Intelligent multi-axis error compensation control system and method for aluminum material machining

Abstract

The invention discloses an intelligent multi-axis error compensation control system and method for aluminum machining, relates to the technical field of numerical control machining precision control, and deconstructs a current signal sequence of a multi-axis servo system into low-frequency, intermediate-frequency and high-frequency components reflecting heat, force and material characteristics by utilizing a variational mode decomposition algorithm through synchronously collecting the current signal sequence. Identifying a material according to the power spectrum fingerprint of the high-frequency component and correcting a material rigidity characteristic matrix; and a material confidence coefficient is utilized to adjust a thermal displacement weight, and a thermal softening gain coefficient based on a Sigmoid function is combined to regulate a nonlinear deformation compensation amount. A spatial position compensation vector is output through weighted synthesis, the problem of nonlinear mismatch caused by heat flow evolution triggered by cutting energy consumption and load fluctuation is solved, and a consistent mapping relation is maintained between a compensation instruction and the evolution process of an actual physical field.

Description

Technical Field

[0001] This invention relates to the field of CNC machining accuracy control technology, and in particular to an intelligent multi-axis error compensation control system and method for aluminum processing. Background Technology

[0002] In multi-axis machining of aluminum alloys, the electromagnetic current sequence of the feed axis servo system contains dynamic evolution information on cutting energy consumption, mechanical load, and material properties. Existing error compensation techniques use pre-defined analytical models to correlate current, temperature, and displacement. However, when facing highly dynamic cutting and multi-field coupling conditions, the machining system exhibits the following failure mechanisms: Nonlinear mismatch mechanism at the critical softening point: Aluminum alloys are thermally sensitive. When the transient heat flow accumulated in the processing area triggers the material to enter the thermal softening stage, the mechanical constitutive relationship of the aluminum alloy undergoes a nonlinear abrupt change. Since existing compensation algorithms are mostly based on linear mapping logic, they cannot provide continuous and matched gain adjustment near the thermal softening critical point, causing the compensation vector to produce physical overshoot, or inducing dynamic oscillations in the servo control loop due to step switching.

[0003] The mapping mismatch mechanism caused by material inhomogeneity: Microscopic constitutive differences within aluminum alloy parts (such as uneven hardness or microstructure segregation) induce high-frequency disturbances in cutting torque, which are simultaneously reflected in fluctuations in the electromagnetic current sequence. Traditional compensation models typically treat material stiffness as a steady parameter, lacking the perception and feedback dimension for material inhomogeneity disturbances. Under conditions of large material variance, the output of the preset thermal prediction model in the machining system is difficult to maintain a stable and consistent mapping relationship with the actual structural deformation state, leading to overall failure of compensation accuracy.

[0004] Failure Mechanism of Non-Stationary Signal Feature Stripping: Signals generated by multi-axis linkage machining exhibit non-stationary and time-varying characteristics. Low-frequency thermal evolution characteristics triggered by energy consumption and high-frequency fingerprint characteristics caused by material differences overlap in the time-frequency space. Existing frequency domain analysis methods are limited by the assumption of global stationarity and lack dynamic constraints on the local time-domain characteristics of the signal. This leads to spectral interference and feature mismatch in the decoupled physical components, making it impossible to provide pure physical input for subsequent spatial position compensation vector synthesis.

[0005] Based on the above analysis of physical characteristics, the transient heat flow evolution triggered by cutting energy consumption, dynamic load fluctuations, and material surface stiffness degradation exhibit nonlinear mismatch characteristics in the spatiotemporal dimension. This makes it difficult for the preset compensation vector in the machining system to maintain a stable and consistent mapping relationship with the evolution process of the actual physical field, thereby causing the overall failure of the steady-state characteristics of the machining system. Summary of the Invention

[0006] This invention provides an intelligent multi-axis error compensation control system and method for aluminum processing, which solves the problem that in the multi-axis linkage machining of aluminum alloys, the transient heat flow evolution triggered by cutting energy consumption, dynamic load fluctuations and material surface stiffness degradation exhibit nonlinear mismatch characteristics in the spatiotemporal dimension, making it difficult for the preset compensation vector in the machining system to maintain a stable and consistent mapping relationship with the evolution process of the actual physical field, thereby causing the overall failure of the steady-state characteristics of the machining system.

[0007] In view of the above problems, the present invention provides an intelligent multi-axis error compensation control method for aluminum processing, the method comprising the following steps: S1. Acquire current signal sequence: Synchronously acquire the current signal sequence of the multi-axis servo system and the machining trajectory information of the multi-axis CNC system, and parse the machining trajectory information to obtain the machining path normal unit vector; use a decomposition algorithm to decompose the current signal sequence into low-frequency components, mid-frequency components and high-frequency components in the frequency domain; S2. Generate physical feature components: Generate a material stiffness feature matrix based on the power spectrum energy distribution of the high-frequency components; calculate the dynamic blade deformation scalar using the material stiffness feature matrix and the mid-frequency components; perform time-domain integration on the low-frequency components to generate a transient thermal deformation vector. S3. Calculate the weighting coefficients: Generate the material confidence coefficient based on the dispersion of the material stiffness characteristic matrix; generate the thermal softening gain coefficient based on the magnitude of the transient thermal deformation vector; S4. Synthesized Output Vector: Based on the material confidence coefficient and the thermal softening gain coefficient, the transient thermal deformation vector and the dynamic tool deformation scalar are weighted and synthesized to output a spatial position compensation vector for correcting trajectory deviation.

[0008] Further, generating the material stiffness feature matrix in step S2 includes: Obtain the power spectral density of the high-frequency component; Identify the characteristic energy peak frequency in the power spectral density and match the characteristic energy peak frequency with a preset fingerprint database to correct the parameters in the material stiffness feature matrix.

[0009] Further, in step S4, the spatial position compensation vector is synthesized. Follow the formula below: in, The confidence coefficient for the material is... Let be the transient thermal deformation vector. The thermal softening gain coefficient is mentioned above. The preset gain coefficient, To make the dynamic change of the blade shape a scalar, This is the unit vector normal to the processing path.

[0010] Further, in step S3, the confidence coefficient of the material is generated. Follow the formula below: in, Let V be the variance of the high-frequency components. For adjustment coefficients, This is the smoothing constant.

[0011] Further, in step S3, the thermal softening gain coefficient is generated. Follow the formula below: in, Let the magnitude of the transient thermal deformation vector be denoted as . For reference, For the preset threshold, For gain factor, This represents the maximum gain boundary.

[0012] The present invention provides an intelligent multi-axis error compensation control system for aluminum processing. The system includes a processor and a memory. The memory stores computer program instructions, and the processor executes the instructions to implement the steps of the above method.

[0013] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0014] The technical solution provided in this application has at least the following technical effects: By establishing the analytical relationship between the time-domain local features and frequency-domain subband distribution of the servo current sequence, the variational mode decomposition algorithm can adaptively peel off the thermal field trend, dynamic load and material fingerprint components according to the nonlinear energy evolution in the aluminum cutting process, thereby eliminating the feature component mismatch caused by the overlap of physical signal features at the perception level.

[0015] By associating and mapping the power spectrum fingerprint features of high-frequency components with physical parameters in a preset fingerprint database, the material stiffness feature matrix can be corrected online as the material properties of the processing area fluctuate, thus ensuring that the mechanical deformation calculation logic is always within the constraint framework of the actual physical constitutive model.

[0016] By constructing a nonlinear thermal softening gain tuning logic based on the Sigmoid function, the spatial position compensation vector can generate a continuous and smooth weight evolution as the physical properties of the material evolve during the thermal softening stage. This suppresses overshoot and servo oscillation caused by compensation commands under cross-threshold conditions, ensuring that the multi-axis CNC system always operates in a steady-state range near the thermo-mechanical coupling critical point.

[0017] The synergistic realization of the above technical features enables the energy consumption and heat flow, dynamic load and material disturbance effects in the aluminum processing process to be unified within a physical response logic framework based on servo current sequence decoupling. It can simultaneously take into account the weighted suppression of material non-homogeneous interference and the dynamic adaptation to thermal softening nonlinear deformation under complex linkage conditions. Attached Figure Description

[0018] Figure 1 Flowchart of an intelligent multi-axis error compensation control method for aluminum processing; Figure 2 This is a diagram of the architecture of an intelligent multi-axis error compensation control system for aluminum processing. Detailed Implementation

[0019] 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.

[0020] Example Please see Figures 1 to 2 This invention provides an intelligent multi-axis error compensation control method for aluminum processing. The method consists of... Figure 2 The intelligent multi-axis error compensation control system for aluminum processing shown is implemented. The method includes the following steps: S1. Acquire current signal sequence: Synchronously acquire the current signal sequence of the multi-axis servo system and the machining trajectory information of the multi-axis CNC system, and parse the machining trajectory information to obtain the machining path normal unit vector; use a decomposition algorithm to decompose the current signal sequence into low-frequency components, mid-frequency components and high-frequency components in the frequency domain; S2. Generate physical feature components: Generate a material stiffness feature matrix based on the power spectrum energy distribution of the high-frequency components; calculate the dynamic blade deformation scalar using the material stiffness feature matrix and the mid-frequency components; perform time-domain integration on the low-frequency components to generate a transient thermal deformation vector. S3. Calculate the weighting coefficients: Generate the material confidence coefficient based on the dispersion of the material stiffness characteristic matrix; generate the thermal softening gain coefficient based on the magnitude of the transient thermal deformation vector; S4. Synthesized Output Vector: Based on the material confidence coefficient and the thermal softening gain coefficient, the transient thermal deformation vector and the dynamic tool deformation scalar are weighted and synthesized to output a spatial position compensation vector for correcting trajectory deviation.

[0021] The intelligent multi-axis error compensation control system for aluminum processing uses its internal logic processor as the command core to schedule the entire process of signal acquisition, data parsing, and instruction synthesis when performing aluminum processing error compensation tasks. During execution, the logic processor initiates a signal acquisition task, instructing the servo drives in the multi-axis CNC system to synchronously upload electromagnetic current data for each feed axis via an industrial Ethernet bus (e.g., EtherCAT) with a fixed sampling period of 250 microseconds. This electromagnetic current data enters the logic processor's circular buffer for timing alignment, forming a current signal sequence. Based on this homogeneous acquisition method, the current data of each axis has phase consistency on the time axis, and the perceptual deviations caused by asynchronous sampling are eliminated before the data enters the physical feature decoupling step.

[0022] The logic processor invokes a pre-defined variational mode decomposition (VMD) program to decouple the current signal sequence in the frequency domain. In the initialization parameter settings of the VMD program, the number of mode decompositions is configured to 3, the quadratic penalty factor is configured to 2000, and the convergence tolerance is configured as follows: The variational mode decomposition (VMD) program constructs a variational constraint model and iteratively calculates the current signal sequence in the frequency domain using the alternating direction multiplier method. The VMD program continuously updates each mode function and its corresponding center frequency until the update amount of the mode function reaches a preset convergence tolerance. This process separates the current signal sequence into three characteristic components with independent physical meaning: a low-frequency component, a mid-frequency component, and a high-frequency component.

[0023] The low-frequency components generated by the variational mode decomposition (VMD) program mainly cover the frequency band from 0 to 10 Hz, corresponding to the time-dependent drift characteristics of cutting energy consumption. The mid-frequency components mainly cover the frequency band from 10 to 500 Hz, corresponding to the mechanical load fluctuation characteristics during machining. The high-frequency components mainly cover the frequency band above 500 Hz, corresponding to the physical disturbance characteristics caused by the microstructural differences in aluminum. The logic processor stores the decoupled low-frequency, mid-frequency, and high-frequency components in separate data storage areas.

[0024] The variational mode decomposition algorithm provides time-domain localization and frequency analysis of the current signal sequence by adaptively dividing it into sub-bands. Through Wiener filtering, the algorithm extracts the high-frequency components of the energy spectrum distribution from the current signal sequence. During high-frequency component extraction, narrow-band center frequency localization filters out broadband electromagnetic noise generated by the machine tool environment. The mid-frequency components are determined by adaptive adjustment of the mode function center to lock the frequency characteristics of motor torque fluctuations. The low-frequency components, after removing high-frequency pulse interference, retain the energy trend reflecting the evolution of the accumulated thermal field.

[0025] The high-frequency components extracted by the logic processor are input to the material fingerprint quantization subunit, which calculates the power spectral density of the high-frequency components. The characteristic energy peak frequency in the power spectral density of the high-frequency components is identified and extracted. The logic processor calls a preset fingerprint library, which stores the mapping relationship between different aluminum alloy grades and characteristic frequency bands. For example, for 7075 aluminum, the characteristic energy peak frequency range in the preset fingerprint library is set to 850 Hz to 1200 Hz; for 6061 aluminum, the characteristic energy peak frequency range is set to 600 Hz to 800 Hz. The logic processor compares the characteristic energy peak frequency with the reference frequency spectrum characteristics of various aluminum alloy grades stored in the preset fingerprint library. When the characteristic energy peak frequency falls within the characteristic window of a specific material in the preset fingerprint library, the material fingerprint quantization subunit retrieves the physical parameters corresponding to that specific material from the preset fingerprint library. These physical parameters are input into the material stiffness feature matrix to complete the online correction of the stiffness constant term. The real-time state of the material stiffness characteristic matrix serves as the calculation benchmark for subsequent mechanical inversion steps.

[0026] The intermediate frequency component is transmitted to the deformation inversion subunit to extract the real-time electromagnetic torque. The deformation inversion subunit performs a subtraction operation between the real-time electromagnetic torque and a preset no-load torque reference stored in non-volatile memory, thereby generating a torque difference. This torque difference is input to the elastic deformation analysis operator integrated within the deformation inversion subunit, which simultaneously calls the material stiffness feature matrix corrected by the material fingerprint quantization subunit. By mapping the torque difference to the stiffness parameters in the material stiffness feature matrix, the deformation inversion subunit outputs a dynamic tool deformation scalar. The dynamic tool deformation scalar numerically represents the instantaneous deviation of the tool tip from the nominal trajectory due to the machining load.

[0027] The low-frequency components obtained by decoupling from the variational mode decomposition program are guided to the thermal potential energy prediction subunit. The thermal potential energy prediction subunit performs a time-domain square integral operation on the low-frequency components, obtaining the cumulative energy consumption of the current processing interval by calculating the sum of the square of the current amplitude and the sampling period. The cumulative energy consumption is then used as an excitation signal input to a preset thermal impedance model. The thermal impedance model simulates the heat transfer process from the heat source to the machine tool structure through a Laplace transfer function, converting the thermal response results into geometric displacement deviations in a three-dimensional coordinate system. The thermal potential energy prediction subunit ultimately generates a transient thermal deformation vector, which includes positioning error components caused by the thermal elongation of the structure in each coordinate axis direction.

[0028] In the process of correcting the parameters of the material stiffness feature matrix, there are multiple alternative feature extraction paths. Besides power spectral density analysis, high-frequency components can also be fed into a wavelet packet decomposition module for three-level decomposition. By extracting the wavelet packet energy normalization coefficients of the high-frequency components in each sub-band, a feature vector characterizing the constitutive differences of the material is formed. This feature vector is input into a preset fingerprint database for Euclidean distance matching to complete the identification of material properties. Furthermore, extracting cepstral peaks by performing cepstral transformation on the high-frequency components and simultaneously correcting the parameters in the material stiffness feature matrix based on the cutting frequency characteristics reflected by the cepstral peaks also constitutes an equivalent implementation path of this invention.

[0029] The logic processor performs statistical operations on the high-frequency components to obtain the real-time variance of the high-frequency components. Real-time variance of high-frequency components The weights are input into the weight tuning module of the logic processor to perform the material confidence coefficient calculation. The calculation. The logic processor executes the formula: To generate material confidence coefficients. Adjustment coefficients are performed within this calculation logic. The smoothing constant is preset to 0.5. To prevent non-zero small constants with a denominator of zero from being preset, for example, as Used for real-time variance of uniform and high-frequency components in aluminum materials. Under conditions approaching zero, the denominator is maintained at a non-zero value. When aluminum exhibits non-homogeneous characteristics, it leads to changes in the real-time variance of high-frequency components. As the value increases, the material confidence coefficient... The value of decreases accordingly. This is used to adjust the weight of the transient thermal deformation vector, ensuring the mapping relationship between the thermal impedance model and actual operating conditions is under control.

[0030] Thermal softening gain coefficient The generation is based on the magnitude of the transient thermal deformation vector by the logic processor. Execution. Magnitude of the transient thermal deformation vector. It is extracted and mapped to a predefined Sigmoid function model. The logic processor executes the formula: The thermal softening gain coefficient is calculated using this formula. In this formula, the gain factor... Used to control the growth slope of the thermal softening gain coefficient as a function of the magnitude of the transient thermal deformation vector; reference base. The preset value is the thermal displacement reference value corresponding to the point where the elastic modulus of aluminum decreases; preset threshold. Set as the critical threshold for enabling the thermal softening effect; maximum gain boundary. It is set to 1.5 to limit the upper bound of the gain output.

[0031] The Sigmoid function is applied when the magnitude of the transient thermal deformation vector approaches a preset threshold. Within a certain range, it provides continuously varying gain output. This ensures that the servo driver output remains continuous during gain switching, and that the displacement deviation remains within a preset threshold envelope during the aluminum material's thermal softening stage.

[0032] When setting the thermal softening gain coefficient In this process, there are multiple equivalent nonlinear mapping paths. Besides the Sigmoid function, the logic processor can also call the hyperbolic tangent function (tanh) for gain tuning, or use a piecewise linear saturation function to simulate the increased deformation trend of aluminum after thermal softening. Furthermore, using a specific exponential distribution function to adjust the weighting ratio of the dynamic tool deformation scalar in real time according to the magnitude evolution of the transient thermal deformation vector also constitutes an equivalent implementation scheme. These equivalent paths provide a preset threshold... The nonlinear gain characteristics in the vicinity maintain the system's ability to respond to the degradation of material stiffness under thermo-coupling effects.

[0033] The synthesis output unit receives the material confidence coefficient. Thermal softening gain coefficient Transient thermal deformation vector And dynamically transforming the blade shape into a scalar Then, a vector weighting algorithm is executed to generate a spatial location compensation vector. Spatial position compensation vector The calculation follows the formula: In this formula, the gain coefficient is preset. Used to adjust the dynamic transformation of the blade shape into a scalar. The original weighting ratio in the composite vector; the unit vector normal to the processing path. Provides geometric guidance for the compensation action; the negative sign before the formula indicates the spatial position compensation vector. The distribution is opposite to the direction of the evolution of processing errors.

[0034] Specifically, the processing path normal unit vector The generation process is completed by the logic processor performing geometric analysis on the machining trajectory information. The logic processor extracts the three-dimensional spatial coordinate sequence of the current machining point from the real-time interpolator of the multi-axis CNC system. The logic processor obtains the instantaneous tangent vector of the machining trajectory by performing a first-order difference operation on the three-dimensional spatial coordinate sequence. Using a preset orthogonalization operator, the logic processor resolves the unit vector pointing to the normal of the machining contour in the plane perpendicular to the instantaneous tangent vector. This unit vector is defined as the machining path normal unit vector. Processing path normal unit vector The real-time calculation provides a geometric reference for the spatial projection of the spatial position compensation vector in subsequent steps.

[0035] Under normal processing conditions, if the real-time variance of the high-frequency components remains within a preset range and the magnitude of the transient thermal deformation vector does not reach a preset threshold, then the material confidence coefficient... The thermal softening gain coefficient is close to 1. It is in a state approaching 0. At this time, the spatial position compensation vector... It consists of a linear combination of the transient thermal deformation vector and the original dynamic knife deformation scalar.

[0036] When specific conditions such as material fluctuation points and thermal softening points occur during aluminum processing, the logic processor adjusts the weighting coefficients to modify the composition of the spatial position compensation vector. When the material fluctuation point is reached, the real-time variance of the high-frequency components increases, causing the material confidence coefficient to decrease from 0.95 to 0.40. At this point, the contribution weight of the transient thermal deformation vector to the spatial position compensation vector is suppressed. When the processing reaches the thermal softening point, i.e., the magnitude of the transient thermal deformation vector exceeds a preset threshold, the thermal softening gain coefficient increases nonlinearly. In a typical thermal softening case, the thermal softening gain coefficient increases from 0 to 0.60. In this state, the proportional coefficient of the dynamically variable tool shape is amplified, and the magnitude of the spatial position compensation vector increases accordingly to match the enhanced elastic deformation characteristics of the aluminum material after thermal softening. Through this data-guided adaptive adjustment, the spatial position compensation vector completes the adjustment of its direction and amplitude under different physical conditions.

[0037] When calculating the thermal softening gain coefficient, the logic processor limits the output result by using a preset maximum gain boundary. This maximum gain boundary is set as a safety threshold corresponding to the physical stiffness limit of the machine tool structural components. Under conditions where extreme high temperatures cause a significant jump in the magnitude of the transient thermal deformation vector, the calculated thermal softening gain coefficient is forcibly constrained within the maximum gain boundary. This boundary verification logic ensures that the spatial position compensation vector generated by the synthesized output unit remains within the linear execution range of the servo system, keeping the multi-axis CNC system within the preset safe operating boundaries.

[0038] The numerical result of the spatial position compensation vector is injected into the servo control loop of the CNC system in real time to correct machining trajectory deviations. The logic processor extracts the path normal unit vector by acquiring the current machining path information. The spatial position compensation vector and the path normal unit vector are geometrically synthesized so that the compensation action acts on the physical normal of the machining error. The synthesized correction command is sent to the servo driver of each feed axis through a dedicated data register. After receiving the correction command, the servo driver changes the instantaneous position setpoint of the motor. The generation and injection process of the spatial position compensation vector is completed within a single interrupt cycle of the logic processor.

[0039] The processor and memory constitute the computing platform for an intelligent multi-axis error compensation control system used in aluminum processing. The memory pre-stores computer program instructions for the steps of the intelligent multi-axis error compensation control method applied to aluminum processing. The processor retrieves and executes the computer program instructions from the memory via the instruction bus. During instruction scheduling, the memory is divided into a data input area for storing the original current signal sequence, an operation cache area for storing intermediate calculation operators, and an instruction output area for storing the spatial position compensation vector. When executing the computer program instructions, the processor interacts with the memory through registers to logically tune the material stiffness characteristic matrix, material confidence coefficient, and thermal softening gain coefficient within a preset real-time period. The memory also stores a preset fingerprint database and the physical constants required for the thermal impedance model, for the processor to perform address addressing and parameter retrieval.

[0040] Computer-readable storage media are configured as non-transitory carriers for storing computer programs. The computer programs are stored in the computer-readable storage media in the form of compiled binary code. During the initialization phase of the multi-axis CNC system, the binary code stored in the computer-readable storage media is loaded into memory via input / output interfaces. The physical entity of the computer-readable storage media includes, but is not limited to, flash memory, hard disk storage, or optical disk storage. The processor establishes a task flow encompassing signal acquisition, frequency decoupling, weight tuning, and vector synthesis by accessing the task image in the computer-readable storage media. This hardware deployment method ensures that the logical execution of the multi-axis error compensation control method at the computer hardware level is stable.

[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 intelligent multi-axis error compensation control method for aluminum processing, characterized in that, The method includes the following steps: S1. Acquire current signal sequence: Synchronously acquire the current signal sequence of the multi-axis servo system and the machining trajectory information of the multi-axis CNC system, and parse the machining trajectory information to obtain the machining path normal unit vector; use a decomposition algorithm to decompose the current signal sequence into low-frequency components, mid-frequency components and high-frequency components in the frequency domain; S2. Generate physical feature components: Generate a material stiffness feature matrix based on the power spectrum energy distribution of the high-frequency components; calculate the dynamic blade deformation scalar using the material stiffness feature matrix and the mid-frequency components; perform time-domain integration on the low-frequency components to generate a transient thermal deformation vector. S3. Calculate the weighting coefficients: Generate the material confidence coefficient based on the dispersion of the material stiffness characteristic matrix; generate the thermal softening gain coefficient based on the magnitude of the transient thermal deformation vector; S4. Synthesized Output Vector: Based on the material confidence coefficient and the thermal softening gain coefficient, the transient thermal deformation vector and the dynamic tool deformation scalar are weighted and synthesized to output a spatial position compensation vector for correcting trajectory deviation.

2. The method according to claim 1, characterized in that, Step S2, generating the material stiffness feature matrix, includes: Obtain the power spectral density of the high-frequency component; Identify the characteristic energy peak frequency in the power spectral density and match the characteristic energy peak frequency with a preset fingerprint database to correct the parameters in the material stiffness feature matrix.

3. The method according to claim 1, characterized in that, In step S4, the spatial position compensation vector is synthesized. Follow the formula below: in, The confidence coefficient for the material is... Let be the transient thermal deformation vector. The thermal softening gain coefficient is mentioned above. The preset gain coefficient, To make the dynamic change of the blade shape a scalar, This is the unit vector normal to the processing path.

4. The method according to claim 1, characterized in that, In step S3, the confidence coefficient of the material is generated. Follow the formula below: in, Let V be the variance of the high-frequency components. For adjustment coefficients, This is the smoothing constant.

5. The method according to claim 1, characterized in that, In step S3, the thermal softening gain coefficient is generated. Follow the formula below: in, Let the magnitude of the transient thermal deformation vector be denoted as . For reference, For the preset threshold, For gain factor, This represents the maximum gain boundary.

6. An intelligent multi-axis error compensation control system for aluminum processing, characterized in that, The system includes a processor and a memory, the memory storing computer program instructions, and the processor executing the instructions to implement the steps of the method according to any one of claims 1 to 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images