Dynamic thermal deformation compensation method of numerical control grinding machine

Abstract

The invention discloses a dynamic thermal deformation compensation method for a numerical control grinding machine. The dynamic thermal deformation compensation method comprises the following steps that three-dimensional temperature field data of a mechanical head are obtained in real time through a distributed temperature monitoring unit; inputting the three-dimensional temperature field data into a numerical control grinding machine thermal deformation digital twin model, and predicting thermal deformation and spatial distribution of the mechanical head in a future time period through coupling temperature field-stress field-deformation field simulation calculation; and generating a compensation strategy according to the thermal deformation and the spatial distribution, driving the radiator to move to a thermal deformation sensitive area for directional heat dissipation, and adjusting the position and posture of the grinding head through a five-axis linkage system of the numerical control grinding machine to offset the predicted thermal deformation. A three-dimensional temperature field of the mechanical head is reconstructed in real time through the distributed temperature monitoring unit, and accurate real-time prediction of thermal deformation and spatial distribution is achieved; and meanwhile, a five-axis system is cooperated to pre-adjust the position and posture of the grinding head, and the thermal deformation restraining precision and the machining stability are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of numerical control grinding machine control, and in particular relates to a dynamic thermal deformation compensation method of a numerical control grinding machine. BACKGROUND

[0002] In the field of high-precision machining of five-axis linkage numerical control grinding machines, thermal deformation error has become the core bottleneck restricting machining precision. During long-time operation of the grinding machine, factors such as heat generated by the motor inside the mechanical head, heat generated by friction between the polishing head and the workpiece, and the like cause uneven temperature distribution of key components, causing non-uniform thermal expansion. The traditional compensation scheme relies on temperature sensors at fixed positions to monitor local temperature and estimates the overall deformation amount through empirical formulas. However, the thermal deformation of the mechanical head, as a complex three-dimensional structure, has significant spatial nonlinearity and time-varying characteristics - a single measurement point cannot capture the gradient distribution of the temperature field, and a static compensation model cannot adapt to the dynamic thermal accumulation effect under high-speed rotating conditions. In the prior art, the heat sink is usually fixedly installed or can only be manually adjusted in position, resulting in a disconnection between the heat dissipation efficiency and the heat source migration, and the precise directional temperature control cannot be achieved.

[0003] Further, current thermal deformation prediction mostly uses offline simulation or simplified mathematical models, and fails to deeply couple with the real-time state of the equipment. Due to the lack of continuous reconstruction capability of the three-dimensional temperature field of the mechanical head, the system cannot predict the spatial evolution trend of the thermal deformation. When the polishing head performs high-speed directional motion, airflow disturbance can distort the temperature monitoring data, and the traditional method does not consider the influence of motion inertia on temperature measurement accuracy, resulting in amplification of prediction deviation. In addition, the coordination between compensation action and machine tool motion control is insufficient: the position adjustment of the heat sink relies on mechanical transmission (such as rack drive), but the path planning does not combine the dynamic migration law of the heat-sensitive area; the pose compensation of the five-axis linkage system does not real-time fuse the kinematic effects such as centrifugal force, and is prone to compensation lag in complex trajectory machining.

[0004] In recent years, digital twin technology has provided a new idea for thermal deformation control, but the existing research has two major defects: first, the twin model mostly runs on the host computer, and it is difficult to match the real-time control cycle with the numerical control system, so the simulation results cannot directly drive the actuator; second, the model parameters are fixed, and there is no online correction mechanism for material thermal properties, and the prediction accuracy declines after long-term operation. These limitations make high-value workpieces (such as precision molds) still face the risk of size out-of-tolerance due to thermal deformation during long-time machining, and an integrated solution that deeply embeds the control system, fuses multi-physical field real-time simulation and dynamic compensation strategy is urgently needed. SUMMARY

[0005] The embodiment of the application aims to provide a dynamic thermal deformation compensation method of a numerical control grinding machine, which reconstructs a three-dimensional temperature field of a mechanical head in real time through a distributed temperature monitoring unit, dynamically couples a temperature field-stress field-deformation field simulation based on an embedded digital twin model, realizes accurate real-time prediction of thermal deformation and spatial distribution, drives a radiator to move to a heat-sensitive area for dynamic heat dissipation, and cooperates with a five-axis system to pre-adjust the pose of a grinding head, effectively overcomes the limitations of traditional single-point monitoring and the lag problem of static model compensation, and significantly improves the thermal deformation suppression precision and processing stability under high-speed variable working conditions.

[0006] To solve the above technical problems, the first aspect of the embodiment of the application provides a dynamic thermal deformation compensation method of a numerical control grinding machine, the numerical control grinding machine comprising: a distributed temperature monitoring unit arranged at a plurality of preset positions of a mechanical head and a radiator arranged at a corresponding position of the mechanical head, the radiator being movable along a straight line, and the compensation method comprising the following steps:

[0007] real-time acquisition of three-dimensional temperature field data of the mechanical head through the distributed temperature monitoring unit;

[0008] inputting the three-dimensional temperature field data into a thermal deformation digital twin model of the numerical control grinding machine, predicting the thermal deformation and spatial distribution of the mechanical head in a future time period through coupling temperature field-stress field-deformation field simulation calculation;

[0009] generating a compensation strategy according to the thermal deformation and spatial distribution, driving the radiator to move to a heat deformation sensitive area for directional heat dissipation, and adjusting the pose of the grinding head through a five-axis linkage system of the numerical control grinding machine to offset the predicted thermal deformation;

[0010] The thermal deformation digital twin model of the grinding machine is a real-time simulation module embedded in the control system of the numerical control grinding machine.

[0011] Further, the real-time acquisition of three-dimensional temperature field data of the mechanical head through the distributed temperature monitoring unit comprises:

[0012] acquiring discrete temperature measurement values at intersection points of different heat conduction paths in the mechanical head;

[0013] reconstructing a spatial field based on a three-dimensional heat conduction model of the mechanical head, the three-dimensional heat conduction model containing thermal expansion coefficients and specific heat capacity parameters of materials in each region of the mechanical head, and mapping the discrete point temperature to a continuously distributed three-dimensional temperature field through a spatial weighted interpolation algorithm;

[0014] The three-dimensional temperature field is dynamically compensated in combination with the real-time motion state of the mechanical head. When the mechanical head performs high-speed rotation or direction changing action, the temperature measurement deviation caused by airflow disturbance is corrected according to a preset inertia temperature rise compensation coefficient, and three-dimensional dynamic temperature field data that are time and space synchronous are generated.

[0015] The three-dimensional dynamic temperature field data are subjected to noise reduction processing, an adaptive Kalman filter is used to eliminate electromagnetic interference and mechanical vibration noise, and three-dimensional temperature field data that match the control period are output.

[0016] Further, the three-dimensional temperature field is dynamically compensated in combination with the real-time motion state of the mechanical head. When the mechanical head performs high-speed rotation or direction changing action, the temperature measurement deviation caused by airflow disturbance is corrected according to a preset inertia temperature rise compensation coefficient, and three-dimensional dynamic temperature field data that are time and space synchronous are generated, including:

[0017] Real-time motion parameters of the mechanical head are acquired, the real-time motion parameters include a current speed value and an acceleration direction, a compensation coefficient matrix pre-stored in a numerical control system is called to match the current motion state, and a compensation coefficient corresponding to the current motion state is obtained;

[0018] The three-dimensional temperature field is dynamically corrected based on the matched compensation coefficient. The compensation coefficient linearly increases with the increase of the speed, and the direction weight factor is adjusted to correct the amount when the acceleration direction changes;

[0019] The three-dimensional temperature field that is dynamically corrected is subjected to time axis alignment processing based on the thermal conduction delay parameter of the material of the mechanical head, so that the temperature field data and the actual physical state of the mechanical head are kept time and space synchronous.

[0020] Further, the three-dimensional temperature field data are input into a thermal deformation digital twin model of the numerical control grinding machine, and through coupling simulation calculation of the temperature field-stress field-deformation field, the thermal deformation amount and the spatial distribution of the mechanical head in a future time period are predicted, including:

[0021] A material attribute database of the thermal deformation digital twin model of the numerical control grinding machine is initialized. The material attribute database pre-stores the thermal conductivity, elastic modulus and thermal expansion coefficient of each component of the mechanical head, and constructs a transient temperature boundary condition in association with the real-time three-dimensional temperature field data;

[0022] Based on the geometric topological structure of the mechanical head, non-uniform grids are dynamically divided, the grid node density is increased in the area of the moving path of the heat sink, and the transient temperature boundary condition is mapped to the grid nodes to generate an initial temperature field distribution;

[0023] A coupling solver is called to synchronously and iteratively calculate the temperature field and the stress field. The heat sink position parameter is introduced as a convection heat exchange boundary in the temperature field calculation, and the thermal stress distribution inside the mechanical head is calculated according to the temperature gradient and the material attribute in the stress field calculation;

[0024] mapping the thermal stress distribution into a deformation variable through a deformation field conversion model, combining a kinematic constraint condition to correct the deformation direction, and outputting a three-dimensional deformation prediction value of a key point on the surface of the mechanical head;

[0025] spatially interpolating and reconstructing the three-dimensional deformation prediction value to generate a continuously distributed thermal deformation and spatial distribution data, a time resolution of which matches the control period.

[0026] Further, after the spatial interpolation and reconstruction of the three-dimensional deformation prediction value, the method further comprises:

[0027] real-time collection of actual deformation variables of key points on the surface of the mechanical head through a laser displacement sensor installed on the side wall of the mechanical head, and construction of a residual sequence with the deformation prediction value at the same position;

[0028] when the sliding variance of the residual sequence exceeds a tolerance threshold, based on a digital twin model parameter self-correction module, the thermal expansion coefficient in the material attribute database is optimized in reverse in combination with the residual distribution;

[0029] the optimized thermal expansion coefficient is used as a new input for thermal stress distribution calculation, and the coupled simulation calculation step is iteratively executed until the residual sequence is restored to the tolerance range.

[0030] Further, the method of generating a compensation strategy according to the thermal deformation and spatial distribution, and driving the heat sink to move to the thermal deformation sensitive area for directional heat dissipation, comprises:

[0031] identifying an area where the temperature gradient in the spatial distribution exceeds a preset threshold as a thermal deformation sensitive area, and calculating the shortest straight line movement path from the mechanical head to the heat sink based on the spatial coordinates of the thermal deformation sensitive area;

[0032] generating a heat dissipation intensity control instruction according to the thermal deformation, triggering a high-speed operation mode of the heat sink fan when the thermal deformation reaches a first critical value, and generating a rack drive pulse signal at the coordinates of the thermal deformation sensitive area;

[0033] coupling and checking the rack drive pulse signal with real-time motion parameters of the mechanical head, prolonging the residence time of the heat sink when the mechanical head is in a high-speed rotating state, and ensuring that the directional heat dissipation action is completed synchronously with the mechanical motion state.

[0034] Further, the method of calculating the shortest straight line movement path from the mechanical head to the heat sink based on the spatial coordinates of the thermal deformation sensitive area, comprises:

[0035] obtaining a migration rate by dividing the displacement difference of the center point of the thermal deformation sensitive area in adjacent control periods by the time interval, and starting a dynamic path planning mode when the migration rate exceeds a preset threshold.

[0036] An optimization objective function of heat sink movement energy consumption and thermal compensation efficiency is constructed, a temperature gradient descent rate is taken as a weight factor, and a curve movement path with the lowest energy consumption and optimal compensation efficiency is solved by a gradient descent algorithm;

[0037] The curve movement path is discretized into a rack driving pulse sequence, and pulse interval time is dynamically adjusted according to real-time rotating speed of the mechanical head, so that the movement speed of the heat sink is synchronized with the migration of the heat-sensitive area.

[0038] Further, the adjustment of the position of the grinding head by the five-axis linkage system of the numerical control grinding machine to offset the predicted thermal deformation amount comprises:

[0039] A compensation vector of the position of the grinding head is constructed based on the thermal deformation amount and spatial distribution, the thermal deformation amount is converted into a three-dimensional offset amount in a working coordinate system of the grinding head by a deformation space mapping model, and the three-dimensional offset amount comprises a linear displacement compensation value and an angle deflection compensation value;

[0040] The three-dimensional offset amount is dynamically coupled with the real-time movement parameters, the centrifugal force deformation effect amount is calculated according to real-time rotating speed and the mass center position parameters of the mechanical head when the mechanical head rotates at high speed, the angle deflection compensation value is corrected, the position adjustment instruction of the grinding head is generated, and the compensation action is synchronously executed by driving the servo motor.

[0041] Correspondingly, a second aspect of the embodiment of the present application provides an electronic device, comprising: at least one processor; and a memory connected with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the dynamic thermal deformation compensation method of the numerical control grinding machine.

[0042] Correspondingly, a fourth aspect of the embodiment of the present application provides a computer readable storage medium, which stores computer instructions, and the instructions are executed by a processor to implement the dynamic thermal deformation compensation method of the numerical control grinding machine.

[0043] The above technical solutions of the embodiment of the present application have the following beneficial technical effects:

[0044] 1. Through the deep coupling of temperature field-stress field-deformation field simulation by embedded digital twin model, combined with non-uniform dynamic mesh technology to locally encrypt calculation nodes in the heat dissipation path area, and introducing the position parameters of the heat sink as the convective heat transfer boundary condition, the high-fidelity quantization of thermal stress distribution is realized; simultaneously, the residual feedback of the measured deformation and the predicted value is used to online reverse optimize the material thermal expansion coefficient, solving the problem of model precision degradation caused by long-term operation, finally outputting the spatio-temporal synchronous deformation prediction data within the control cycle of the numerical control system, reducing the thermal deformation prediction error by more than 60%, breaking through the disconnection bottleneck of traditional offline model and real-time control;

[0045] 2. Based on the real-time motion parameters (rotation speed / acceleration) of the mechanical head, the inertia temperature rise compensation coefficient is dynamically called to correct the temperature measurement deviation caused by air flow disturbance under high-speed variable direction working condition, and a spatio-temporal synchronous three-dimensional dynamic temperature field is generated; then, through the migration rate prediction of the heat sensitive area, the heat sink curve path with optimal energy consumption is planned with the temperature gradient decline rate as the weight factor, and the path is discretized into pulse sequence to dynamically adjust the moving rhythm, realizing the real-time tracking of the heat dissipation position and the heat source migration; at the same time, according to the threshold value of thermal deformation, the heat dissipation intensity and the residence time are linked and controlled, the heat dissipation time of the heat sink is prolonged during high-speed rotation, and the overall heat dissipation efficiency is improved by 40%, completely solving the disconnection problem between traditional fixed heat dissipation mode and dynamic migration of heat source;

[0046] 3. Through the deformation space mapping model, the thermal deformation is converted into three-dimensional offset (including linear displacement and angle deflection) in the working coordinate system of the polishing head, and the centrifugal force deformation effect is dynamically calculated by combining real-time motion parameters to correct the angle compensation value under high-speed rotation working condition; simultaneously driving the heat sink movement and the five-axis linkage system, the pose adjustment instructions are executed by the servo motor within the same control cycle, so that the heat dissipation action and the pose compensation cooperatively offset the thermal deformation and the centrifugal force effect, reducing the size tolerance risk in complex trajectory machining by 70%, realizing the global dynamic cooperation of thermal deformation suppression and kinematic effect. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 It is the dynamic thermal deformation compensation method flow chart of the numerical control grinding machine provided by the embodiment of the present application; Figure 2 It is a schematic diagram of the numerical control grinding machine including the dynamic thermal deformation compensation assembly provided by the embodiment of the present application; Figure 3 It is a schematic diagram of the dynamic thermal deformation compensation assembly provided by the embodiment of the present application; Figure 4 It is a schematic diagram of the distributed temperature monitoring unit provided by the embodiment of the present application; Figure 5 It is a schematic diagram of the dust reduction assembly provided by the embodiment of the present application.

[0052] Reference signs:

[0053] 1. CNC grinding machine main body; 2. Dynamic thermal deformation compensation assembly; 21. Mechanical head; 22. Polishing head; 23. Distributed temperature monitoring unit; 24. Driving motor; 25. First gear; 26. Second gear; 27. Third gear; 28. Fourth gear; 29. Rack; 209. Fixed plate; 208. Radiator; 3. Dust reduction assembly; 31. Dust collector; 32. Stretching pipe; 33. Air outlet; 34. Rubber band; 35. Dust filter bag. DETAILED DESCRIPTION

[0054] In order to make the objectives, technical solutions and advantages of the present application clearer and more comprehensible, the present application will be further described in detail below with reference to specific embodiments and drawings. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present application. In addition, in the following description, the description of well-known structures and techniques is omitted to avoid unnecessary confusion of the concept of the present application.

[0055] As Figure 2 , Figure 3 and Figure 4As shown, a five-axis linkage numerical control grinding machine includes a numerical control grinding machine body 1 and a dynamic thermal deformation compensation assembly 2; the dynamic thermal deformation compensation assembly 2 is arranged on the numerical control grinding machine body 1, and contains a mechanical head 21, a grinding head 22, a distributed temperature monitoring unit 23, a driving assembly, a gear transmission set, a rack 29 and a radiator 208, and errors caused by thermal deformation of the grinding machine are compensated through the cooperation of various components. In the dynamic thermal deformation compensation assembly 2, the mechanical head 21 is installed on the numerical control grinding machine body 1, the grinding head 22 is installed below the mechanical head 21, the distributed temperature monitoring unit 23 is installed on one side of the mechanical head 21, and is used for monitoring thermal deformation data in real time; the driving assembly contains a driving motor 24, and the gear transmission set includes a first gear 25, a second gear 26, a third gear 27 and a fourth gear 28; the driving motor 24 is connected with the first gear 25, the first gear 25 is engaged with the second gear 26, the second gear 26 is engaged with the third gear 27, the third gear 27 is engaged with the fourth gear 28, and the fourth gear 28 is engaged with the rack 29; the rack 29 is connected with the radiator 208, and is used for driving the radiator 208 to adjust the position of the workpiece, and cooperates with the distributed temperature monitoring unit 23 to monitor and compensate the thermal deformation of the grinding head 22 and other tools; the dynamic thermal deformation compensation assembly 2 further includes a fixed plate 209, the fixed plate 209 is used for fixing the radiator 208, and ensures the stability of the radiator 208 during the adjustment of the position; the radiator 208 includes a fan, air flow is accelerated through the rotation of the fan, and the mechanical head 21, the grinding head 22 and other components are cooled, and the position of the radiator 208 can be changed along with the driving of the rack 29, so that the thermal deformation compensation demand is better adapted; the distributed temperature monitoring unit 23 is a high-precision temperature sensor, can monitor the temperature change of the mechanical head 21 and the grinding head 22 in real time and accurately, and feeds back data to a control system, so as to provide accurate basis for thermal deformation compensation; the driving motor 24 is a servo motor, can accurately control the rotation of the gear according to the data fed back by the distributed temperature monitoring unit 23, and then accurately adjusts the position of the rack 29, so that the position of the radiator 208 is accurately adjusted, and the thermal deformation is effectively compensated.

[0056] Please refer to Figure 1 and Figure 2 , the first aspect of the embodiment of the application provides a dynamic thermal deformation compensation method of a numerical control grinding machine, the numerical control grinding machine includes: a distributed temperature monitoring unit arranged at a plurality of preset positions of a mechanical head 21 and a radiator 208 arranged at a corresponding position of the mechanical head 21, the radiator 208 can move along a straight line, and the compensation method includes the following steps:

[0057] Step S100, three-dimensional temperature field data of the mechanical head 21 is acquired in real time through the distributed temperature monitoring unit.

[0058] Distributed temperature monitoring units (such as high-precision thermocouples, thermal resistors, or infrared temperature measurement sensor arrays) are strategically pre-arranged at multiple key and preset locations of the machine head 21, forming a sensing network covering the key structures of the machine head 21. These sensors continuously and synchronously collect real-time temperature values at their installation points. Through high-speed data acquisition channels, the temperature readings of each discrete point are collected and processed, and using spatial interpolation or temperature field reconstruction algorithms based on the geometry of the machine head 21, a continuous three-dimensional temperature field model reflecting the temperature distribution of the entire machine head 21 at the current time is constructed. This three-dimensional temperature field data is no longer isolated single-point temperature, but accurately depicts the heat accumulation area, temperature gradient distribution, and overall thermal state of the machine head 21 in the spatial dimension.

[0059] Step S200, input the three-dimensional temperature field data into the thermal deformation digital twin model of the CNC grinding machine, and through coupled temperature field-stress field-deformation field simulation calculation, predict the thermal deformation amount and spatial distribution of the machine head 21 in the future time period. The thermal deformation digital twin model of the CNC grinding machine is a real-time simulation module embedded in the control system of the CNC grinding machine.

[0060] Based on the real-time three-dimensional temperature field data obtained in step S100, it is input as an initial condition and boundary condition into the "thermal deformation digital twin model" embedded in the control system of the CNC grinding machine. The model is a high-fidelity mapping of the physical entity of the CNC grinding machine in virtual space, and its core is the ability to perform real-time simulation calculation of multiple physical fields. First, according to the input temperature field data, the model simulates the conduction and convection process of heat in the structure material (such as cast iron, steel, etc.) of the machine head 21, and deduces the dynamic evolution of the temperature field (temperature field simulation) in the future short time (such as several seconds to tens of seconds). Then, the model maps the calculated temperature distribution to the geometry of the machine head 21, calculates the thermal expansion / contraction amount (thermal strain) caused by temperature unevenness according to the thermal expansion coefficient (CTE) and other properties of the material, and further considers the constraint conditions (such as fixed connection points) and material mechanics performance (elastic modulus, Poisson's ratio) of the machine head 21 structure, and through stress field simulation calculation, the internal stress distribution caused by thermal strain. Finally, the stress field results drive the deformation field simulation, accurately calculating the small elastic deformation amount (thermal deformation amount) and its specific distribution form (such as deflection, bending, elongation, etc.) in three-dimensional space of each part of the machine head 21 (especially the key shaft system and spindle head) caused by thermal stress in the future predicted time period.

[0061] Step S300, generate a compensation strategy according to the thermal deformation amount and spatial distribution, drive the heat sink 208 to move to the thermal deformation sensitive area for directional heat dissipation, and adjust the pose of the grinding head 22 through the five-axis linkage system of the CNC grinding machine to offset the predicted thermal deformation amount.

[0062] Based on the future thermal deformation amount and its spatial distribution information predicted in step S200, the compensation strategy generation module will make comprehensive judgment and analysis. The strategy contains two levels of coordinated actions: first, the active thermal management level: identify the "thermal deformation sensitive area" that has the greatest impact on the overall thermal deformation (i.e. the area with significant predicted deformation or high deformation sensitivity coefficient). Then, the control system drives the heat sink 208 (such as micro fans, semiconductor cooling fins, etc.) installed on the mechanical head 21 with linear motion capability to quickly and accurately move to the above or near these predicted sensitive areas. The heat sink 208 starts to perform local and intensive "directed heat dissipation" on the target area, aiming to actively intervene in the temperature rise process of the area, slow down its temperature rising speed or promote its cooling, thereby weakening the driving force of future thermal deformation in the area from the source. Second, the real-time compensation of geometric error level: at the same time, the predicted thermal deformation amount (including size and direction) is converted into motion instruction offset of the five-axis linkage system (X, Y, Z linear axes and A, C rotary axes) of the numerical control grinding machine. During the machining process, the control system superimposes this offset to the G code instructions of the original machining trajectory. The five-axis linkage system adjusts the actual pose (position and attitude) of the polishing head 22 (grinding wheel) in space accordingly, so that the motion trajectory of the polishing head 22 relative to the workpiece occurs a displacement equal in size and opposite in direction to the predicted thermal deformation amount, thereby directly offsetting the spatial position error of the spindle and the polishing head 22 caused by the thermal deformation of the mechanical head 21 in the geometric space, ensuring that the actual position of the machining point always meets the programming expectation.

[0063] By real-time sensing of the three-dimensional temperature field of the mechanical head 21, accurately predicting the future thermal deformation trend by using the embedded digital twin model, and taking the double coordinated compensation mechanism of directional movement of the heat sink 208 (active inhibition of heat source) and dynamic adjustment of the five-axis linkage pose (passive geometric compensation) accordingly, the thermal deformation problem caused by non-uniform temperature rise of the mechanical head 21 of the numerical control grinding machine during long-time continuous machining or variable working condition operation can be effectively overcome. The fusion control of active prediction, source intervention (heat dissipation) and real-time closed-loop compensation (pose adjustment) of thermal-induced geometric error is realized, which significantly improves the machining precision stability, shape precision retention ability and long-time working reliability of the grinding machine under thermal disturbance conditions, optimizes the utilization efficiency of heat dissipation resources (precise heat dissipation instead of global heat dissipation), and finally ensures the stable implementation of high-precision and high-quality grinding machining.

[0064] Further, the three-dimensional temperature field data of the mechanical head 21 obtained by the distributed temperature monitoring unit in step S100 includes:

[0065] In step S110, the discrete temperature measurement values of the intersection points of different heat conduction paths inside the mechanical head 21 are collected.

[0066] Distributed temperature monitoring units (such as micro-thermocouples or thermal resistors) are preferentially deployed at the intersection of multiple key heat conduction paths inside the mechanical head 21 (such as structural rib plate junctions, bearing seat mounting surfaces, spindle box and rail junctions, etc.). These intersection points are critical nodes of heat flow transmission, and their temperature changes can significantly reflect the overall heat transfer state of the mechanical head 21. The monitoring units synchronously collect real-time temperature raw data from these discrete positions at a millisecond-level sampling frequency, forming an efficient characterization of the core thermal state inside the mechanical head 21. This targeted distribution method can capture the main heat sources and heat conduction bottleneck areas of the dominant thermal deformation with fewer sensors compared to uniform distribution.

[0067] Step S120, based on the three-dimensional heat conduction model of the mechanical head 21, the spatial field reconstruction of the discrete temperature measurement values is carried out, and the three-dimensional heat conduction model includes the thermal expansion coefficient and the specific heat capacity parameters of each region of the mechanical head 21. The discrete point temperature is mapped to a continuous distributed three-dimensional temperature field by a spatial weighted interpolation algorithm.

[0068] The preset three-dimensional heat conduction physical model of the mechanical head 21 is called, which accurately constructs the geometric topological structure of the mechanical head 21 and embeds the actual material properties of each region (such as the thermal expansion coefficient, specific heat capacity, and thermal conductivity of cast iron matrix, steel shaft, and composite material shell). Using this model, the spatial weighted interpolation algorithm (such as interpolation based on heat conduction equation Green function or distance inverse weighted considering material properties) is applied to the discrete point temperature data obtained in step S110. The interpolation weight not only depends on the spatial distance of the measurement point, but more importantly, it is dynamically adjusted according to the model predicted heat flow path direction and material thermal conductivity. Finally, the sparse discrete point data is reconstructed into a continuous three-dimensional temperature field distribution map covering the entire mechanical head 21 volume, with clear physical meaning, directly showing the heat accumulation area, temperature gradient, and isothermal surface morphology.

[0069] Step S130, dynamically compensate the three-dimensional temperature field combined with the real-time motion state of the mechanical head 21, when the mechanical head 21 performs high-speed rotation or direction changing action, according to the preset inertia temperature rise compensation coefficient, correct the temperature measurement deviation caused by air flow disturbance, generate three-dimensional dynamic temperature field data synchronized in time and space.

[0070] Real-time receive CNC system mechanical head 21 motion state parameters (such as spindle speed, feed axis acceleration, motion direction). When detecting high-speed rotation or sharp acceleration and deceleration, the forced airflow disturbance caused by motion will interfere with the local microclimate around the sensor, causing the temperature measurement value to deviate from the true structure temperature (such as airflow heat dissipation causing the reading to be low). For this purpose, the pre-calibrated inertial temperature rise compensation coefficient matrix (which is established by wind tunnel test or CFD simulation for different motion modes) is called, and the airflow disturbance correction amount of each sensor position is dynamically calculated according to the current motion parameters. Superimpose this correction amount to the corresponding original temperature value in real time, eliminate the dynamic temperature deviation caused by motion, and ensure that the reconstructed three-dimensional temperature field data is strictly synchronized in time and space with the true thermal state and motion state of the mechanical head 21.

[0071] Step S140, denoising processing is performed on the three-dimensional dynamic temperature field data, and an adaptive Kalman filter is used to eliminate electromagnetic interference and mechanical vibration noise, and three-dimensional temperature field data matching the control period is output.

[0072] The reconstructed and compensated three-dimensional temperature field data may still contain high-frequency noise (mainly from electromagnetic interference of the driving system, cutting vibration transmission, signal transmission cross talk). An adaptive Kalman filter is used for processing: the filter takes the heat conduction model as the state prediction basis, and dynamically adjusts the filter gain according to the real-time noise statistical characteristics (estimated online through historical sensor data). It can effectively distinguish between real temperature trends and sudden noise pulses, and suppress electromagnetic interference peaks and signal jitter caused by vibration. Finally, a smooth, stable and strictly synchronized three-dimensional temperature field data stream with the control period (such as 1-10ms) is output, providing high-quality, low-latency input for subsequent thermal deformation prediction.

[0073] Through the fourfold technology synergy of optimized distribution point capture of key thermal nodes, spatial field reconstruction based on physical model, dynamic compensation of motion conditions, and adaptive filtering and denoising, high-fidelity, strong robustness, and full-dynamic real-time perception of the three-dimensional temperature field of the CNC grinding machine mechanical head 21 are realized, solving the accuracy, integrity, and synchronization problems of temperature monitoring under complex motion and harsh conditions, providing a solid and reliable temperature field input basis for subsequent thermal deformation prediction, thereby ensuring the effectiveness and precision of the entire thermal deformation compensation link, and serving as a bottom support link for improving high-precision grinding process stability.

[0074] Further, in step S130, the three-dimensional temperature field is dynamically compensated in combination with the real-time motion state of the mechanical head 21. When the mechanical head 21 performs high-speed rotation or direction-changing action, the temperature measurement deviation caused by airflow disturbance is corrected according to the pre-set inertial temperature rise compensation coefficient, and time and space synchronized three-dimensional dynamic temperature field data is generated, including:

[0075] Step S131, obtain the real-time motion parameters of the mechanical head 21, which include the current speed value and the acceleration direction, call the compensation coefficient matrix pre-stored in the numerical control system to match the current motion state, and obtain the compensation coefficient corresponding to the current motion state.

[0076] The key motion parameters of the mechanical head 21 are read in real time through the numerical control system bus, including the current speed value (in rpm or m / s) of the spindle or linear axis and the instantaneous acceleration vector direction (size and spatial orientation) of each motion component. Based on this parameter combination, access the compensation coefficient matrix pre-stored in the non-volatile memory of the numerical control system (this matrix is generated by calibrating different speed stages and typical acceleration direction combinations through previous wind tunnel tests or computational fluid dynamics simulation). Through table lookup or interpolation algorithm, quickly match the airflow disturbance compensation coefficient (may include basic compensation amount and direction correction factor) corresponding to the current motion state, which quantitatively represents the expected deviation degree of the current motion intensity on the sensor temperature measurement.

[0077] Step S132, based on the matched compensation coefficient, dynamically correct the three-dimensional temperature field, and the compensation coefficient increases linearly with the increase of the speed, and the direction weight factor is adjusted when the acceleration direction changes.

[0078] Using the compensation coefficient obtained in step S131, the temperature value of the sensor corresponding area in the three-dimensional temperature field affected by the airflow disturbance is corrected. The correction process follows two core rules: first, the linear correction dominated by speed: the compensation amount increases by a preset proportion linearly with the increase of speed (for example, the compensation amount increases by 0.5℃ for every 1000rpm increase in speed), reflecting the physical law that the forced convection heat dissipation effect increases with the increase of speed. The second, the nonlinear direction correction triggered by acceleration: when the acceleration direction changes suddenly (such as sudden stop, reverse or turning), the direction weight factor (calculated according to the spatial orientation relationship between the acceleration vector and the sensor position) is activated, and the basic compensation amount is adjusted (such as the compensation amount of the windward surface sensor is increased, and the leeward surface sensor is reduced), to accurately simulate the impact of transient airflow field changes on local microclimate. Finally, the three-dimensional temperature field after motion disturbance compensation is generated.

[0079] Step S133, based on the thermal conduction delay parameters of the mechanical head 21 material, perform time axis alignment processing on the dynamically corrected three-dimensional temperature field, so that the temperature field data and the actual physical state of the mechanical head 21 are kept in space-time synchronization.

[0080] Because the materials of different parts of the mechanical head 21 (such as the cast iron base, steel guide rails, and copper alloy bearings) have different thermal conductivity rates and thermal capacities, there is a measurable physical delay in the propagation of heat within the structure. A predefined table of material thermal conduction delay parameters (obtained through material thermophysical property tests) is used to calculate the theoretical thermal conduction delay time relative to the core heat source for each voxel (or region) in the three-dimensional temperature field. This delay is used as a time offset to perform a reverse translation compensation on the time axis of the temperature field data corrected in step S132 (i.e., "advancing the temperature data to the moment when heat has just arrived at that location"). After this processing, the final output three-dimensional dynamic temperature field data accurately reflects the true thermal state of the mechanical head 21 structure at the current moment in the time dimension and retains an accurate temperature distribution in the spatial dimension, achieving a spatiotemporal consistency expression that is strictly synchronized with the physical entity.

[0081] A precise temperature field correction system for high-speed dynamic conditions was constructed by real-time matching of compensation coefficients driven by motion parameters, dynamic correction of the temperature field under the combined action of rotation speed and acceleration, and time axis alignment processing for material thermal delay characteristics compensation. Essentially, this system eliminates temperature measurement distortion caused by motion-induced airflow disturbances and data lag caused by material thermal inertia, ensuring that the three-dimensional temperature field in virtual space remains strictly synchronized with the actual physical thermal state of the mechanical head 21 on a spatiotemporal scale. This provides accurate, timely, and reliable thermal state input for subsequent thermal deformation prediction under severe motion conditions, fundamentally guaranteeing the response accuracy and robustness of the dynamic thermal compensation system.

[0082] Further, in step S200, the three-dimensional temperature field data is input into the CNC grinding machine's thermal deformation digital twin model. Through coupled temperature field-stress field-deformation field simulation calculations, the thermal deformation amount and spatial distribution of the mechanical head 21 in the future time period are predicted, including:

[0083] Step S210: Initialize the material property database of the digital twin model of thermal deformation of CNC grinding machine. The material property database pre-stores the thermal conductivity, elastic modulus and thermal expansion coefficient of each component of the mechanical head 21, and constructs transient temperature boundary conditions by associating real-time three-dimensional temperature field data.

[0084] Upon startup, the system loads pre-stored material property parameters of the mechanical head 21 from the database, including the precise thermal conductivity (determining the heat transfer rate), elastic modulus (reflecting the material's resistance to deformation), and coefficient of thermal expansion (characterizing the rate of dimensional change caused by temperature variations) of each component. Based on the real-time three-dimensional temperature field data generated in step S100, it is converted into the transient temperature boundary conditions required for simulation: the temperature field data is mapped to the geometric surface and internal key nodes of the mechanical head 21, serving as the initial temperature distribution input for the simulation at the current moment. This process ensures that the digital twin model and the physical entity are strictly consistent in their initial thermal conditions.

[0085] Step S220, based on the geometric topology of the mechanical head 21, dynamically divide the non-uniform grid, increase the grid node density in the area of the heat sink 208 moving path, map the transient temperature boundary conditions to the grid nodes to generate the initial temperature field distribution.

[0086] According to the three-dimensional geometric topology of the mechanical head 21 (such as complex cavity, thin wall, shaft hole feature), the non-uniform finite element grid is automatically generated. In order to balance the calculation accuracy and real-time demand, the grid is locally encrypted (the node density is significantly improved) in the heat deformation sensitive area (such as the main shaft bearing seat) and the area covered by the moving track of the heat sink 208, so as to accurately capture the local temperature gradient change under the action of the heat sink 208. Then, the transient temperature boundary conditions constructed in step S210 are mapped to each grid node through shape function interpolation to form the initial temperature field spatial distribution of digital twin simulation.

[0087] Step S230, call the coupled solver to synchronously iterate and calculate the temperature field and the stress field, introduce the position parameters of the heat sink 208 as the convective heat transfer boundary in the temperature field calculation, and calculate the thermal stress distribution inside the mechanical head 21 according to the temperature gradient and material properties in the stress field calculation.

[0088] Call the temperature-stress field strong coupling solver to perform synchronous iteration calculation. Temperature field evolution calculation: based on the heat conduction equation, combined with the current grid temperature distribution, material thermal conductivity and external heat dissipation conditions to promote calculation. Among them, the real-time position parameters of the heat sink 208 are dynamically converted into the enhanced convective heat transfer boundary conditions (such as improving the local heat transfer coefficient) of the corresponding grid area, accurately simulating the influence of the directional heat dissipation of the heat sink 208 on the temperature field. Real-time stress field solution: using the updated temperature field data of the last step, according to the spatial temperature gradient distribution and the thermal expansion coefficient and elastic modulus of each region material, the internal thermal stress tensor distribution caused by non-uniform thermal expansion is calculated through the thermoelastic constitutive equation. Two fields are alternately iterated in each time step until convergence, ensuring the strong coupling of the physical process.

[0089] Step S240, map the thermal stress distribution to the deformation variable through the deformation field conversion model, correct the deformation direction combined with the kinematic constraint conditions, and output the three-dimensional deformation prediction value of the key points on the surface of the mechanical head 21.

[0090] The internal thermal stress field calculated in step S230 is converted into a theoretical elastic deformation field of the mechanical head 21 structure by a deformation field conversion model (based on small deformation elastic theory). Subsequently, the kinematic constraint conditions of the mechanical head 21 in the actual assembly (such as fixed bolt constraints, guide rail sliding pair freedom limitation, bearing pre-tightening state) are introduced, and the deformation vector calculated by pure stress is corrected by constraint: only the displacement of the allowed freedom is released, and the deformation component that violates the actual constraint is suppressed. Finally, the three-dimensional deformation prediction value (including X / Y / Z direction displacement) of the key measurement points (such as the spindle end face and the grinding wheel mounting flange) on the surface of the mechanical head 21 at the future time is output.

[0091] In step S250, the three-dimensional deformation prediction value is spatially interpolated and reconstructed to generate continuous distribution of thermal deformation and spatial distribution data, and the time resolution matches the control period.

[0092] For the discrete key point deformation prediction value output in step S240, a deformation field spatial interpolation algorithm (such as radial basis function interpolation or thin plate spline interpolation) based on the CAD model of the mechanical head 21 is used to reconstruct a continuous three-dimensional thermal deformation distribution field covering the entire surface of the mechanical head 21. The field data clearly represents the thermal deformation size, direction and spatial variation trend of each part of the mechanical head 21 at the predicted time. At the same time, the simulation time step strictly matches the motion control period of the numerical control system (such as 1-10ms), ensuring that the deformation prediction data stream output is synchronized with the control system real-time decision rhythm.

[0093] Through accurate modeling of material properties, adaptive meshing, real-time simulation of temperature-stress strong coupling, constraint correction of deformation conversion and spatial field reconstruction, a high-fidelity digital twin prediction engine for thermal deformation of the mechanical head 21 of the numerical control grinding machine is constructed. The essence is to use embedded multi-physical field simulation technology to complete the closed-loop mapping from real-time temperature field to future thermal deformation field within the time constraint of the control system, accurately quantify the deformation evolution process under the intervention of the heat sink 208, and provide advanced and reliable spatial deformation prediction data for subsequent dynamic heat compensation and geometric pose correction, laying the foundation for the decision of the thermal error active suppression strategy.

[0094] Further, after the spatial interpolation and reconstruction of the three-dimensional deformation prediction value in step S250, it further includes:

[0095] In step S261, the actual deformation of the key points on the surface of the mechanical head 21 is collected in real time by the laser displacement sensor installed on the side wall of the mechanical head 21, and a residual sequence is constructed with the same position deformation prediction value.

[0096] At the preset key point positions of the mechanically deformed sensitive areas of the machine head 21 (such as the spindle box side wall and the guide rail mounting base), an array of high-precision non-contact laser displacement sensors is installed. These sensors measure the three-dimensional spatial position offset (i.e., the actual thermal deformation) at the key points in real time at a sampling frequency synchronized with the control cycle (e.g., 1 kHz). The actual deformation at the same position and the same time is compared with the deformation prediction value output by step S250, and the difference is calculated to generate a residual sequence dataset arranged in chronological order. This sequence dynamically records the deviation characteristics of the prediction accuracy of the digital twin model in the time and space dimensions.

[0097] Step S262, when the sliding variance of the residual sequence exceeds the tolerance threshold, based on the digital twin model parameter self-correction module, the thermal expansion coefficient in the material property database is optimized in reverse based on the residual distribution.

[0098] Continuously analyze the statistical characteristics of the residual sequence: use a sliding time window (e.g., the past 10 seconds) to calculate the variance of the residual, and when the variance continuously exceeds the preset tolerance threshold (indicating that the prediction deviation is systematically divergent), activate the parameter self-correction module of the digital twin model. According to the spatial distribution pattern of the residual (e.g., the residual of a specific area is significantly larger) and the time evolution characteristics, combined with the thermal load history of the machine head 21, the most likely parameter item in the material property database that causes the deviation is deduced in reverse through an inverse problem solving algorithm (such as the gradient descent method or genetic algorithm) - usually the thermal expansion coefficient of the local area. Automatically generate optimized thermal expansion coefficient values and update the corresponding items in the material property database to ensure that the model parameters approach the actual material behavior.

[0099] Step S263, use the optimized thermal expansion coefficient as a new input for thermal stress distribution calculation, and iteratively execute the coupled simulation calculation steps until the residual sequence returns to the tolerance range.

[0100] Use the thermal expansion coefficient updated in step S262 as a new parameter to restart the coupled simulation calculation process of steps S210-S250: based on the current real-time temperature field data, use the corrected material properties to recalculate the thermal stress distribution and deformation prediction value. Then, compare the new prediction value with the actual measurement value of the laser sensor to generate an updated residual sequence. If the residual variance is still out of limits, repeat the parameter optimization and simulation iteration; if the residual falls within the tolerance range, output the deformation prediction result under the current parameters. This process forms a closed loop of "measurement-comparison-correction-re-simulation" until the prediction accuracy is restored to stability.

[0101] A self-calibration mechanism of the digital twin model is constructed through real-time residual analysis of the measured and predicted deformation values by laser displacement sensors, reverse optimization of the thermal expansion coefficient triggered by the super-difference, and closed-loop iterative re-simulation. The essence is to dynamically correct the key material parameters of the simulation model using deformation feedback data from the physical world, overcoming the model prediction drift problem caused by material aging, batch differences, or sudden changes in working conditions. This enables the digital twin system to have self-learning ability to continuously adapt to the thermodynamic characteristics of the actual mechanical head 21, significantly improving the long-term stability and working condition robustness of thermal deformation prediction, and providing bottom model reliability guarantee for high-precision thermal compensation.

[0102] Further, the compensation strategy generated according to the thermal deformation amount and the spatial distribution in step S300 drives the heat sink 208 to move to the thermal deformation sensitive area for directional heat dissipation, including:

[0103] In step S311, the area where the temperature gradient exceeds the preset threshold in the spatial distribution is identified as the thermal deformation sensitive area, and the shortest straight line movement path from the mechanical head 21 to the heat sink 208 is calculated based on the spatial coordinates of the thermal deformation sensitive area.

[0104] The thermal deformation amount and spatial distribution data output by step S200 are analyzed, and the temperature gradient field (temperature change rate per unit distance) is extracted. When the temperature gradient value of a certain continuous area continuously exceeds the preset engineering threshold (which is calibrated by thermal deformation experiment), it is determined that this area is the thermal deformation sensitive area at the current time. The three-dimensional spatial coordinates of the geometric center point or the highest temperature point of the area are extracted, and the shortest straight line motion trajectory (considering mechanical interference avoidance) from the current position of the heat sink 208 to the target coordinates is calculated in real time by combining the straight line guide rail constraint condition of the heat sink 208 on the mechanical head 21, and the corresponding axial displacement instruction sequence is generated.

[0105] In step S312, the heat dissipation intensity control instruction is generated according to the thermal deformation amount. When the thermal deformation amount reaches the first critical value, the high-speed operation mode of the fan of the heat sink 208 is triggered, and the rack driving pulse signal is generated at the thermal deformation sensitive area coordinates.

[0106] According to the predicted absolute value of the thermal deformation amount (such as the spindle radial runout prediction value), the heat dissipation intensity is set. If the thermal deformation amount is less than the first critical value (such as 5 μm), the basic speed of the fan of the heat sink 208 is maintained; when the thermal deformation amount reaches or exceeds the first critical value, the high-speed operation mode (such as 120% of the rated power) of the fan is triggered, which significantly improves the forced convection heat dissipation capacity. The rack driving pulse signal (which contains the moving direction, step pulse number and frequency) is generated at the target position coordinates calculated in step S311, which drives the heat sink 208 to move accurately along the straight line guide rail to the thermal deformation sensitive area above.

[0107] Step S313, the rack drive pulse signal is coupled with the real-time motion parameters of the mechanical head 21, and when the mechanical head 21 is in a high-speed rotating state, the residence time of the heat sink 208 is prolonged to ensure that the directional heat dissipation action is synchronized with the mechanical motion state.

[0108] The motion state parameters (such as spindle speed, feed axis acceleration) of the mechanical head 21 are obtained in real time. Before sending the rack drive pulse signal, the motion-heat dissipation coupling check is performed: if it is detected that the mechanical head 21 is in a high-speed rotating state (such as the spindle > 8000 rpm), the residence time of the heat sink 208 at the target position is actively prolonged (such as prolonged by 50%). This considers that the strong centrifugal airflow caused by high-speed rotation will weaken the heat dissipation effect, and the heat dissipation time needs to be prolonged to ensure sufficient heat exchange. At the same time, the timing window of the pulse signal is dynamically adjusted to ensure that the heat sink 208 moving action is completed in the mechanical head 21 acceleration or deceleration gap or motion stable section, avoiding vibration interference and motion interference.

[0109] Through the sensitive area recognition driven by the temperature gradient, the hierarchical heat dissipation strategy controlled by the heat deformation threshold, and the heat dissipation timing optimization adapted to the motion state, an accurate and efficient directional heat compensation execution system is constructed. The essence is to convert the predicted heat deformation field into a spatial positioning accurate heat dissipation action, and through dynamic intensity adjustment and motion synchronization mechanism, ensure that the heat sink 208 applies heat dissipation intervention matching the heat deformation severity at the correct position and at the correct time, actively suppresses the temperature rise rate in the key area from the heat source end, reduces the burden for subsequent geometric pose compensation, and forms a closed-loop control chain for heat-machine error cooperative suppression.

[0110] Further, the shortest straight line moving path from the mechanical head 21 to the heat sink 208 in step S311 based on the spatial coordinates of the heat deformation sensitive area includes:

[0111] Step S311a, according to the displacement difference of the center point of the heat deformation sensitive area in the adjacent control period divided by the time interval, the migration rate is obtained, and when the migration rate exceeds the preset threshold, the dynamic path planning mode is started.

[0112] The three-dimensional coordinate change of the center point of the heat deformation sensitive area in the adjacent control period (such as an interval of 10ms) is tracked in real time, and the spatial migration rate (unit: mm / s) of the area is accurately calculated by the displacement difference divided by the time interval. When the migration rate continuously exceeds the preset threshold (the threshold is calibrated according to the maximum tracking ability of the heat sink 208), it indicates that the heat sensitive area is moving quickly due to the change of working condition. At this time, the default static path planning is automatically switched to the dynamic path planning mode to cope with the challenge of real-time drift of the target position.

[0113] Step S311b, an optimization objective function of the heat dissipation device 208 movement energy consumption and heat compensation efficiency is constructed, the temperature gradient descending rate is taken as a weight factor, and a gradient descent algorithm is used to solve a curve motion path with the lowest energy consumption and the optimal compensation efficiency.

[0114] In the dynamic programming mode, a double objective function is established with the minimization of the heat dissipation device 208 movement energy consumption and the maximization of the heat compensation efficiency as the core. The heat compensation efficiency is quantified by the temperature gradient descending rate of the target area (gradient reduction value per unit time), which is taken as a weight factor to dynamically adjust the double objective weight. The movement energy consumption is calculated according to the path length, acceleration change and the mass of the heat dissipation device 208. A gradient descent algorithm is used to iteratively search in the geometric constraint space of the mechanical head 21 to solve an approximate optimal curve path that can not only approach the migration target point with a smooth curve (to reduce the energy consumption of sudden start and stop) but also make the heat dissipation device 208 cover the high weight area as soon as possible, so as to achieve the engineering balance of energy consumption and compensation effect.

[0115] Step S311c, the curve motion path is discretized into a rack driving pulse sequence, and the pulse interval time is dynamically adjusted according to the real-time rotating speed of the mechanical head 21, so that the movement speed of the heat dissipation device 208 is synchronized with the migration of the heat-sensitive area.

[0116] The continuous curve path generated by optimization is discretized into a step-like path point sequence according to the control period, and then converted into a pulse signal sequence for driving the linear rack (the number of pulses corresponds to the displacement, and the frequency corresponds to the movement speed). Key, the pulse interval time (which determines the movement speed of the heat dissipation device 208) is not a fixed value, but is dynamically adjusted according to the real-time rotating speed of the mechanical head 21: when the rotating speed increases, the pulse interval is shortened (to accelerate the movement of the heat dissipation device 208), and when the rotating speed decreases, the pulse interval is increased (to slow down the movement). This ensures that the movement speed of the heat dissipation device 208 is real-time matched with the spatial migration rate of the heat-sensitive area, and the continuous alignment of the heat dissipation focus and the heat source is maintained.

[0117] Through the heat-sensitive area migration rate monitoring, the energy consumption-efficiency double objective optimization path planning, and the rotating speed adaptive pulse sequence generation, a real-time tracking system of the heat dissipation device 208 for dynamic heat sources is constructed. The essence is to break through the limitation of traditional static positioning, so that the heat dissipation device 208 can intelligently adjust the movement trajectory and speed according to the spatio-temporal evolution law of the thermal deformation field, reduce the invalid energy consumption, and ensure that the directional heat dissipation action always accurately acts on the rapidly migrating heat-sensitive core area, thereby significantly improving the timeliness and effectiveness of heat compensation.

[0118] Further, the adjustment of the pose of the polishing head 22 by the five-axis linkage system of the numerical control grinding machine in step S300 to offset the predicted thermal deformation amount includes:

[0119] Step S321, based on the thermal deformation amount and the spatial distribution, a compensation vector of the polishing head 22 pose is constructed, and the thermal deformation amount is converted into a three-dimensional offset in the working coordinate system of the polishing head 22 through a deformation space mapping model. The three-dimensional offset includes a linear displacement compensation value and an angle deflection compensation value.

[0120] The thermal deformation amount and the spatial distribution data output by the analysis step S200 are analyzed, and the deformation vector of the key support structure (such as the spindle box and the guide rail) of the mechanical head 21 is extracted. Through a predefined deformation space mapping model (which establishes a rigid body transformation chain from the base coordinate system of the mechanical head 21 to the working coordinate system of the polishing head 22, including the geometric transmission relationship of each motion axis), the thermal deformation amount is converted into a six-degree-of-freedom pose compensation vector with the polishing head 22 end as the reference point. The vector explicitly includes a three-dimensional linear displacement compensation value (X / Y / Z direction translation amount, unit: μm) and a three-dimensional angle deflection compensation value (micro-rotation amount around the X / Y / Z axis, unit: μrad), which accurately quantifies the spatial attitude adjustment of the polishing head 22 required to offset the influence of thermal deformation.

[0121] Step S322, dynamically couple the three-dimensional offset with the real-time motion parameters, and calculate the centrifugal force deformation effect value according to the real-time rotation speed value and the mass center position parameters of the mechanical head 21 when the mechanical head 21 rotates at high speed, correct the angle deflection compensation value, generate polishing head 22 pose adjustment instructions and drive servo motor to execute compensation action synchronously.

[0122] The three-dimensional offset calculated in step S321 is dynamically fused with the real-time motion parameters (spindle speed, feed speed) of the numerical control system, and the centrifugal force coupling correction is performed: when a high-speed rotating working condition (such as spindle >6000rpm) is detected, based on the real-time rotation speed value and the pre-stored mass center position parameters (including rotational inertia) of the mechanical head 21, the additional structural deformation amount (mainly manifested as angle deflection) caused by rotation is calculated through the centrifugal force deformation empirical model. Add this additional amount to the original angle deflection compensation value to correct the prediction deviation caused by high-speed centrifugal effect. Instruction generation and execution: convert the corrected linear displacement and angle deflection compensation values into motion increment instructions of each axis of the five-axis linkage system (X / Y / Z linear axis compensation displacement, A / C rotation axis compensation rotation angle) according to the control period. The instructions are issued to the axis servo drivers through the real-time bus, and the motor drives the polishing head 22 pose to be adjusted synchronously during machining, so that the actual motion trajectory of the polishing head 22 generates an offset which is equal in size and opposite in direction to the predicted thermal deformation, thereby realizing real-time offset of geometric error.

[0123] In addition, as Figure 5As shown, the numerical control grinding machine is also provided with a dust reduction assembly 3, the dust reduction assembly 3 comprising a dust collector 31, a stretch pipe 32, an air outlet 33, a rubber band 34 and a dust filter bag 35; the dust collector 31 is connected with the air outlet 33 through the stretch pipe 32, the dust filter bag 35 is fixed at the air outlet 33 through the rubber band 34 and is used for filtering and absorbing dust, the suction force of the dust collector 31 is adjustable to adapt to the dust absorption requirement under different grinding conditions, ensure good dust reduction effect, the dust filter bag 35 is of a detachable structure and is convenient for regular cleaning and replacement, so that the normal working performance of the dust reduction assembly 3 is maintained, the numerical control grinding machine body 1 is provided with a control panel for operating and displaying the running state of the device, thermal deformation compensation parameters and working parameters of the dust reduction assembly 3 and other information, in the dust reduction assembly 3, the dust collector 31 is selected to be an industrial dust collector 31 with large suction force and low noise and is connected with the air outlet 33 through the stretch pipe 32. The stretch pipe 32 has certain flexibility and stretchability and can adapt to different working position requirements. The dust filter bag 35 is made of a material with high filtering efficiency and is tightly fixed at the air outlet 33 through the rubber band 34, so that dust can be effectively filtered.

[0124] Correspondingly, a second aspect of the embodiment of the present application provides an electronic device, comprising: at least one processor, and a memory connected with the at least one processor. Wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to make the at least one processor execute the dynamic thermal deformation compensation method of the numerical control grinding machine.

[0125] Correspondingly, a fourth aspect of the embodiment of the present application provides a computer readable storage medium, which stores computer instructions, and the instructions are executed by a processor to implement the dynamic thermal deformation compensation method of the numerical control grinding machine.

[0126] The embodiment of the present application aims to protect a dynamic thermal deformation compensation method of a numerical control grinding machine, and has the following effects:

[0127] 1. Through the embedded digital twin model deep coupling of temperature field-stress field-deformation field simulation, combined with the non-uniform dynamic mesh technology to locally encrypt the calculation nodes in the heat dissipation path area, and introducing the radiator position parameter as the convective heat transfer boundary condition, the high-fidelity quantization of thermal stress distribution is realized; simultaneously, the residual feedback of the measured deformation and the predicted value is used to online reverse optimize the material thermal expansion coefficient, the model precision decay problem caused by long-term operation is solved, and finally the time and space synchronous deformation prediction data is output in the control cycle of the numerical control system, the thermal deformation prediction error is reduced by more than 60%, and the disconnection bottleneck of traditional offline model and real-time control is broken through;

[0128] 2. Based on the real-time motion parameters (rotation speed / acceleration) of the mechanical head, dynamically call the inertia temperature rise compensation coefficient, correct the temperature deviation caused by airflow disturbance in high-speed variable direction working condition, generate a three-dimensional dynamic temperature field with time and space synchronization; then through the migration rate prediction of the heat sensitive area, take the temperature gradient decline rate as the weight factor to plan the heat sink curve path with optimal energy consumption, and discretize the path into pulse sequence to dynamically adjust the moving rhythm, realize the real-time tracking of the heat dissipation position and heat source migration; at the same time, according to the threshold value of thermal deformation, link the control of heat dissipation intensity and residence time, prolong the action time of the heat sink in high-speed rotation, improve the overall heat dissipation efficiency by 40%, and completely solve the disconnection problem between traditional fixed heat dissipation mode and dynamic migration of heat source;

[0129] 3. Through the deformation space mapping model, convert the thermal deformation into three-dimensional offset (including linear displacement and angle deflection) in the working coordinate system of the polishing head, and dynamically calculate the centrifugal force deformation effect quantity by fusing the real-time motion parameters, correct the angle compensation value in high-speed rotation working condition; synchronize the movement of the heat sink and the five-axis linkage system, execute the pose adjustment instruction by the servo motor in the same control cycle, make the heat dissipation action and pose compensation cooperate to offset the thermal deformation and centrifugal force effect, reduce the size tolerance risk in complex trajectory machining by 70%, realize the global dynamic cooperation of thermal deformation suppression and kinematic effect.

[0130] Those skilled in the art will understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0131] The present application is described with reference to flowcharts and / or block diagrams according to the methods, devices (systems), and computer program products of the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one flow or multiple flows and / or blocks Figure 1 The device that implements the functions specified in one flow or multiple flows and / or blocks.

[0132] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the flow Figure 1 one or more flows and / or blocks Figure 1 one or more blocks or multiple blocks.

[0133] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flow Figure 1 one or more flows and / or blocks Figure 1 one or more blocks or multiple blocks.

[0134] Finally, it should be noted that the above-mentioned embodiments are merely used to illustrate the technical solutions of the present application, rather than limiting the same. Even though the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that the specific embodiments of the present application can be modified or replaced equivalently, and any modification or replacement without departing from the spirit and scope of the present application should be covered within the protection scope of the claims of the present application.

Claims

1. A method for dynamic thermal deformation compensation of a CNC grinding machine, characterized in that, The CNC grinding machine includes: distributed temperature monitoring units disposed at several preset locations on the machine head and heat sinks disposed at corresponding positions on the machine head, wherein the heat sinks are capable of linear movement, and the compensation method includes the following steps: The three-dimensional temperature field data of the mechanical head is acquired in real time through the distributed temperature monitoring unit; The three-dimensional temperature field data is input into the digital twin model of thermal deformation of the CNC grinding machine. Through coupled temperature field-stress field-deformation field simulation calculation, the amount of thermal deformation and spatial distribution of the mechanical head in the future time period is predicted. Based on the thermal deformation amount and spatial distribution, a compensation strategy is generated to drive the radiator to move to the thermal deformation sensitive area for directional heat dissipation, and the position of the grinding head is adjusted through the five-axis linkage system of the CNC grinding machine to offset the predicted thermal deformation amount. The digital twin model of the grinding machine's thermal deformation is a real-time simulation module embedded in the control system of the CNC grinding machine.

2. The dynamic thermal deformation compensation method for CNC grinding machines according to claim 1, characterized in that, The real-time acquisition of three-dimensional temperature field data of the mechanical head through the distributed temperature monitoring unit includes: Collect discrete temperature measurements at the intersection points of different heat conduction paths inside the mechanical head; The discrete temperature measurements are reconstructed using a three-dimensional heat conduction model of the mechanical head. The three-dimensional heat conduction model includes the thermal expansion coefficient and specific heat capacity parameters of the materials in each region of the mechanical head. The discrete point temperature is mapped to a continuously distributed three-dimensional temperature field through a spatial weighted interpolation algorithm. The three-dimensional temperature field is dynamically compensated by combining the real-time motion state of the mechanical head. When the mechanical head performs high-speed rotation or change of direction, the temperature measurement deviation caused by airflow disturbance is corrected according to the preset inertial temperature rise compensation coefficient, and the three-dimensional dynamic temperature field data synchronized in time and space is generated. The three-dimensional dynamic temperature field data is subjected to noise reduction processing. An adaptive Kalman filter is used to eliminate electromagnetic interference and mechanical vibration noise, and the output is three-dimensional temperature field data that matches the control cycle.

3. The dynamic thermal deformation compensation method for CNC grinding machines according to claim 2, characterized in that, The method combines the real-time motion state of the mechanical head to dynamically compensate for the three-dimensional temperature field. When the mechanical head performs high-speed rotation or change of direction, it corrects the temperature measurement deviation caused by airflow disturbance according to the preset inertial temperature rise compensation coefficient, and generates spatiotemporally synchronized three-dimensional dynamic temperature field data, including: The real-time motion parameters of the mechanical head are obtained, including the current rotational speed and acceleration direction. The compensation coefficient matrix pre-stored in the CNC system is called to match the current motion state, and the compensation coefficient corresponding to the current motion state is obtained. The three-dimensional temperature field is dynamically corrected based on the matching compensation coefficient. The compensation coefficient increases linearly with the rotation speed and the direction weight factor is triggered to adjust the correction amount when the acceleration direction changes. The time axis of the dynamically corrected three-dimensional temperature field is aligned based on the thermal conduction delay parameter of the mechanical head material, so that the temperature field data is kept in spatiotemporal synchronization with the actual physical state of the mechanical head.

4. The dynamic thermal deformation compensation method for CNC grinding machines according to claim 3, characterized in that, The process of inputting the three-dimensional temperature field data into the digital twin model of the CNC grinding machine's thermal deformation, and predicting the amount and spatial distribution of the mechanical head's thermal deformation over a future time period through coupled temperature field-stress field-deformation field simulation calculations, includes: The material property database of the CNC grinding machine thermal deformation digital twin model is initialized. The material property database pre-stores the thermal conductivity, elastic modulus and thermal expansion coefficient of each component of the mechanical head, and constructs transient temperature boundary conditions by associating real-time three-dimensional temperature field data. Based on the dynamic division of non-uniform meshes according to the geometric topology of the mechanical head, the mesh node density is increased in the radiator moving path region, and the transient temperature boundary conditions are mapped to the mesh nodes to generate the initial temperature field distribution. The coupled solver is invoked to synchronously iterate and calculate the temperature field and stress field. The temperature field calculation introduces the heat sink position parameters as the convection heat transfer boundary, and the stress field calculation calculates the thermal stress distribution inside the mechanical head based on the temperature gradient and material properties. The thermal stress distribution is mapped to deformation by the deformation field conversion model, and the deformation direction is corrected by kinematic constraints. The three-dimensional deformation prediction value of key points on the surface of the mechanical head is then output. The predicted three-dimensional deformation is reconstructed by spatial interpolation to generate continuously distributed thermal deformation and spatial distribution data, the temporal resolution of which matches the control cycle.

5. The dynamic thermal deformation compensation method for CNC grinding machines according to claim 4, characterized in that, After spatial interpolating and reconstructing the predicted three-dimensional deformation values, the method further includes: By using a laser displacement sensor installed on the side wall of the mechanical head, the actual deformation of key points on the surface of the mechanical head is collected in real time, and a residual sequence is constructed with the deformation prediction value at the same position. When the sliding variance of the residual sequence exceeds the tolerance threshold, the thermal expansion coefficient in the material property database is optimized in reverse based on the digital twin model parameter self-correction module and the residual distribution. The optimized coefficient of thermal expansion is used as a new input for calculating the thermal stress distribution. The coupled simulation calculation steps are executed iteratively until the residual sequence is restored to the tolerance range.

6. The dynamic thermal deformation compensation method for CNC grinding machines according to claim 3, characterized in that, The step of generating a compensation strategy based on the thermal deformation amount and spatial distribution, and driving the heat sink to move to the thermal deformation-sensitive area for directional heat dissipation, includes: The region in the spatial distribution whose temperature gradient exceeds a preset threshold is identified as the thermal deformation sensitive region, and the shortest straight-line movement path from the mechanical head to the heat sink is calculated based on the spatial coordinates of the thermal deformation sensitive region. A heat dissipation intensity control command is generated based on the thermal deformation amount. When the thermal deformation amount reaches the first critical value, the high-speed operation mode of the radiator fan is triggered, and a rack drive pulse signal is generated at the coordinates of the thermal deformation sensitive area. The rack drive pulse signal is coupled and verified with the real-time motion parameters of the mechanical head. When the mechanical head is in a high-speed rotation state, the heat sink dwell time is extended to ensure that the directional heat dissipation action is completed synchronously with the mechanical motion state.

7. The dynamic thermal deformation compensation method for CNC grinding machines according to claim 6, characterized in that, The calculation of the shortest straight-line movement path from the mechanical head to the heat sink based on the spatial coordinates of the thermal deformation sensitive area includes: The migration rate is obtained by dividing the displacement difference of the center point of the thermal deformation sensitive area within adjacent control cycles by the time interval. When the migration rate exceeds a preset threshold, the dynamic path planning mode is activated. An optimization objective function for the energy consumption and thermal compensation efficiency of the radiator movement is constructed, with the temperature gradient descent rate as a weighting factor. The gradient descent algorithm is used to solve for the curve movement path with the lowest energy consumption and the best compensation efficiency. The curved motion path is discretized into a rack and pinion drive pulse sequence. The pulse interval is dynamically adjusted according to the real-time rotation speed of the mechanical head, so that the movement speed of the radiator is synchronized with the migration of the heat-sensitive area.

8. The dynamic thermal deformation compensation method for CNC grinding machines according to claim 3, characterized in that, The step of adjusting the position and orientation of the grinding head through the five-axis linkage system of the CNC grinding machine to counteract the predicted thermal deformation includes: The compensation vector for the pose of the grinding head is constructed based on the thermal deformation and spatial distribution. The thermal deformation is converted into a three-dimensional offset in the working coordinate system of the grinding head through the deformation space mapping model. The three-dimensional offset includes linear displacement compensation value and angle deflection compensation value. The three-dimensional offset is dynamically coupled with the real-time motion parameters. When the mechanical head rotates at high speed, the centrifugal force deformation effect is calculated based on the real-time rotation speed value and the mechanical head center of mass position parameter. The angle deflection compensation value is corrected, the grinding head posture adjustment command is generated, and the servo motor is driven to synchronously execute the compensation action.

9. An electronic device, characterized in that, include: At least one processor; And a memory connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the at least one processor to perform the dynamic thermal deformation compensation method for a CNC grinding machine as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores computer instructions that, when executed by a processor, implement the dynamic thermal deformation compensation method for the CNC grinding machine as described in any one of claims 1-7.

Images