Casting machining and polishing treatment device and method for precision casting

Abstract

The invention discloses a casting machining and polishing treatment device and method for precision casting, and belongs to the technical field of automatic manufacturing and industrial robots. The multi-axis mechanical arm is mounted on the machining base; the combined treatment head is mounted at the tail end of the multi-axis mechanical arm; the composite processing head comprises an electric spindle grinding machine, a three-dimensional scanner, a structured light projector and a macro camera; and the three-dimensional scanner, the structured light projector and the macro camera are all positioned and mounted by taking the spindle axis of the motorized spindle grinding machine as a reference. According to the motorized spindle grinding machine, a plurality of procedures such as macroscopic scanning, microscopic detection, grinding execution and quality evaluation are physically integrated.

Description

Technical Field

[0001] This invention relates to the field of automated manufacturing and industrial robots, specifically to a device and method for machining and grinding castings for precision casting. Background Technology

[0002] In the subsequent processing of precision castings, traditional finishing methods mainly rely on manual grinding or use equipment with low automation and segmented processes. These methods have a series of technical bottlenecks when dealing with flash, gating and risers, and micro-defects on the surface of castings.

[0003] The current situation and its shortcomings are specifically reflected in: Separating macro and micro processing steps leads to inefficiency: In traditional processes, the removal of macro excesses such as flash and risers and the fine grinding of micro defects such as pits and protrusions are usually separate steps. This separate operation mode not only requires repeated loading, unloading and positioning of workpieces, which significantly prolongs the processing cycle, but also increases the cumulative error introduced by repeated positioning, affecting the final processing accuracy.

[0004] Optical inspection has poor adaptability to complex surfaces: casting surfaces, especially metal surfaces, often have high reflectivity or specular reflection characteristics, which can cause local overexposure when optical inspection methods such as structured light scanning acquire three-dimensional data, resulting in data loss or serious distortion. In addition, surfaces with complex curvature are also prone to generating false defect signals due to optical effects. If these are not identified, automated equipment will perform unnecessary grinding actions, which wastes time and also poses a risk of damaging the workpiece.

[0005] The lack of adaptive force control in the grinding process leads to poor quality consistency: Traditional automated grinding equipment mostly uses trajectory control based on preset paths, which cannot adjust the grinding contact force in real time according to the specific defect shape and size of the grinding area. This rigid execution method is prone to over-cutting of small defects or insufficient removal of larger allowances when facing defects of different sizes, resulting in uneven surface quality and difficulty in ensuring product consistency.

[0006] Quality inspection is disconnected from the processing and lacks closed-loop feedback: In traditional processes, quality inspection after polishing is usually carried out offline as an independent process. This lagging quality feedback mechanism makes it impossible to detect and correct quality problems that occur during processing in a timely manner. Once defective products are found, rework is often required, which not only affects production efficiency but also increases the scrap rate, making it difficult to achieve efficient online quality control.

[0007] The information disclosed in the background section above is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0008] The purpose of this invention is to provide a precision casting casting processing and grinding apparatus and method to solve the problems mentioned in the background art.

[0009] The technical solution of the present invention is as follows: a processing base; a multi-axis robotic arm mounted on the processing base; and a composite processing head mounted at the end of the multi-axis robotic arm. The composite processing head includes an electric spindle grinder, a 3D scanner, a structured light projector, and a macro camera. The 3D scanner, the structured light projector, and the macro camera are all positioned and installed with the spindle axis of the electric spindle grinder as a reference.

[0010] Preferably, a rotary worktable is mounted on the top surface of the processing base, and the rotary worktable is used to support the casting to be processed.

[0011] Preferably, a three-jaw self-centering chuck is fixed on the upper surface of the rotary table.

[0012] Preferably, the composite processing head further includes a torque sensor, which is installed in series between the end of the multi-axis robotic arm and the composite processing head.

[0013] A method for machining and grinding castings for precision casting includes: acquiring global three-dimensional data of the casting to be processed using a three-dimensional scanner, comparing the global three-dimensional data with a preset standard CAD model to identify macroscopic allowance areas, thereby generating a reference grinding path; controlling a multi-axis robotic arm to drive the composite processing head to move along the reference grinding path; before the electric spindle grinder contacts the casting to be processed, acquiring a reference brightness image of the pre-processed area using a macro camera to set an exposure suppression coefficient, and adjusting the projection brightness of the structured light projector and the exposure time of the macro camera based on the exposure suppression coefficient, and then acquiring a deformed grating image of the area to calculate microscopic defect features; generating a desired contact force feedback model based on the microscopic defect features and real-time contact force data fed back by a torque sensor, and using it to dynamically adjust the grinding posture and contact force of the electric spindle grinder; after grinding is completed, scanning the processed surface again using the structured light projector and the macro camera to obtain surface flatness data, and evaluating the grinding quality based on the surface flatness data.

[0014] Preferably, after acquiring the deformed grating image to solve the micro-defect features, the method further includes: driving the multi-axis robotic arm to generate a slight attitude deflection, performing a second structured light scan and three-dimensional reconstruction of the pre-processed area from a new angle; calculating the positional deviation of the two three-dimensional reconstruction contours in space, and comparing the positional deviation with a preset geometric tolerance threshold to eliminate false defects.

[0015] Preferably, the step of generating a desired contact force feedback model based on the micro-defect features and the real-time contact force data fed back by the torque sensor includes: calculating the material surplus or deficit volume between the micro-defect features and the standard surface based on the geometric dimensions of the micro-defect features; and converting the volume difference into a desired contact force change curve along the grinding path according to the material removal rate characteristic parameters of the currently used grinding wheel, so as to establish the desired contact force feedback model.

[0016] Preferably, after evaluating the polishing quality based on the surface flatness data, the method further includes: when the evaluated polishing quality does not meet the preset standard, generating a compensation polishing path for a specific substandard area; controlling the multi-axis robotic arm to execute the compensation polishing path for secondary fine polishing.

[0017] This invention provides an improved apparatus and method for machining and grinding castings for precision casting, which has the following improvements and advantages compared with the prior art: 1. This solution integrates an electric spindle grinder, a 3D scanner, a structured light projector, and a macro camera into a single composite processing head, and calibrates and installs all optical inspection units with the grinder spindle as the reference. This design physically integrates multiple processes such as macroscopic scanning, microscopic inspection, grinding execution, and quality assessment into one. 2. This solution eliminates the need for workpieces to be transferred between different devices, fundamentally eliminating positioning errors caused by repeated clamping. Through the coordinated movement of the rotary table and the multi-axis robotic arm, scanning and processing of all surfaces of the casting can be completed in a single clamping, greatly shortening auxiliary time and improving overall processing efficiency and automation level. 3. The torque sensor in this solution can monitor the contact force and torque generated during the grinding process in real time, providing key real-time data input for subsequent adaptive force control, enabling the grinding process to change from rigid execution to flexible interaction, and adapting to the subtle differences between different castings; 4. By driving the robotic arm to perform minute posture deflections, performing two structured light scans and 3D reconstructions, and comparing the positional deviations of the two reconstructed contours, false defects caused by optical effects can be effectively eliminated. This combined method ensures that subsequent polishing is performed on real physical defects, significantly improving the accuracy and reliability of defect identification. Attached Figure Description

[0018] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a schematic diagram of the overall structure of the device; Figure 2 This is a schematic diagram of a multi-axis robotic arm and its connection structure; Figure 3This is a schematic diagram of the composite processing head; Figure 4 This is a schematic diagram of the method flow of the present invention; In the diagram: 100, machining base; 110, rotary table; 111, three-jaw self-centering chuck; 200, multi-axis robotic arm; 300, composite processing head; 310, 3D scanner; 320, structured light projector; 330, macro camera; 340, electric spindle grinder; 350, torque sensor. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0020] Example 1 Please see Figure 1-3 This invention provides a precision casting casting processing and grinding device, comprising: a processing base 100; a multi-axis robotic arm 200 mounted on the processing base 100; and a composite processing head 300 mounted at the end of the multi-axis robotic arm 200. The composite processing head 300 includes an electric spindle grinder 340, a 3D scanner 310, a structured light projector 320, and a macro camera 330. The 3D scanner 310, the structured light projector 320, and the macro camera 330 are all positioned and installed with the spindle axis of the electric spindle grinder 340 as a reference. The precision casting grinding and polishing apparatus of this embodiment integrates an electric spindle grinder 340, a 3D scanner 310, a structured light projector 320, and a macro camera 330 in its composite processing head 300. The 3D scanner 310, structured light projector 320, and macro camera 330 are positioned and installed with the spindle axis of the electric spindle grinder 340 as a reference. This reference-based positioning structure allows the 3D scanner 310 to acquire complete 3D data of the casting during the global scanning phase. By comparing this data with a standard CAD model, it identifies macroscopic excess areas such as flash and risers, thereby generating a reference grinding path. During the grinding process, the structured light projector 320 and macro camera 330 continuously perform microscopic inspection of the area to be processed in front of the grinder, enabling the identification and location of minute defects. This ensures that the identification and processing of microscopic defects are carried out simultaneously while the macroscopic grinding path is being executed, avoiding the separation of macroscopic and microscopic processing in traditional manual grinding, and helping to improve the overall efficiency and accuracy of the process. More importantly, since all detection units are uniformly calibrated with the axis of the electric spindle grinder 340 as the reference, the entire composite processing head 300 operates in the same high-precision coordinate system. This enables seamless conversion between the macroscopic allowance coordinates found by the 3D scanner 310, the microscopic defect coordinates located by the macro camera 330, and the grinding path coordinates executed by the electric spindle grinder 340. This design fundamentally avoids the complex data conversion process between different devices and different coordinate systems, which is prone to introducing cumulative errors. It is the key to achieving high-precision, integrated processing.

[0021] A rotary worktable 110 is mounted on the top surface of the machining base 100. The rotary worktable 110 is used to support the casting to be processed. In this embodiment, the precision casting machining and grinding apparatus for castings includes a rotary worktable 110 mounted on the top surface of the machining base 100 to support the casting to be processed. A servo motor is mounted on the bottom of the rotary worktable 110, which is connected to the rotating shaft of the turntable via a precision worm gear reduction mechanism. This structure provides stable rotational torque and positioning accuracy. During the global three-dimensional data acquisition phase, the controller can drive the multi-axis robotic arm 200 to move the composite processing head 300 to perform multi-angle scanning around the casting fixed on the rotary worktable 110. The addition of the rotary worktable 110 eliminates the need for the multi-axis robotic arm 200 to perform large-scale... The spatial movement of the rotary table 110, the three-jaw chuck, and the multi-axis robotic arm 200, working in coordination, is particularly important during the global 3D data acquisition phase. The controller drives the rotary table 110 to rotate at a preset angular velocity while simultaneously controlling the multi-axis robotic arm 200 to maintain a relatively fixed posture, thereby enabling the scanning of all surfaces of the casting. This coordinated control avoids large spatial movements of the robotic arm, reducing its travel distance. During the grinding process, the three-jaw self-centering chuck 111 ensures the concentricity of the casting, allowing the controller to accurately map the grinding path onto the rotating casting surface through simple coordinate transformations. Even if there is slight clamping eccentricity in the casting, the controller can detect changes in contact force in real time through the torque sensor 350 and fine-tune the posture of the robotic arm to compensate for this.

[0022] All surfaces to be processed can be scanned by rotating the casting itself, which helps to shorten scanning time and reduce the movement of the robotic arm, thereby reducing energy consumption and mechanical wear while ensuring data integrity.

[0023] A three-jaw self-centering chuck 111 is fixed on the upper surface of the rotary table 110; In this embodiment, a three-jaw self-centering chuck 111 is fixed on the upper surface of the rotary table 110. This three-jaw self-centering chuck 111 can clamp castings of different sizes and shapes, providing stable clamping and positioning for the castings. During the processing, the three-jaw self-centering chuck 111 can ensure that the castings maintain accurate concentricity on the rotary table 110, which is crucial for the multi-axis robotic arm 200 and the composite processing head 300 to perform precision grinding. Through the clamping of the chuck, the castings will not loosen or shift during high-speed grinding, ensuring the stability of the processing and the final accuracy, and avoiding grinding deviations or rework caused by inaccurate positioning.

[0024] The composite processing head 300 also includes a torque sensor 350, which is connected in series between the end of the multi-axis robotic arm 200 and the composite processing head 300. In this embodiment, a torque sensor 350 is connected in series between the composite processing head 300 and the multi-axis robotic arm 200. This torque sensor 350 can measure the contact force and torque generated in various directions during the grinding process in real time. During the grinding stage, the controller continuously receives data from the torque sensor 350 and compares it with a pre-generated expected contact force feedback model. For example, when the grinding wheel contacts a harder macroscopic allowance or a small protrusion, the reading of the torque sensor 350 will increase accordingly; when the grinding wheel passes through a pit, the reading will decrease instantly. Based on these real-time feedbacks, the controller can dynamically adjust the grinding posture and feed speed of the electric spindle grinder 340 to maintain a constant contact force. This helps to achieve uniform grinding of different areas, prevent over-cutting or under-cutting, and thus improve the surface quality of the grinding.

[0025] Example 2 Please see Figure 4A method for machining and grinding castings for precision casting includes: acquiring global three-dimensional data of the casting to be processed using a 3D scanner 310, comparing the global three-dimensional data with a preset standard CAD model to identify macroscopic allowance areas, thereby generating a reference grinding path; controlling a multi-axis robotic arm 200 to drive a composite processing head 300 to move along the reference grinding path; before the electric spindle grinder 340 contacts the casting to be processed, acquiring a reference brightness image of the pre-processed area using a macro camera 330 to set an exposure suppression coefficient, and based on the exposure suppression coefficient... After adjusting the projection brightness of the structured light projector 320 and the exposure time of the macro camera 330, a deformable grating image of the area is acquired to calculate the micro-defect features. Based on the micro-defect features and the real-time contact force data fed back by the torque sensor 350, a desired contact force feedback model is generated and used to dynamically adjust the grinding posture and contact force of the electric spindle grinder 340. After grinding is completed, the processed surface is scanned again using the structured light projector 320 and the macro camera 330 to obtain surface flatness data, and the grinding quality is evaluated based on the surface flatness data. This method uses a Creaform HandySCAN700 3D scanner to perform a global 3D scan of the casting, acquiring point cloud data. A Siemens 840DSL controller reconstructs the point cloud data into a complete 3D digital model of the casting and compares it with a preset standard CAD model. This comparison automatically identifies macroscopic excess areas such as flash and risers, and generates a baseline grinding path covering the area to be processed. This process solves the problems of time-consuming and manual identification of macroscopic excess areas in traditional methods.

[0026] A multi-axis robotic arm 200 (model: ABBIRB6700) moves a composite processing head 300 along the reference grinding path. Before the electric spindle grinder 340 contacts a specific processing point, a macro camera 330 (model: Baslerac A2040-180kc) acquires a reference brightness image of that area. The controller analyzes the brightness distribution of this image to determine if high reflectivity exists and calculates an exposure suppression coefficient. The calculation of the exposure suppression coefficient is not based on a single statistical feature, but rather on analyzing the brightness histogram distribution of the reference brightness image. When the proportion of high-brightness pixels in the histogram is high, it indicates the presence of specular reflection areas. The controller then calculates a coefficient positively correlated with the proportion of high-brightness pixels. This coefficient... The relationship between the projection brightness of the structured light projector 320 and the exposure time of the macro camera 330, used for quantitative adjustment, can be expressed as follows: Adjusted projection brightness

[0027] Adjusted exposure time

[0028] in, The maximum brightness of the projector. This is the camera's maximum exposure time. This is the exposure suppression factor, and its value ranges from... Between; this histogram-based analysis method can more accurately reflect the degree of highly reflective areas, thereby achieving more precise exposure control.

[0029] This coefficient is used to quantitatively instruct the structured light projector 320 (model: FisheyeLensT1.435mm) to reduce the projection brightness and simultaneously adjust the exposure time of the macro camera 330, thereby effectively avoiding data acquisition failure caused by specular reflection.

[0030] The controller then activates the structured light projector 320 to project the grating pattern, and the macro camera 330 simultaneously captures the deformed grating image; the controller calculates the geometric features of the suspected micro defects based on the deformed image; this series of steps ensures that micro defects can be accurately identified and located before polishing, rather than blindly processing them; Compared with traditional fixed-parameter scanning or time-consuming HDR (High Dynamic Range) imaging technology, the method proposed in this invention, which sets the exposure suppression coefficient in real time based on a reference brightness image, is an instantaneous and local adaptive optimization for the processing area. It can quickly generate the best optical acquisition parameters for each area to be processed within the grinding cycle, thereby greatly improving the detection success rate and data quality of the microscopic features of complex high-reflectivity castings without sacrificing processing efficiency.

[0031] During the grinding process, this method generates a desired contact force feedback model based on the solved microscopic defect characteristics and real-time contact force data fed back by a torque sensor 350 (model: ATIGamma). The model is established as follows: the controller calculates the material surplus or deficit volume between the defect and the standard surface based on the defect's geometric dimensions, and transforms this volume difference into a curve representing the expected contact force change along the grinding path according to the material removal rate characteristic parameters of the currently used grinding wheel. The material removal rate characteristic parameters can be obtained through prior experimental calibration. These parameters are related to factors such as the grinding wheel material, rotational speed, and the hardness of the casting to be processed. For example, a function model reflecting the material removal rate can be established by conducting grinding experiments on a standard sample and recording the material removal volume under specific contact forces and grinding times. ,in For material removal rate, For contact force, For polishing speed.

[0032] The controller calculates the material surplus / deficit volume based on the characteristics of micro-defects. And combined with the material removal rate characteristic parameters of the grinding wheel The process of converting volume differences into a curve representing the expected change in contact force along the grinding path can be simplified as follows: Expected polishing time

[0033] According to the preset polishing speed Based on the characteristics of the grinding wheel, the desired contact force required at different positions is calculated. .

[0034] The controller dynamically adjusts the grinding posture and contact force of the electric spindle grinder 340 by comparing the data from the torque sensor 350 with the curve in real time. This dynamic adjustment mechanism enables the grinding process to automatically adapt to different defect areas, ensuring the uniformity and accuracy of the grinding.

[0035] After grinding, the composite processing head 300 moves slightly and uses the structured light projector 320 and macro camera 330 to quickly scan the processed surface again to obtain surface flatness data. The controller evaluates the grinding quality based on this data. This online evaluation step provides instant quality feedback, prevents substandard castings from flowing into the next stage, and helps improve the overall quality control level of the production line.

[0036] After acquiring deformable grating images to solve micro-defect features, the process also includes: driving the multi-axis robotic arm 200 to generate a small attitude deflection, performing a second structured light scan and three-dimensional reconstruction of the pre-processed area from a new angle; calculating the positional deviation of the two three-dimensional reconstruction contours in space, and comparing the positional deviation with a preset geometric tolerance threshold to eliminate false defects. In this embodiment, after acquiring the deformable grating image and initially calculating the microscopic defect features, the multi-axis robotic arm 200 immediately generates a slight attitude deflection to perform a second structured light scan and 3D reconstruction of the pre-processed area from a new angle. The controller calculates the positional deviation of the two 3D reconstructed contours in space. When the deviation value is compared with a preset geometric tolerance threshold, such as ±0.05 mm, the geometric tolerance threshold is not a simple empirical value, but is determined by statistical methods based on the equipment calibration stage. The repeatability accuracy of the multi-axis robotic arm 200 is obtained through repeated positioning experiments. By performing multiple 3D reconstructions on a standard geometric object, the systematic error of the 3D reconstruction algorithm is obtained. The determination of the geometric tolerance threshold is a comprehensive reflection of these two types of errors, and can be expressed as:

[0037] Among them is The confidence coefficient is typically between 2 and 3 to ensure that, in most cases, such as at a confidence level of 95% or 99%, it can correctly distinguish between real defects and spurious defects.

[0038] If the deviation is extremely small, it can be identified as a genuine physical defect; if the deviation increases significantly, it can be identified as a pseudo-defect caused by specular reflection or complex curvature, and thus eliminated. This geometric consistency verification method can effectively improve the accuracy of defect identification, avoid ineffective polishing of optical pseudo-defects, and thereby improve processing efficiency. The advantage of this method is that it does not rely on software filtering based on experience or algorithmic assumptions, but utilizes the basic principles of the physical world: the position of real geometric features in three-dimensional space is fixed, while the position of optical effects, such as the bright spot produced by specular reflection, will change significantly or even disappear with the change of observation angle; therefore, by comparing physical secondary observations, it can identify genuine and fake defects more reliably and robustly than pure software algorithms, and significantly reduce the false judgment rate.

[0039] Based on the micro-defect characteristics and real-time contact force data fed back by torque sensor 350, a desired contact force feedback model is generated, including: calculating the material surplus or deficit volume between the micro-defect characteristics and the standard surface based on the geometric dimensions of the micro-defect characteristics; and converting the volume difference into a desired contact force change curve along the grinding path according to the material removal rate characteristic parameters of the currently used grinding wheel, so as to establish the desired contact force feedback model. In this embodiment, the process of generating the desired contact force feedback model is as follows: The controller first calculates the material surplus or deficit volume between the defect and the standard surface based on the microscopic defect characteristics, such as the geometric dimensions of depth and width. For example, for a protrusion with a depth of 0.2 mm and a diameter of 1.0 mm, the controller calculates its approximate volume. The controller then converts this volume difference into a desired contact force variation curve along the grinding path based on the material removal rate characteristic parameters of the grinding wheel used on the current electric spindle grinder 340. This curve provides the controller with a reference for the contact force during the grinding process. For example, the contact force needs to be increased when passing through a protrusion, while the contact force needs to be decreased when passing through a pit. The establishment of this model transforms the grinding process from simple path execution to a dynamic control based on defect characteristics, which helps to achieve more refined grinding. Compared to traditional constant force control strategies, the expected contact force feedback model of this invention achieves a higher level of adaptive control. Constant force control cannot distinguish the size of defects, and applying the same force to defects of different sizes can easily lead to over-grinding of small defects or low removal efficiency for larger defects. This invention, by tailoring a dynamically changing expected force curve for each defect volume, ensures that the contact force of the grinding tool at each point on the path is precisely matched with the amount of material to be removed at that point, thereby maximizing processing efficiency while ensuring uniform grinding quality.

[0040] After evaluating the polishing quality based on surface flatness data, the process also includes: when the evaluated polishing quality does not meet the preset standard, generating a compensation polishing path for specific substandard areas; controlling the multi-axis robotic arm 200 to execute the compensation polishing path for secondary fine polishing; In this embodiment, after grinding is completed, the controller evaluates the grinding quality based on surface smoothness data. If the evaluation result shows that the smoothness of a certain local area does not meet the preset standard, for example, the surface roughness Ra value is greater than 3.2 micrometers, the controller immediately generates a compensatory grinding path for that specific substandard area. The generation of the compensatory grinding path is a closed-loop control process based on the online evaluation result. When the evaluation result shows that the surface smoothness of a certain local area does not meet the standard, the controller uses surface data obtained from structured light scanning and 3D reconstruction to compare the substandard area with a standard CAD model. Through geometric difference analysis, a more targeted fine grinding path covering the substandard area is generated. This new path typically has a smaller step size and a slower feed rate to achieve precise correction of the area. This process is repeated until the surface quality of the area meets the standard.

[0041] The multi-axis robotic arm 200 then executes the compensation path to perform a secondary fine grinding of the area; this compensation grinding process can be repeated until the online evaluation result of the area meets the standard. This cyclical grinding and evaluation mechanism helps ensure that the surface quality of each processed area meets the requirements, reduces the need for manual rework, and improves the product qualification rate.

[0042] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A precision casting casting processing and grinding device, characterized in that, include: A processing base (100); a multi-axis robotic arm (200) mounted on the processing base (100); a composite processing head (300) mounted at the end of the multi-axis robotic arm (200); the composite processing head (300) includes an electric spindle grinder (340), a 3D scanner (310), a structured light projector (320), and a macro camera (330); the 3D scanner (310), the structured light projector (320), and the macro camera (330) are all positioned and installed with the spindle axis of the electric spindle grinder (340) as a reference.

2. The precision casting casting processing and grinding device according to claim 1, characterized in that, A rotary table (110) is mounted on the top surface of the processing base (100), and the rotary table (110) is used to support the casting to be processed.

3. The precision casting casting processing and grinding device according to claim 2, characterized in that, A three-jaw self-centering chuck (111) is fixed on the upper surface of the rotary table (110).

4. The precision casting casting processing and grinding device according to claim 1, characterized in that, The composite processing head (300) also includes a torque sensor (350), which is connected in series between the end of the multi-axis robotic arm (200) and the composite processing head (300).

5. A method for machining and grinding castings for precision casting, applied to the precision casting machining and grinding apparatus according to any one of claims 1 to 4, characterized in that, include: The three-dimensional scanner (310) acquires the global three-dimensional data of the casting to be processed, and compares the global three-dimensional data with the preset standard CAD model to identify the macro-scale allowance area, thereby generating a reference grinding path; the multi-axis robotic arm (200) is controlled to drive the composite processing head (300) to move along the reference grinding path. Before the electric spindle grinder (340) contacts the casting to be processed, the macro camera (330) acquires the reference brightness image of the pre-processed area to set the exposure suppression coefficient. Based on the exposure suppression coefficient, the projection brightness of the structured light projector (320) and the exposure time of the macro camera (330) are adjusted, and the deformed grating image of the area is acquired to solve the micro-defect features. Based on the micro-defect characteristics and the real-time contact force data fed back by the torque sensor (350), a desired contact force feedback model is generated and used to dynamically adjust the grinding posture and contact force of the electric spindle grinder (340). After grinding is completed, the processed surface is scanned again using the structured light projector (320) and the macro camera (330) to obtain surface flatness data, and the grinding quality is evaluated based on the surface flatness data.

6. The method for machining and grinding castings for precision casting according to claim 5, characterized in that, After acquiring the deformed grating image to solve the micro-defect features, the method further includes: driving the multi-axis robotic arm (200) to generate a small attitude deflection, performing a second structured light scan and three-dimensional reconstruction of the pre-processed area from a new angle; calculating the positional deviation of the two three-dimensional reconstruction contours in space, and comparing the positional deviation with a preset geometric tolerance threshold to eliminate false defects.

7. The method for machining and grinding castings for precision casting according to claim 5, characterized in that, The process of generating a desired contact force feedback model based on the micro-defect features and the real-time contact force data fed back by the torque sensor (350) includes: calculating the material surplus or deficit volume between the micro-defect features and the standard surface based on the geometric dimensions of the micro-defect features; and converting the volume difference into a desired contact force change curve along the grinding path according to the material removal rate characteristic parameters of the currently used grinding wheel, so as to establish the desired contact force feedback model.

8. The method for machining and grinding castings for precision casting according to claim 5, characterized in that, After evaluating the polishing quality based on the surface flatness data, the process further includes: when the evaluated polishing quality does not meet the preset standard, generating a compensation polishing path for a specific substandard area; controlling the multi-axis robotic arm (200) to execute the compensation polishing path for secondary fine polishing.

Images