diagram illustrating an error compensation system architecture for controlling the machining trajectory of composite components, provided as an embodiment of the present invention. Detailed Implementation

[0014] The above technical solutions will now be described in detail with reference to the accompanying drawings and specific embodiments to provide a better understanding of them. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments used only to explain the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. Furthermore, it should be noted that, for ease of description, only the parts related to the present invention are shown in the drawings, not all of them.

[0015] Example: Please refer to Figure 1 and Figure 2The independent edge computing gateway establishes a physical communication link with the underlying execution terminal and spindle driver. The independent edge computing gateway integrates a bypass interception module and a direct memory access interface. Within the standard interpolation control cycle of the underlying execution terminal, the independent edge computing gateway intercepts the theoretical trajectory 3D coordinate set of the composite component and the operating status of the underlying execution terminal through the physical communication link, and writes the actual machining coordinate set into the interpolation register of the underlying execution terminal before the end of the control cycle. The actual machining coordinate set is then converted into machining servo actions within the underlying execution terminal.

[0016] Triggered by the timing clock of the control cycle, the independent edge computing gateway acquires the theoretical trajectory three-dimensional coordinate set of the composite component, the local fiber orientation vector of the current control cycle, and the feed speed vector. The independent edge computing gateway extracts the transient active electrical parameters fed back from the spindle driver via the physical communication link. These transient active electrical parameters are input to a discrete-time integrator, which performs time-discrete integration calculations on them to generate a work integral scalar. Simultaneously, the independent edge computing gateway scans a preset look-ahead trajectory segment, extracts the trajectory geometry features of the look-ahead trajectory segment, and synchronously acquires the original three-axis cutting force vectors of the current control cycle.

[0017] After the trajectory geometry features, the three-axis original cutting force vector, the local fiber orientation vector, and the work integral scalar are aggregated in parallel into the memory space of the independent edge computing gateway, the gateway initiates a multiphysics decoupling and interferometric solution mechanism. Based on the local fiber orientation vector and feed velocity vector, the gateway extracts the spatial anisotropic projection operator. Based on the work integral scalar, it extracts the corresponding modulus retention parameter through a preset mapping relationship. Based on the trajectory geometry features and cutting kinematics features, the gateway extracts the size effect resistance operator.

[0018] After feature extraction is completed, the independent edge computing gateway performs temporal difference operations on the original three-axis cutting force vector to calculate the temporal gradient of its magnitude. When the temporal gradient of the original three-axis cutting force vector's magnitude is less than 0 and the modulus retention rate parameter is less than a preset softening threshold, the independent edge computing gateway determines that the current machining interface meets the feature conflict condition and triggers the generation of uncertainty penalty weights. The magnitude of the original three-axis cutting force vector and the uncertainty penalty weights are multiplied in a multiplier to output the effective force equivalent.

[0019] The output effective force equivalent, spatial anisotropic projection operator, and size effect resistance operator are jointly input into the interferometric solver. The interferometric solver performs interferometric calculations based on these parameters, generating a deformation compensation scalar. This scalar is then introduced into a preset basic control Jacobian matrix, driving an algebraic update to generate a dynamic compensation matrix. The independent edge computing gateway uses this dynamic compensation matrix to perform coordinate transformation on the theoretical trajectory 3D coordinate set of the composite component, generating an actual processing coordinate set. This actual processing coordinate set is packaged into trajectory interpolation instructions, which are then sent by the independent edge computing gateway along the physical communication link to the underlying execution terminal.

[0020] As a computational branch for extracting the spatial anisotropic projection operator, the independent edge computing gateway performs a spatial inner product operation on the local fiber orientation vector and the feed velocity vector in a three-dimensional Cartesian coordinate system. The absolute value of the spatial inner product of the local fiber orientation vector and the feed velocity vector is extracted by the mathematical coprocessor of the independent edge computing gateway. The absolute value of the spatial inner product is determined by the independent edge computing gateway as the spatial anisotropic projection operator. Under the condition of introducing a preset first scaling factor, the absolute value of the spatial inner product is multiplied by the preset first scaling factor, and the product value obtained by multiplying the absolute value of the spatial inner product by the preset first scaling factor is determined as the spatial anisotropic projection operator.

[0021] The absolute value of the spatial inner product between the local fiber orientation vector and the feed velocity vector establishes a physical mapping between the microscopic geometric topology and the macroscopic material stiffness. The absolute value of the spatial inner product transforms the directional characteristics of the elastic tensor into scalar constraints. It characterizes the component projection of the orthotropic stiffness of the composite component along the feed velocity vector direction. The spatial anisotropic projection operator, as a scalar constraint, directly participates in the interferometric solution.

[0022] With polling of the underlying control interface, the independent edge computing gateway acquires the current feedback signal of the spindle driver in discrete time frames to obtain transient active power parameters. These transient active power parameters are continuously pushed into a time integration register deployed within the independent edge computing gateway. Within each interpolation time window, the time integration register performs time-discrete integration calculations on the transient active power parameters. The discrete integration operation converts the current waveform in the time dimension into an energy accumulation value, and the accumulated result of the time-discrete integration calculation is formatted in the time integration register to generate a work integral scalar.

[0023] The work integral scalar corresponds physically to the cumulative heat loss generated at the machining interface. This accumulated heat causes the temperature of the polymer matrix in the composite component to rise towards its glass transition temperature (GLT). Once the temperature in the cutting region exceeds the GLT, the polymer matrix undergoes a high-elasticity phase transition. This transition causes a decrease in the macroscopic mechanical stiffness of the composite component. To quantify the change in flexibility at the computational level, a pre-defined mapping matrix is ​​established, containing the correspondence between the cumulative electrical work value and the modulus retention rate. This pre-defined mapping matrix is ​​calibrated using a constant-temperature thermomechanical analyzer, recording the softening curve gradient of the composite component material from room temperature to the GLT. The independent edge computing gateway inputs the work integral scalar as a search index into the pre-defined mapping matrix. The pre-defined mapping matrix performs an index lookup and table matching to extract the corresponding value as the modulus retention rate parameter. The output modulus retention rate parameter is dimensionless, establishing a quantitative correspondence between the actuator electric field data and the thermodynamic state of the material.

[0024] In the parallel processing layer of the trajectory kinematics control logic, the independent edge computing gateway scans the look-ahead trajectory segment through a pre-read buffer queue. The independent edge computing gateway extracts the curvature derivative of the look-ahead trajectory segment as the trajectory's geometric feature. The comparator module built into the independent edge computing gateway performs real-time Boolean logic comparison between the extracted curvature derivative and a preset dimensionality reduction threshold. When the curvature derivative is greater than the preset dimensionality reduction threshold, the Boolean logic comparison outputs a high-level toggling signal. This high-level toggling signal triggers the underlying operating architecture to enter a feed deceleration state and drives the independent edge computing gateway's instruction generator to synchronously generate feed deceleration commands.

[0025] The feed rate reduction command causes a reduction in the tool's depth of cut per unit time. The independent edge computing gateway calculates the current cutting thickness based on the constraint parameters of the feed rate reduction command and the current spindle speed parameter. This current cutting thickness is used as a cutting kinematic feature in the interference solution. The reduction in the spatial dimension of the current cutting thickness causes the physical contact area between the tool edge and the composite component matrix to approximate the geometric dimensions of the tool tip blunt radius. As the current cutting thickness decreases, the mechanical boundary evolves from shear fracture behavior to extrusive frictional deformation behavior. This extrusive frictional deformation behavior leads to an exponential increase in the unit cutting resistance required to push the material.

[0026] The current cutting thickness is substituted into a preset single-edge cutting thickness attenuation model for calculation. This model incorporates an exponential operator characterizing the rise of the friction curve in the blunt region. Based on the substituted current cutting thickness, the model performs mathematical calculations to obtain the resistance mutation increment. The output resistance mutation increment corresponds in physical dimensions to the increased local resistance vector magnitude in the extrusion friction zone. The independent edge computing gateway determines the calculated resistance mutation increment as a size effect resistance operator and pushes it into the memory buffer stack as input parameters for the interference solution.

[0027] After the multiphysics decoupling operation is completed, the independent edge computing gateway performs a first-order time-difference operation on the magnitude of the original three-axis cutting force vector. The first-order difference operation outputs the temporal gradient of the original three-axis cutting force vector's magnitude. The independent edge computing gateway determines whether the temporal gradient of the original three-axis cutting force vector's magnitude is less than 0. When the feature conflict condition is met, the independent edge computing gateway extracts the absolute value of the temporal gradient of the original three-axis cutting force vector's magnitude. The absolute value of the temporal gradient of the original three-axis cutting force vector's magnitude is assigned by the independent edge computing gateway and defined as the descent gradient. Simultaneously, the independent edge computing gateway calls constant 1 and calculates the difference between constant 1 and the modulus retention parameter. The difference between constant 1 and the modulus retention parameter is assigned as the imbalance parameter. The descent gradient and the imbalance parameter are sent to the hardware multiplication register by the independent edge computing gateway. The hardware multiplication register multiplies the descent gradient and the imbalance parameter to obtain the trend back dispersion. The quantized value of the trend back dispersion reflects the feature conflict state of the accumulation of the descent gradient and the increase of the imbalance parameter.

[0028] The trend back dispersion is transmitted to subsequent computing nodes via the underlying data bus. The trend back dispersion is multiplied by a preset sensitivity coefficient to obtain the first product. This first product is input into a preset convergence function, which performs a spatial transformation to obtain the basic mapping value. The basic mapping value is multiplied by a preset compensation amplification factor, and then a constant 1 is added to calculate the uncertainty penalty weight. The magnitude of the original three-axis cutting force vector and the uncertainty penalty weight are multiplied in the arithmetic logic unit of the independent edge computing gateway. The magnitude of the original three-axis cutting force vector multiplied by the uncertainty penalty weight yields the effective force equivalent. The independent edge computing gateway multiplies the effective force equivalent, the preset system basic compliance constant, the spatial anisotropy projection operator, and the size effect resistance operator to output the first composite parameter. When processing division operations, the independent edge computing gateway calls the underlying maximum value function. The maximum value function performs a comparison operation between the modulus retention rate parameter and the preset safety minimum value constant. The maximum value function extracts the larger value between the modulus retention rate parameter and the preset safety minimum constant, defining it as the cutoff retention rate parameter. The first composite parameter is then divided by the cutoff retention rate parameter, and the division operation node solves to obtain the deformation compensation scalar. The entire calculation process for obtaining the deformation compensation scalar strictly follows the deformation constraint formula, which is expressed as the ratio between the first composite parameter and the cutoff retention rate parameter.

[0029] The deformation compensation scalar is obtained by solving the deformation constraint formula: in, For deformation compensation scalar, This represents the original cutting force vector of the three axes. The modulus of the original triaxial cutting force vector. As the uncertainty penalty weight, The preset system basic compliance constant, For spatial anisotropic projection operators, For size effect resistance operator, The modulus retention parameter, This is a preset safety minimum constant. This is the function for finding the maximum value.

[0030] In the algebraic structure of the deformation constraint formula, the modulus of the triaxial original cutting force vector has the nominal dimension of Newtonian mechanics. The uncertainty penalty weight is a dimensionless mathematical scaling factor. The product of the modulus of the triaxial original cutting force vector and the uncertainty penalty weight maintains the nominal dimension of Newtonian mechanics. The preset system fundamental compliance constant has the dimension of mechanical displacement strain rate per millimeter per Newton. The spatial anisotropy projection operator and the size effect resistance operator are both dimensionless parameters extracted through inner product operation and exponential decay operation. The modulus retention rate parameter distributed in the denominator of the division and the preset safety minimum constant together constitute a dimensionless numerical envelope. Substituting the parameters into the deformation constraint formula for multiplication and division operations, the Newtonian mechanical dimension and the strain rate dimension per millimeter per Newton cancel each other out. The final calculation result after cancellation presents the nominal dimension of displacement in millimeters. The nominal dimension of displacement matches the data type requirements of the position feedforward control command of the underlying execution terminal.

[0031] Before the deformation compensation scalar is incorporated into the online real-time control loop, the preset system fundamental compliance constant, preset compensation amplification factor, and preset sensitivity factor are determined by offline physical calibration to establish their value boundaries. The preset system fundamental compliance constant is obtained through a static test calibration method of the machine tool foundation structure. The calibration environment for the preset system fundamental compliance constant is set at a constant temperature of 20℃. Under the constant temperature of 20℃, the calibration instrument applies a stepped increasing static cutting load to the machine tool spindle, and the measuring instrument simultaneously measures and records the radial inherent deformation rate of the tool tip induced by the corresponding static cutting load. The linear regression slope formed by fitting the radial inherent deformation rate of the tool tip to the static cutting load is extracted and stored in the non-volatile memory of the independent edge computing gateway as the preset system fundamental compliance constant.

[0032] The preset compensation amplification factor corresponds to the machine tool spindle's ultimate stiffness redundancy. The sensing test matrix acquires the overload torque threshold of the machine tool spindle motor and the yield strength parameters of the machine tool's cast iron bed material. The overload torque threshold of the spindle motor and the yield strength parameters of the machine tool's cast iron bed material are converted into dimensionless safety margin scalars via a structural mechanics conversion module. These dimensionless safety margin scalars are assigned to the preset compensation amplification factor. The preset compensation amplification factor limits the maximum value of the uncertainty penalty weight. The preset sensitivity coefficient determines the response period for entering the saturation region of the preset convergence function. The value of the preset sensitivity coefficient is calibrated based on the frequency of matrix fiber fracture in different composite components, and the preset sensitivity coefficient is written into the system configuration file of the underlying driver and loaded during power-on initialization.

[0033] Under the condition of sensor communication link failure, the failure of the physical hardware of the mechanical sensor triggers a decrease in the magnitude of the triaxial original cutting force vector. The absolute value of the time gradient of the magnitude of the triaxial original cutting force vector undergoes a step change and approaches the extreme value of the data acquisition of the underlying execution terminal. As the absolute value of the time gradient of the magnitude of the triaxial original cutting force vector approaches the extreme value of the data acquisition of the underlying execution terminal, the decreasing gradient drives the first product to tend towards positive infinity. When the first product, which tends towards positive infinity, is input into the preset convergence function, the output of the preset convergence function remains at the base mapping value of constant 1. The base mapping value is multiplied by the preset compensation amplification factor and then added to the constant 1 to calculate the uncertainty penalty weight. The maximum value of the uncertainty penalty weight is fixed within the sum of the constant 1 and the preset compensation amplification factor. When the uncertainty penalty weight is multiplied by the magnitude of the triaxial original cutting force vector to calculate the effective force equivalent, the trajectory interpolation command value of the underlying execution terminal remains within the travel tolerance range of the physical machine tool.

[0034] The continuous accumulation of thermodynamic energy in the composite component processing area causes the polymer matrix to soften and melt. This softening and melting causes the extracted modulus retention parameter to decay and approach zero. A maximum value function compares the modulus retention parameter approaching zero with a preset safety minimum constant. The maximum value function then truncates the preset safety minimum constant as the output truncated retention parameter. The modulus retention parameter approaching zero is isolated from the division operation bus by the maximum value function. The calculation operation of dividing the first composite parameter by the truncated retention parameter is converted into dividing the first composite parameter by the preset safety minimum constant. The underlying execution terminal maintains a normal periodic control cycle.

[0035] For the spatial transformation of the preset convergence function, the log-odds mapping model is introduced into the replacement instruction set of the independent edge computing gateway. The log-odds mapping model constructs the mapping relationship based on the logistic function of the natural base. After the first product is received by the log-odds mapping model, it extracts the negative form of the first product and uses it as the exponential component of the natural base to generate the decaying exponent term. Constant 1 is added to the decaying exponent term to construct the denominator feature of the log-odds mapping model. Constant 2 is divided by the denominator feature of the log-odds mapping model, and after subtracting constant 1, the output benchmark is generated. The output benchmark is configured as the base mapping value. The log-odds mapping model calculates the base mapping value.

[0036] The standalone edge computing gateway incorporates a lookup table-based architecture. A one-dimensional discrete data array is pre-programmed into the gateway's read-only memory. The addressing index of this array is mapped to a discrete scalar of the trend back dispersion, and the contents of the memory cells store the pre-calculated values ​​of this mapping. Within a control cycle, the trend back dispersion is converted into the memory address offset of the one-dimensional discrete data array. When an address mismatch results in a truncation error, the gateway invokes the underlying arithmetic logic unit to perform a two-point linear interpolation operation. This operation uses the stored values ​​of two adjacent memory addresses to calculate the base mapping value for the current control cycle. The lookup table-based architecture transforms function computation into memory reads and basic multiply-accumulate operations, outputting uncertainty penalty weights.

[0037] In a scenario where the step-level logic is deployed within the hardware logic gates of an independent edge computing gateway, the first product is input to a hardware comparator. The hardware comparator performs a single-cycle level-to-level comparison between the first product and a preset deviation threshold. When the first product is greater than the preset deviation threshold, the hardware comparator outputs a latched high-level signal, which drives the base mapping value to output a constant of 1. When the first product is equal to or less than the preset deviation threshold, the hardware comparator is pulled down to a low-level state, which drives the base mapping value to output a constant of 0. The step-level logic directly outputs a constant through a cutoff threshold that is either 0 or 1. The uncertainty penalty weight output by the step-level logic is a constant at the upper limit of the saturation boundary.

[0038] The instruction set constituting the error compensation method is physically embedded in a specific electronic device. This specific electronic device is embodied as an external edge computing gateway. The motherboard of the external edge computing gateway integrates an industrial-grade microcontroller unit and a memory array. The industrial-grade microcontroller unit integrates a digital signal processing coprocessor. When the external edge computing gateway is powered on and reset, the machine code instructions contained in the memory array, used to implement logical steps such as anisotropic projection operator extraction and deformation compensation scalar solution, are loaded into the high-speed instruction cache of the industrial-grade microcontroller unit via the bus. The high-low level transition of the industrial-grade microcontroller unit triggers the switching of semiconductor transistors inside the arithmetic logic unit, completing the electrical signal mapping and logical operations of multi-source physical field decoupling on the physical circuit.

[0039] The deformation compensation scalar output by the industrial-grade microcontroller is packaged and converted into low-level machine code messages conforming to the industrial fieldbus protocol specification by the compiler at the data link layer. These low-level machine code messages are sent to the low-level servo driver via the Ethernet physical layer chip of an independent external gateway. The low-level machine code messages bypass the high-level syntax parser of the CNC system and are injected into the position loop feedforward control register of the servo driver. Changes in the charge state of the position loop feedforward control register are converted into stator current increments in the servo motor, driving the machine tool spindle to produce physical displacement. Message interaction and register state overwriting complete the compensation algorithm calculation and the output of the machine tool spindle's physical displacement.

[0040] The control logic, mapping matrix, and compensation algorithm are compiled and solidified into a computer program product. This computer program product is non-volatilely physically recorded onto a computer-readable storage medium. The specific physical form of the computer-readable storage medium includes read-only memory, electrically erasable programmable read-only memory, static random access memory, flash memory chips, or high-speed magnetic media arrays. When the processor of the underlying execution terminal or external edge device reads and executes the physical level state sequence contained in the computer-readable storage medium, the physical level state sequence drives the device to execute data flow and register control actions for the error compensation method used for composite component machining trajectory control.

[0041] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. 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 of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An error compensation method for machining trajectory control of composite components, characterized in that, Includes the following steps: Obtain the three-dimensional coordinate set of the theoretical trajectory of the composite component, the local fiber orientation vector of the current control cycle, and the feed speed vector; Obtain the preset compensation amplification factor and the preset sensitivity factor; Obtain transient active electrical parameters and perform time discrete integration calculation on the transient active electrical parameters to generate a work integral scalar; Obtain the trajectory geometry features of the preset look-ahead trajectory segment, as well as the three-axis original cutting force vector of the current control cycle; Based on the local fiber orientation vector and the feed velocity vector, calculate the spatial anisotropic projection operator; Based on the work integral scalar, the corresponding modulus retention parameter is extracted through a preset mapping relationship; Based on the trajectory geometry and cutting kinematics features, a size effect resistance operator is extracted. When the time gradient of the modulus of the original triaxial cutting force vector is less than 0 and the modulus retention rate parameter is less than the preset softening threshold, the feature conflict condition is determined to be met, and the uncertainty penalty weight is generated by combining the preset sensitivity coefficient and the preset compensation amplification coefficient. The effective force equivalent is obtained by multiplying the magnitude of the original triaxial cutting force vector with the uncertainty penalty weight. Based on the effective force equivalent, the spatial anisotropic projection operator, and the size effect resistance operator, an interference solution is performed to generate a deformation compensation scalar. The deformation compensation scalar is introduced as an amplitude gain factor into a preset basic control Jacobian matrix to generate a dynamic compensation matrix. The dynamic compensation matrix is ​​used to transform the three-dimensional coordinate set of the theoretical trajectory to generate the actual machining coordinate set, and the actual machining coordinate set is output as a trajectory interpolation command.

2. The error compensation method for machining trajectory control of composite components according to claim 1, characterized in that: The step of calculating the spatial anisotropic projection operator based on the local fiber orientation vector and the feed velocity vector includes: calculating the absolute value of the spatial inner product of the local fiber orientation vector and the feed velocity vector, and determining the absolute value or the value obtained by multiplying the absolute value by a preset first scaling factor as the spatial anisotropic projection operator.

3. The error compensation method for machining trajectory control of composite components according to claim 1, characterized in that: The step of extracting the corresponding modulus retention rate parameter based on the work integral scalar through a preset mapping relationship includes: pre-establishing a preset mapping matrix containing the correspondence between the cumulative electrical work value and the modulus retention rate; inputting the work integral scalar as a retrieval index into the preset mapping matrix, and extracting the corresponding value as the modulus retention rate parameter, wherein the modulus retention rate parameter is a dimensionless parameter.

4. The error compensation method for machining trajectory control of composite components according to claim 1, characterized in that: The extraction of the size effect resistance operator based on the trajectory geometry features and cutting kinematic features includes: The curvature derivative of the look-ahead trajectory segment is extracted as the geometric feature of the trajectory; When the curvature derivative is greater than the preset dimensionality reduction threshold, the feed deceleration state is triggered and a feed deceleration command is generated. The current cutting thickness is calculated based on the feed rate reduction command and used as the cutting kinematic characteristic. The current cutting thickness is substituted into the preset single-edge cutting thickness attenuation model for calculation to obtain the resistance mutation increment, and the resistance mutation increment is determined as the size effect resistance operator.

5. The error compensation method for machining trajectory control of composite components according to claim 1, characterized in that, The process of generating a deformation compensation scalar based on the effective force equivalent, the spatial anisotropic projection operator, and the size effect resistance operator through interferometric solution includes the following calculation steps: Obtain the preset system basic flexibility constant and the preset safety minimum constant; When it is determined that the feature conflict condition is met, the absolute value of the time gradient of the magnitude of the original three-axis cutting force vector is extracted as the descent gradient. The difference between constant 1 and the modulus retention parameter is calculated and used as the imbalance parameter; Multiplying the descent gradient by the imbalance parameter yields the trend divergence. Multiply the trend back dispersion by the preset sensitivity coefficient to obtain a first product, and input the first product into a preset convergence function to obtain the basic mapping value; The uncertainty penalty weight is calculated by multiplying the basic mapping value by the preset compensation amplification factor and adding a constant 1. The effective force equivalent, the preset system basic flexibility constant, the spatial anisotropic projection operator, and the size effect resistance operator are multiplied together to obtain the first composite parameter. The maximum value between the modulus retention parameter and the preset safety minimum constant is used to obtain the cutoff retention parameter; Divide the first composite parameter by the cutoff retention rate parameter to obtain the deformation compensation scalar.

6. The error compensation method for machining trajectory control of composite components according to claim 5, characterized in that, The calculation of the deformation compensation scalar obtained by the solution strictly follows the following deformation constraint formula: in, Let be the deformation compensation scalar. The original cutting force vector of the three axes, The modulus of the original triaxial cutting force vector is given. The uncertainty penalty weight is... The preset system basic compliance constant is... For the spatial anisotropic projection operator, For the size effect resistance operator, The modulus retention parameter is... The preset safety minimum constant is... This is the function for finding the maximum value.

7. The error compensation method for machining trajectory control of composite components according to claim 5, characterized in that, The mapping model corresponding to the preset convergence function is any one of the following: Hyperbolic tangent mapping model: Calculate the hyperbolic tangent value of the first product as the base mapping value; Log-odds mapping model: The first product is calculated based on the Sigmoid function with the natural base, and used as the underlying mapping value; Step-level mapping model: When the first product is greater than the preset deviation threshold, the basic mapping value is output as a constant 1; otherwise, it is output as a constant 0. The preset deviation threshold is a system preset parameter.

8. An error compensation system for machining trajectory control of composite components, characterized in that, It includes a memory and a processor, the memory being used to store computer instructions, and the processor being used to execute the computer instructions to implement an error compensation method for machining trajectory control of composite components as described in any one of claims 1 to 7.

Images