A multi-process thermal mechanical deformation monitoring and compensation system for an engine support

Abstract

The application discloses a multi-process thermal mechanical deformation monitoring and compensation system for an engine support, and relates to the technical field of image monitoring. The system first collects visible light and infrared data to generate a fusion image, and identifies a chip blocking state of a workpiece surface. If there is blocking, a physical cleaning device is driven to remove the blocking until a determination logic is met, and a thermal history log containing cleaning process data is generated. Then, the current three-dimensional full-field deformation vector of the workpiece is calculated by combining the recovered pure surface temperature field and the thermal history log. Finally, the displacement and rotation amount of key feature points are extracted to generate machine tool coordinate system space transformation instructions for compensation. The application solves the problems of infrared temperature measurement distortion caused by chip accumulation in the inner cavity under the micro-lubrication working condition, and thermal hysteresis prediction deviation caused by internal heat accumulation after chip removal.

Description

Technical Field

[0001] This invention relates to the field of image monitoring technology, specifically to a multi-process thermomechanical deformation monitoring and compensation system for engine brackets. Background Technology

[0002] As the automotive industry moves towards lightweighting and high performance, engine mounts, as core components of the powertrain, widely utilize aluminum alloy materials and complex topology optimization structures (such as numerous thin walls, deep cavities, and reinforcing ribs). To meet environmental protection and cost control requirements, minimum quantity lubrication (MQL) or dry cutting processes are gradually replacing traditional high-flow wet cutting. However, due to the high coefficient of thermal expansion of the material and the extremely uneven distribution of structural stiffness, the workpiece is prone to nonlinear deformation under the coupled effect of cutting heat and clamping force, making it difficult to guarantee machining accuracy.

[0003] For this type of deformation, existing technologies typically rely on post-hoc detection by machine tool probes or simple compensation based on the principle of linear thermal expansion. These methods largely assume a uniform and static thermal environment, and cannot effectively address the complex dynamic behavior of engine mounts under MQL conditions, particularly exhibiting insurmountable technical shortcomings in the following two aspects: First, there's the issue of random heat source interference (chip accumulation in the internal cavity) caused by structural hot spots. Traditional wet cutting relies on coolant to flush away chips in deep cavities, but under MQL (Multi-Level Liquid) aerosol conditions, the airflow struggles to penetrate the densely packed reinforcing ribs of the support frame. High-temperature chips easily accumulate in the cavity like a bird's nest, forming randomly located, extremely hot localized heating pads that continuously heat the workpiece. Existing monitoring models typically only calculate tool cutting heat and cannot detect these random external heat sources, leading to sudden thermal deformations in the workpiece that the algorithm cannot explain, rendering predictions based on standard models completely ineffective.

[0004] Secondly, there is the problem of asymmetric thermal torsion caused by non-uniform stiffness. Engine mounts typically exhibit an asymmetric structure with a thick flange (high stiffness) at one end and a thin cantilever (low stiffness) at the other. Under the effect of thermal accumulation, its deformation is not a simple linear elongation or widening, but rather the thin-walled arm elongates due to heat but is restrained by the thick flange, causing the cantilever to undergo complex spatial warping similar to a twisted rope. Existing three-axis machine tool compensation systems can only correct translational errors and cannot characterize and correct this spatial torsion.

[0005] Therefore, the present invention provides a multi-process thermomechanical deformation monitoring and compensation system for engine brackets. Summary of the Invention

[0006] The purpose of this invention is to provide a multi-process thermomechanical deformation monitoring and compensation system for engine mounts, so as to solve the existing problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a multi-process thermomechanical deformation monitoring and compensation system for engine brackets, comprising: The multimodal fusion sensing module is configured to acquire visible light image data and infrared thermal imaging data of the processing area in real time, and generate a fused image tensor containing texture information and temperature information based on data fusion processing, thereby identifying the heat distribution state and chip occlusion state of the workpiece surface. The active boundary reconstruction module is configured to drive the physical cleaning device to perform directional cleaning action on the obstructed area when the identification result shows that there is chip obstruction, until the workpiece meets the hot spot attribute determination logic and generates a thermal history log containing cleaning process data. The thermo-mechanical coupling solution module is configured to obtain the pure surface temperature field after the boundary is restored, and calculate the three-dimensional full-field deformation vector of the workpiece at the current moment by combining the thermal history log. The adaptive compensation execution module is configured to extract the displacement and rotation of key feature points from the three-dimensional full-field deformation vector of the workpiece, generate coordinate system space transformation instructions for the CNC machine tool, and drive the machine tool to execute them.

[0008] A further improvement of the present invention is that the multimodal fusion sensing module specifically includes: Dual-spectrum acquisition unit, including an industrial camera and an infrared thermal imager calibrated with parallel or coaxial optical axes; The semantic segmentation unit is configured to run a deep learning semantic segmentation algorithm to perform pixel-level classification on the fused image tensor, dividing the pixels into the workpiece body, chip accumulation and background. The hot spot attribute determination logic is configured to determine areas with temperatures above a preset threshold and semantic labels of chip accumulation as false hot spots, and areas with temperatures above a preset threshold and semantic labels of the workpiece body as true temperature rises.

[0009] A further improvement of the present invention is that the active boundary reconstruction module specifically includes: A spatial mapping unit is configured to map the pixel coordinates of the pseudo hotspot in the image coordinate system to three-dimensional physical coordinates in the machine tool machining coordinate system; The vector chip removal execution unit includes a multi-degree-of-freedom servo nozzle and a high-pressure air valve, and is configured to adjust the nozzle angle according to the three-dimensional physical coordinates to apply pulsed high-pressure airflow to the chip accumulation area. The closed-loop verification logic is configured to trigger the multimodal fusion sensing module to collect data again and perform a judgment after the airflow application ends: calculate the number of pixels in the field of view that simultaneously meet the semantic classification of chip accumulation and temperature above a preset thermal threshold; if the number is greater than the preset convergence threshold, generate a control command to drive the vector chip removal execution unit to repeatedly perform the removal action until the number is less than or equal to the convergence threshold.

[0010] A further improvement of the present invention is that the thermal history log includes several record entries, the record entries including the center physical coordinates of the chip accumulation area, the average radiation temperature of the chips before they are removed, and the length of time the chips remain on the workpiece surface.

[0011] A further improvement of this invention is that the thermo-mechanical coupling solution module is configured to perform data assimilation processing, including reading the thermal history log, calculating the cumulative heat flux conducted from the chip to the workpiece interior based on the chip's average radiation temperature and residence time, then calculating the virtual heat source attenuation coefficient, and superimposing it onto the corresponding node of the workpiece's heat conduction equation to solve for the corrected internal core temperature field; subsequently, calling the pre-stored stiffness sensitivity field; performing a full-field weighted integral operation, and obtaining the thermo-stiffness asymmetric response index by coupling the real internal core temperature field and the stiffness sensitivity field. Finally, the thermal-stiffness asymmetric response index is used as a nonlinear gain coefficient and applied to the torsional mode vector of the reduced-order model to obtain the three-dimensional full-field deformation vector.

[0012] A further improvement of the present invention is that the thermo-mechanical coupling solution module also stores the deformation mode matrix extracted from the finite element analysis results of the engine bracket based on intrinsic orthogonal decomposition as a reduced-order model basis function, and a three-dimensional scalar field characterizing the contribution weight of each spatial point of the engine bracket to the torsional deformation of the overall structure under a unit thermal load as a stiffness sensitivity field.

[0013] A further improvement of this invention is that the calculation formula of the thermal-stiffness asymmetric response index is configured to perform a weighted integration of the real-time temperature field and the stiffness sensitivity field. Specifically, it includes multiplying the square of the ratio of the difference between the temperature at each point of the workpiece and the average temperature by the stiffness sensitivity weight of the corresponding point, then multiplying by the temperature gradient direction factor, and finally integrating over the entire field volume. The thermal-stiffness asymmetric response index is used to characterize whether heat is concentrated in the low-stiffness weak area of ​​the workpiece.

[0014] A further improvement of this invention is that the thermo-mechanical coupling solution module uses the following modified model to calculate the three-dimensional full-field deformation vector, which is obtained by summing the basic deformation prediction term based on the superposition of linear modes and the nonlinear torsional correction term; wherein, the basic deformation prediction term is the sum of the products of the modal coefficients and the corresponding order modes; and the nonlinear torsional correction term is the product of the predefined extreme torsional mode vector and the nonlinear gain coefficient based on the thermo-stiffness asymmetric response index.

[0015] A further improvement of this invention is that the adaptive compensation execution module is configured to perform six-degree-of-freedom spatial compensation, specifically including: parsing the three-dimensional full-field deformation vector, calculating the three-dimensional translation compensation amount of the workpiece coordinate system origin, and writing it into the external workpiece zero-point offset register; simultaneously calculating the spatial rotation compensation matrix of the workpiece coordinate system, extracting Euler angles, and adjusting the machining posture through the tilt surface machining command or rotary axis compensation function of the CNC system.

[0016] On the other hand, the present invention provides a method for monitoring and compensating for multi-process thermomechanical deformation of engine brackets, comprising the following steps: S1. Real-time acquisition of visible light image data and infrared thermal imaging data of the processing area, and generation of fused image tensor containing texture information and temperature information based on data fusion processing, thereby identifying the heat distribution state and chip occlusion state of the workpiece surface. S2. When the identification result shows that there is chip obstruction, drive the physical cleaning device to perform directional cleaning action on the obstructed area until the workpiece meets the hot spot attribute judgment logic and generates a thermal history log containing cleaning process data. S3. Obtain the pure surface temperature field after the boundary is restored, and calculate the three-dimensional full-field deformation vector of the workpiece at the current moment by combining the thermal history log. S4. Extract the displacement and rotation of key feature points from the three-dimensional full-field deformation vector of the workpiece, generate coordinate system space transformation instructions for the CNC machine tool, and drive the machine tool to execute them.

[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention first solves the problems of infrared thermometry distortion caused by chip accumulation in the inner cavity under micro-lubrication conditions, and thermal hysteresis prediction bias caused by the cooling of the workpiece surface after the chips are removed but the internal heat is still accumulating, by actively reconstructing the boundary and combining it with the thermal history data assimilation mechanism. It realizes the quantification of chip residence time and radiation intensity into virtual heat source compensation, ensuring that the system can accurately invert the true thermal state of the deep layer of the workpiece even when the surface is at a low temperature, thus guaranteeing the physical authenticity of the model input.

[0018] 2. By constructing a thermo-stiffness asymmetric response index and performing six-degree-of-freedom spatial compensation, the problem of torsional nonlinear warping of engine brackets under non-uniform heating due to the large difference in stiffness between flange and cantilever was solved; the coupled calculation of thermal energy distribution and structural mechanical properties was realized, and the torsional compensation amount can be adaptively adjusted according to whether the heat source acts on the low stiffness weak area, and the workpiece posture can be corrected by five-axis linkage, which significantly improves the machining accuracy of complex irregular components. Attached Figure Description

[0019] Figure 1 This is a framework diagram of a multi-process thermomechanical deformation monitoring and compensation system for engine brackets according to the present invention; Figure 2 This is a flowchart of a multi-stage thermomechanical deformation monitoring and compensation method for engine brackets according to the present invention. Detailed Implementation

[0020] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0021] The term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three cases: A exists alone, A and B exist simultaneously, and B exists alone.

[0022] Example 1 Figure 1 This embodiment illustrates a framework diagram of a multi-stage thermomechanical deformation monitoring and compensation system for engine mounts, including: The multimodal fusion sensing module is configured to acquire visible light image data of the processing area in real time. and infrared thermal imaging data Based on data fusion processing, a fused image tensor containing texture and temperature information is generated. This allows for the identification of the heat distribution and chip obstruction status on the workpiece surface; including: The system first acquires visible light images of the engine bracket machining area in real time through a dual-spectrum acquisition unit mounted on the side of the machine tool spindle. and infrared thermal imaging data And perform spatiotemporal registration.

[0023] Under MQL (Multi-Level Lubrication) conditions, aluminum alloy chips are highly reflective. Infrared thermal imagers are easily affected by changes in metal emissivity, misidentifying reflective, low-temperature chips as high-temperature areas, or failing to detect genuine high-temperature chips accumulated in deep cavities (if there is an oil film on the chip surface). A single data source cannot distinguish between the temperature rise of the workpiece itself and the chips covering it; therefore, the system inputs the registered image into a lightweight semantic segmentation network (such as DeepLabV3+) to process the tensor of the fused image ( Pixel-level classification is performed, dividing pixels into the workpiece body, chip accumulation, and background.

[0024] Semantic segmentation units are configured to segment data at temperatures above a preset threshold. Furthermore, areas with semantically labeled chip accumulation are identified as false hot spots, and temperatures exceeding a preset threshold are considered to be hot spots. Furthermore, regions whose semantic tags are the workpiece body are determined to be true temperature rises.

[0025] Although the interference source (false hot spot) is identified at this point, directly inputting the data containing the interference into the thermal model will cause errors. Therefore, a servo-guided nozzle needs to be driven to perform pulsed high-pressure purging of the area until visual confirmation that no chip residue remains. However, blind purging is an open-loop process; it ends once the purging is complete. Existing procedures discard data during the chip's presence. This means that although the chip is blown away, it has already transferred heat while remaining on the workpiece surface, and this energy data is often ignored in existing technologies. Therefore, this invention retains evidence of chip removal through an active boundary reconstruction module.

[0026] The active boundary reconstruction module is configured to drive the physical cleaning device to perform directional cleaning of the obstructed area when the identification result indicates the presence of chip obstruction, until the workpiece meets the hot spot attribute determination logic, and generates a thermal history log containing cleaning process data. ; Specifically, it includes: The spatial mapping unit is configured to map the pixel coordinates of the pseudo hot spot in the image coordinate system to three-dimensional physical coordinates in the machine tool machining coordinate system; combined with the known geometric surface data of the workpiece CAD model, the pixel coordinates are projected onto the workpiece surface through a hand-eye calibration matrix.

[0027] The vector chip removal execution unit includes a multi-degree-of-freedom servo nozzle and a high-pressure air valve, and is configured to adjust the nozzle angle according to the three-dimensional physical coordinates to apply pulsed high-pressure airflow to the chip accumulation area. The closed-loop verification logic is configured to trigger the multimodal fusion sensing module to collect data again and perform a judgment after the airflow application ends: calculate the number of pixels in the field of view that simultaneously meet the semantic classification of chip accumulation and temperature above a preset thermal threshold; if the number is greater than the preset convergence threshold, generate a control command to drive the vector chip removal execution unit to repeatedly perform the removal action until the number is less than or equal to the convergence threshold.

[0028] The heat history log It includes several record entries, each containing the center physical coordinates of the chip accumulation area. The average radiation temperature of the chips before they are removed and the length of time the chips remain on the workpiece surface. .

[0029] ,in, This represents the cumulative chip dwell time obtained by accumulating the time difference between two consecutive frames in which the chip is detected by the vision system.

[0030] However, when observing the workpiece using an infrared camera, the surface temperature is low due to the cooling effect of the air gun. If this clean but cooled surface temperature is directly input into the heat conduction model, the model will assume the workpiece's interior is also cold. But in reality, because the chips previously remained inside... Over time, heat has seeped into the interior, leading to severe thermal hysteresis errors.

[0031] Therefore, this invention constructs a thermo-mechanical coupling solution module based on data assimilation and virtual heat source injection. This thermo-mechanical coupling solution module is configured to obtain the pure surface temperature field after the boundary is recovered. ), and combined with the aforementioned heat history log Using preset workpiece mechanical property data, calculate the workpiece's three-dimensional full-field deformation vector at the current moment. ; For irregularly shaped components like engine mounts, which have a thick flange (rigid) on one side and a thin cantilever (flexible) on the other, if we only consider linear expansion... The calculations neglect torsional deformation. For example, if heat is concentrated in a thin-walled area, the support will twist like a rope, rather than simply lengthening. Therefore, the thermo-mechanical coupling solution module is also configured to use a reduced-order model combined with the workpiece's stiffness characteristics to calculate the final three-dimensional deformation vector.

[0032] Existing order reduction models (such as standard POD) typically employ the principle of linear superposition: That is, the current temperature field is projected onto several fixed deformation modes.

[0033] However, this method assumes that temperature distribution and structural stiffness are decoupled. But when machining engine brackets with MQL, an extreme bad situation often occurs: high-temperature chips accumulate precisely at the root of the thin-walled cantilever, where the bracket's stiffness is weakest. At this point, even if the total heat is not large, the resulting local torsion is enormous. Standard POD only considers the global temperature and underestimates this localized thermal shock hitting the weakest point, leading to an underestimation of the predicted deformation and inadequate compensation.

[0034] However, existing technologies that directly use the surface temperature measured by an infrared camera as the boundary condition for finite element analysis neglect the time term of heat conduction; for thick-walled supports, surface coldness does not necessarily mean internal coldness; and the deformation calculated by existing models is often less than the actual deformation (undercompensation), resulting in a smaller hole spacing after processing. Therefore, a virtual heat source model that decays over time is introduced for data assimilation.

[0035] To address the aforementioned issues, this invention does not directly use the POD coefficient, but instead constructs a correction term.

[0036] First, in the offline phase, a structural stiffness sensitivity field is calculated. This is represented as a three-dimensional matrix corresponding to the workpiece shape. The value of each point in the matrix represents the sensitivity of that point to the overall distortion caused by heating; the softer the surface, the more critical it is, and the larger the value. Based on this, a stiffness-weighted temperature field system is constructed to obtain the temperature distribution of the workpiece in real time. .

[0037] The solution module is at the corresponding coordinates The following virtual heat source attenuation coefficient is superimposed at the location. , represented as: ,in, , and The results were obtained by fitting an offline laser pulse heating experiment onto a standard test block of the same material.

[0038] This is represented as a one-dimensional unsteady-state heat conduction solution based on a semi-infinite object; the longer the time, the deeper the heat penetration. Through... The model can distinguish between fleeting chips (surface heat) and chips that have accumulated for a long time (deep heat).

[0039] This represents the physical process by which residual heat gradually diffuses and reaches equilibrium inside the metal after the chips are removed. This ensures the continuity of the temperature field and avoids abrupt changes in compensation commands.

[0040] Virtual heat source attenuation coefficient It was configured to simulate the process of heat diffusion from the surface of a workpiece to its deep core. This was achieved by superimposing this onto the heat conduction equation. This allows the solution module to correct errors based solely on the surface temperature at the current moment. The resulting deviation in the internal cooling and external heating model leads to the reconstruction of the true internal core temperature field, which includes thermal hysteresis effects, as follows: .

[0041] Secondly, the corrected actual internal core temperature field was obtained. Then, the solution module calls the pre-stored stiffness sensitivity field to obtain the sensitivity weights. The sensitivity weight is obtained in the offline stage by applying a unit thermal load (the amount of end displacement caused by heating at a certain point) to the digital twin model point by point through finite element analysis (FEA), and is expressed as the sensitivity weight at the thin wall. The value is large, and small in thick-walled areas.

[0042] Subsequently, the real-time thermal-stiffness asymmetric response index was calculated based on the above data: Where n is the normal direction of the thin-walled feature of the workpiece or the main direction of heat flow diffusion, the square term is used to amplify the influence of local high temperature, and the average temperature is ignored and only the peak temperature is considered. This item is used when high temperatures occur in areas with high stiffness. The term will be extremely small, and the thermal-stiffness asymmetric response exponent will increase relatively little if the high temperature occurs in a region of low stiffness. The term becomes extremely large, leading to a surge in the thermal-stiffness asymmetric response exponent. Through This item allows the algorithm to sense whether the heat is being applied to a weak point. The term represents the gradient direction factor, configured to determine whether the heat flow penetrates the thin wall vertically (greater harm) or flows horizontally (less harmful), further eliminating ineffective heat.

[0043] Finally, the solution module uses the thermal-stiffness asymmetric response index as a nonlinear gain coefficient and applies it to the torsional mode vector of the reduced-order model.

[0044] Final three-dimensional full-field deformation vector It consists of two parts: one part is the linear expansion component calculated based on the actual internal core temperature field, and the other part is affected by... The nonlinear torsional correction component of exponential modulation is calculated using the following formula: Among them, the linear term ensures the accuracy of the prediction of overall thermal expansion (lengthening and widening); The gain adjustment coefficient for the torsional mode typically ranges from 0.1 to 1.0 and is determined through reverse iterative optimization based on the residuals of historical processing data. Represents a predefined extreme torsional mode; This utilizes the properties of the exponential function, expressed as when... When the heat is very small (uniform heat distribution or hitting a hard surface), the correction term approaches 0, and the model degenerates into a normal POD to ensure stability; when When the exponential term increases (heat concentrates to attack the weak point), it drives the model to output a high-gain torsional compensation, thereby achieving rapid response and suppression of nonlinear deformation under extreme conditions.

[0045] Through the above-mentioned linkage, the system realizes a complete closed-loop control from chip hysteresis heating to model parameter correction, then to structural sensitivity assessment, and finally to precise coordinate compensation.

[0046] In one possible embodiment, during the finishing of the cantilever end of the engine mount, if there is a reinforcing rib with a thickness of only 3mm at the root of the cantilever, a small clump of chips accumulates at this point during MQL machining, and the temperature is 30°C higher than the surrounding area. Standard POD calculations show that the overall temperature rise is only 2°C (because the chip clump is very small and is averaged out). Therefore, POD predicts a deformation of only 0.005mm. This results in the actual machining process where the cantilever end sags by 0.05mm due to softening at the root caused by heat, leading to out-of-tolerance hole positioning.

[0047] This embodiment of the invention introduces... The system looked up the table and found that the location of the reinforcing rib was... The value is extremely high, indicating a structurally sensitive region; although the hot spot area is small, it is multiplied by a huge weight. Then, the calculated The index rose significantly. (Correction item) The model was activated, and it determined that although the overall heat was not high, the structural asymmetric response was quite strong. At this point, the predicted output value was: 0.005mm (linear) + 0.04mm (correction) = 0.045mm. The result showed that the compensation command accurately counteracted the cantilever sagging, and the hole position was acceptable.

[0048] The adaptive compensation execution module is configured to extract the displacement and rotation of key feature points from the workpiece's three-dimensional full-field deformation vector, generate coordinate system space transformation instructions for the CNC machine tool, and drive the machine tool to execute them. The execution of six-degree-of-freedom spatial compensation specifically includes: parsing the three-dimensional full-field deformation vector, calculating the three-dimensional translation compensation amount of the workpiece coordinate system origin, and writing it into the external workpiece zero-point offset register; simultaneously calculating the spatial rotation compensation matrix of the workpiece coordinate system, extracting Euler angles, and adjusting the machining posture through the CNC system's tilt surface machining instructions or rotary axis compensation function.

[0049] The system is configured to operate according to the following timing logic: before the machine tool performs the finishing process, the active boundary reconstruction module is triggered first; only when the multimodal fusion perception module confirms that there are no false hot spots in the monitoring field of view, the current moment is locked, the thermo-mechanical coupling calculation module is triggered to read the current frame temperature data and the previous thermal history log for calculation, and a compensation command is sent to the adaptive compensation execution module.

[0050] The threshold and weight settings involved in this embodiment can be set by default according to the present invention, or can be set by those skilled in the art.

[0051] Example 2 Figure 2This invention presents a flowchart of a multi-stage thermomechanical deformation monitoring and compensation method for engine mounts, based on the same inventive concept as Embodiment 1. The invention provides a multi-stage thermomechanical deformation monitoring and compensation method for engine mounts, comprising: S1. Real-time acquisition of visible light image data and infrared thermal imaging data of the processing area, and generation of fused image tensor containing texture information and temperature information based on data fusion processing, thereby identifying the heat distribution state and chip occlusion state of the workpiece surface. S2. When the identification result shows that there is chip obstruction, drive the physical cleaning device to perform directional cleaning action on the obstructed area until the workpiece meets the hot spot attribute judgment logic and generates a thermal history log containing cleaning process data. S3. Obtain the pure surface temperature field after the boundary is restored, and calculate the three-dimensional full-field deformation vector of the workpiece at the current moment by combining the thermal history log. S4. Extract the displacement and rotation of key feature points from the three-dimensional full-field deformation vector of the workpiece, generate coordinate system space transformation instructions for the CNC machine tool, and drive the machine tool to execute them.

[0052] Current CNC machining thermal compensation typically employs a three-axis linear compensation mode. This means the system calculates the workpiece's elongation in the X, Y, and Z directions, and then modifies the machine tool's external workpiece zero-point offset.

[0053] However, this method cannot handle attitude errors. Under non-uniform heating, the engine mount not only expands in volume but also undergoes spatial distortion. For example, the cantilever end of the mount warps upward by 0.05mm due to heat, accompanied by a 0.03° twist. This method can only raise the tool by 0.05mm. However, the tool axis remains perpendicular to the horizontal plane, while the workpiece surface is tilted. This results in an elliptical entry point during drilling or stepped tool marks during face milling. Secondly, it can lead to the dynamic loss of the coordinate system reference. When the workpiece undergoes thermal distortion, its positioning reference surface itself also deflects. If the machine tool coordinate system under cold conditions is continued to be used, the positional accuracy of all subsequent machining features relative to the reference surface will fail. Therefore, this embodiment introduces six-degree-of-freedom (6-DOF) rigid body transformation compensation. That is, not only is the origin of the coordinate system (X / Y / Z) moved, but the axes of the coordinate system (A / B / C) are also rotated, so that the virtual machining coordinate system of the machine tool completely coincides with the actual workpiece coordinate system after thermal deformation.

[0054] The specific steps of step S4 include: S41, Transform the three-dimensional full-field deformation vector The input consists of displacement data from tens of thousands of nodes; typically, three points are selected to define the primary datum plane (restricting three degrees of freedom), two points to define the secondary datum pin hole (restricting two degrees of freedom), and one point to define the stop surface (restricting one degree of freedom). Then, the data is extracted... in, These are theoretical coordinates under cold conditions. It is the thermal deformation vector corresponding to that point. These are the actual coordinates under thermal conditions.

[0055] S42. Solve the least squares problem using the Singular Value Decomposition (SVD) algorithm, expressed as: Because different reference points on the engine mount have different levels of importance, for example, the main locating hole (pin hole) has a different weight. Set to 1.0, and the weight of the auxiliary support surface. Set to 0.5. (By...) The compensation strategy prioritizes the accuracy of core positioning features (such as hole positions) and squeezes residual thermal deformation errors that cannot be completely eliminated onto secondary non-mating surfaces, thereby maximizing the yield of finished products under the premise that physical deformation is irreversible.

[0056] This leads to a rigid body transformation matrix. This makes the cold reference point set P, after transformation, similar to the hot reference point set. The sum of squared errors between them is minimized. The thermal deformation of the engine mount is non-rigid (with local warping), so it is impossible to find a matrix that makes all points perfectly coincide. SVD can find a globally optimal solution, distributing the residual nonlinear deformation error evenly across all locations, avoiding large deviations at any critical hole position.

[0057] S43, calculate The homogeneous transformation matrix is ​​decomposed into parameters executable by the machine tool. The translation compensation is then calculated. And the rotation compensation amount, expressed as Euler angles extracted from the rotation matrix R (assuming the machine tool rotation sequence is ZYX): , , And set safety limiting logic: set rotation threshold (optional in this embodiment) If the calculated angle exceeds this value, the system determines that the workpiece deformation is too large or that the clamping has become loose, triggering an emergency stop alarm to prevent a collision.

[0058] S44, CNC instruction generation and driving In one possible embodiment, for writing translation errors (3-axis machine tool or 5-axis universal): the system uses the FOCAS / OPC-UA protocol to... Write directly to the workpiece zero-point offset register of the CNC machine's G54-G59 section.

[0059] In another possible embodiment, for writing rotational errors (for 5-axis machine tools): utilize the tilting surface machining function of a modern CNC system. The system will calculate the Euler angles. This is injected into the G68.2 instruction parameters. G68.2 activates the machine tool's RTCP (Tool Tip Follow) logic. When the machine tool rotates the A / C axes to compensate for workpiece torsion, the controller automatically calculates the compensation motion of the X / Y / Z axes to ensure that the physical position of the tool tip relative to the workpiece remains unchanged. Without this instruction, simply rotating the worktable would cause the workpiece to be thrown off the tool, resulting in serious machine collisions or machining position errors.

[0060] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied 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.

[0061] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0062] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0063] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0064] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A multi-process thermomechanical deformation monitoring and compensation system for engine brackets, characterized in that: The system includes: The multimodal fusion sensing module is configured to acquire visible light image data and infrared thermal imaging data of the processing area in real time, and generate a fused image tensor containing texture information and temperature information based on data fusion processing, thereby identifying the heat distribution state and chip occlusion state of the workpiece surface. The active boundary reconstruction module is configured to drive the physical cleaning device to perform directional cleaning action on the obstructed area when the identification result shows that there is chip obstruction, until the workpiece meets the hot spot attribute determination logic and generates a thermal history log containing cleaning process data. The thermo-mechanical coupling solution module is configured to obtain the pure surface temperature field after the boundary is restored, and calculate the three-dimensional full-field deformation vector of the workpiece at the current moment by combining the thermal history log. The adaptive compensation execution module is configured to extract the displacement and rotation of key feature points from the three-dimensional full-field deformation vector of the workpiece, generate coordinate system space transformation instructions for the CNC machine tool, and drive the machine tool to execute them.

2. The multi-process thermomechanical deformation monitoring and compensation system for engine brackets according to claim 1, characterized in that: The multimodal fusion sensing module specifically includes: Dual-spectrum acquisition unit, including an industrial camera and an infrared thermal imager calibrated with parallel or coaxial optical axes; The semantic segmentation unit is configured to run a deep learning semantic segmentation algorithm to perform pixel-level classification on the fused image tensor, dividing the pixels into the workpiece body, chip accumulation and background. The hot spot attribute determination logic is configured to determine areas with temperatures above a preset threshold and semantic labels of chip accumulation as false hot spots, and areas with temperatures above a preset threshold and semantic labels of the workpiece body as true temperature rises.

3. The multi-process thermomechanical deformation monitoring and compensation system for engine brackets according to claim 2, characterized in that: The active boundary reconstruction module specifically includes: A spatial mapping unit is configured to map the pixel coordinates of the pseudo hotspot in the image coordinate system to three-dimensional physical coordinates in the machine tool machining coordinate system; The vector chip removal execution unit includes a multi-degree-of-freedom servo nozzle and a high-pressure air valve, and is configured to adjust the nozzle angle according to the three-dimensional physical coordinates to apply pulsed high-pressure airflow to the chip accumulation area. The closed-loop verification logic is configured to trigger the multimodal fusion sensing module to collect data again and perform a judgment after the airflow application ends: calculate the number of pixels in the field of view that simultaneously meet the semantic classification of chip accumulation and temperature above a preset thermal threshold; if the number is greater than the preset convergence threshold, generate a control command to drive the vector chip removal execution unit to repeatedly perform the removal action until the number is less than or equal to the convergence threshold.

4. The multi-process thermomechanical deformation monitoring and compensation system for engine brackets according to claim 1, characterized in that: The thermal history log includes several record entries, which include the central physical coordinates of the chip accumulation area, the average radiation temperature of the chips before they are removed, and the duration of the chips' residence on the workpiece surface.

5. A multi-process thermomechanical deformation monitoring and compensation system for engine brackets according to claim 4, characterized in that: The thermo-mechanical coupling solution module is configured to perform data assimilation processing, including reading the thermal history log, calculating the cumulative heat flux conducted from the chip to the workpiece interior based on the chip's average radiation temperature and residence time, and then calculating the virtual heat source attenuation coefficient, which is superimposed on the corresponding node of the workpiece's heat conduction equation to solve for the corrected internal core temperature field; subsequently, it calls the pre-stored stiffness sensitivity field; performs a full-field weighted integral operation, and obtains the thermo-stiffness asymmetric response index by coupling the real internal core temperature field and the stiffness sensitivity field. Finally, the thermal-stiffness asymmetric response index is used as a nonlinear gain coefficient and applied to the torsional mode vector of the reduced-order model to obtain the three-dimensional full-field deformation vector.

6. The multi-process thermomechanical deformation monitoring and compensation system for engine brackets according to claim 5, characterized in that: The thermo-mechanical coupling solution module also stores the deformation mode matrix extracted from the finite element analysis results of the engine bracket based on intrinsic orthogonal decomposition as a reduced-order model basis function, and a three-dimensional scalar field characterizing the contribution weight of each spatial point of the engine bracket to the torsional deformation of the overall structure under unit thermal load as a stiffness sensitivity field.

7. A multi-process thermomechanical deformation monitoring and compensation system for engine brackets according to claim 5, characterized in that: The calculation formula for the thermal-stiffness asymmetric response index is configured to perform a weighted integration of the real-time temperature field and the stiffness sensitivity field. Specifically, it includes multiplying the square of the ratio of the difference between the temperature at each point of the workpiece and the average temperature by the stiffness sensitivity weight of the corresponding point, then multiplying by the temperature gradient direction factor, and finally integrating over the entire field volume. The thermal-stiffness asymmetric response index is used to characterize whether heat is concentrated in the low-stiffness weak area of ​​the workpiece.

8. A multi-process thermomechanical deformation monitoring and compensation system for engine brackets according to claim 7, characterized in that: The thermo-mechanical coupling solution module calculates the three-dimensional full-field deformation vector using the following modified model, which is the sum of the basic deformation prediction term based on the superposition of linear modes and the nonlinear torsional correction term; wherein, the basic deformation prediction term is the sum of the products of the modal coefficients and the corresponding order modes; the nonlinear torsional correction term is the product of the predefined extreme torsional mode vector and the nonlinear gain coefficient based on the thermo-stiffness asymmetric response index.

9. A multi-process thermomechanical deformation monitoring and compensation system for engine brackets according to claim 1, characterized in that: The adaptive compensation execution module is configured to perform six-degree-of-freedom spatial compensation, specifically including: parsing the three-dimensional full-field deformation vector, calculating the three-dimensional translation compensation amount of the workpiece coordinate system origin, and writing it into the external workpiece zero-point offset register; at the same time, calculating the spatial rotation compensation matrix of the workpiece coordinate system, extracting Euler angles, and adjusting the machining posture through the tilt surface machining command or rotary axis compensation function of the CNC system.

10. A multi-process thermomechanical deformation monitoring and compensation system for engine brackets, characterized in that: The system execution method includes the following steps: S1. Real-time acquisition of visible light image data and infrared thermal imaging data of the processing area, and generation of fused image tensor containing texture information and temperature information based on data fusion processing, thereby identifying the heat distribution state and chip occlusion state of the workpiece surface. S2. When the identification result shows that there is chip obstruction, drive the physical cleaning device to perform directional cleaning action on the obstructed area until the workpiece meets the hot spot attribute judgment logic and generates a thermal history log containing cleaning process data. S3. Obtain the pure surface temperature field after the boundary is restored, and calculate the three-dimensional full-field deformation vector of the workpiece at the current moment by combining the thermal history log. S4. Extract the displacement and rotation of key feature points from the three-dimensional full-field deformation vector of the workpiece, generate coordinate system space transformation instructions for the CNC machine tool, and drive the machine tool to execute them.

Images