A milling control system for the rear end cover of a new energy motor

Abstract

This invention relates to the field of new energy vehicle motor manufacturing, and discloses a milling control system for the rear end cover of a new energy motor, including a central control module, a positioning and detection module, an environmental preprocessing module, a vision and laser detection module, a milling execution module, and a data storage module. The positioning and detection module determines the clamping state through both air static pressure values ​​and fluctuation rates; the environmental preprocessing module cleans the workpiece surface before detection using a directional air cutter and a protective shutter; the central control module generates tool compensation parameters based on collected geometric features, combined with a thermal error model and historical machining trends; the milling execution module triggers a spindle speed fluctuation strategy based on the load current variance and executes cutting. This invention employs a closed-loop control logic of semi-finishing, cleaning, measurement, compensation, and finishing, solving the impact of clamping deformation, environmental interference, and cutting chatter on machining accuracy, and improving the dimensional consistency of the finished product.

Description

Technical Field

[0001] This invention relates to the field of new energy vehicle motor manufacturing technology, and in particular to a milling control system for the rear end cover of a new energy motor. Background Technology

[0002] As a core component of the motor assembly, the rear end cover of a new energy motor directly affects the assembly accuracy and operational stability of the motor due to its dimensional tolerances and surface quality. In existing CNC milling processes, workpiece clamping stability is fundamental to ensuring machining accuracy. Traditional hydraulic or pneumatic clamping systems typically determine workpiece clamping based solely on whether the pipeline pressure reaches a set threshold. However, in actual operation, if there are foreign objects such as chips between the positioning surface and the workpiece, or if the workpiece is slightly tilted, the system, while detecting that the pressure is within the threshold, cannot identify the contact surface's fit. This "false clamping" state leads to a shift in the machining datum, resulting in dimensional errors that are difficult to correct in subsequent milling operations.

[0003] Furthermore, the interior of CNC machine tools is filled with high-pressure cutting fluid, oil mist, and metal dust. This harsh machining environment poses a challenge to in-machine inspection technology. Existing in-machine measurement solutions often lack targeted environmental pretreatment and sensor protection mechanisms, making optical sensors susceptible to oil contamination or erroneous data collection due to residual cutting fluid on the workpiece surface. Simultaneously, conventional process flows often employ a continuous roughing and finishing process, lacking independent inspection and feedback steps in the intermediate stages. This results in elastic deformation and residual errors generated during roughing being directly transmitted to the finishing process, making targeted error compensation impossible.

[0004] On the other hand, the rear end cover of the motor is mostly a thin-walled structure, which is prone to regenerative chatter during high-speed milling, affecting surface roughness. Existing control systems mostly use constant speed machining and lack active vibration suppression strategies based on load feedback. At the same time, during long-term batch processing, thermal expansion of the machine tool spindle and progressive wear of the tool can cause systematic dimensional drift. Existing technologies usually use static compensation methods, which are difficult to dynamically adjust compensation parameters according to real-time temperature changes and historical machining trends, resulting in poor dimensional consistency of workpieces in the same batch (low CPK value), making it difficult to meet the manufacturing requirements of high-precision motors. Summary of the Invention

[0005] To address the issues of uncontrollable workpiece clamping status, interference from the processing environment on measurement accuracy, and lack of real-time compensation mechanisms for thermal deformation and cutting chatter in the processing of rear end covers for existing new energy motors, which result in poor dimensional consistency and unstable surface quality of finished products.

[0006] To address the above problems, the present invention provides the following technical solution:

[0007] This invention provides a milling control system for the rear end cover of a new energy motor, which includes a central control module, a positioning detection module, an environmental preprocessing module, a vision detection module, a laser detection module, a milling execution module, and a data storage module; the central control module establishes communication connections with each of the above modules respectively.

[0008] The positioning detection module monitors the workpiece clamping status and transmits the clamping position signal to the central control module; the environmental preprocessing module cleans the workpiece surface before the inspection process; the vision detection module and the laser detection module collect workpiece geometric feature data; the data storage module stores machining standards and algorithm models for the central control module to access; the central control module compares the geometric feature data with the machining standards and generates tool compensation parameters; the milling execution module adjusts the machining instructions using the tool compensation parameters.

[0009] The positioning detection module includes a pneumatic pressure sensing nozzle, a pneumatic pipeline, and a digital pressure sensor unit integrated into the positioning support surface of the fixture; the digital pressure sensor unit collects the static pressure value inside the pneumatic pipeline; the central control module generates the clamping signal when the static pressure value reaches a preset pressure threshold and the pressure fluctuation rate is lower than a preset stability threshold.

[0010] The environmental pretreatment module includes a high-pressure pneumatic purging assembly and a sensor protection assembly; the high-pressure pneumatic purging assembly includes a directional air knife fixed to the side of the spindle, and the air outlet angle of the directional air knife forms a preset angle with the surface of the workpiece to be measured; the sensor protection assembly includes a sealed protective housing and a fast-opening and closing shutter set at the light transmission window, and the vision inspection module and the laser inspection module are encapsulated in the sealed protective housing.

[0011] The vision inspection module includes an industrial area array camera, dual telecentric optical lenses, and a composite illumination unit, used to identify workpiece surface defects and extract the center pixel coordinates of the main reference hole. The central control module converts the center pixel coordinates into physical coordinates in the machine tool machining coordinate system. The laser inspection module includes a laser displacement sensor and a servo-driven telescopic slide. The servo-driven telescopic slide drives the laser displacement sensor to switch positions. The central control module converts the collected local distance values ​​into absolute spatial coordinates in the machine tool machining coordinate system.

[0012] The data storage module stores the thermal error calculation model, the material linear expansion coefficient, and the historical machining error trend table; the central control module uses the thermal error calculation model to restore the measured size to the equivalent size at the standard reference temperature and calculates the dimensional scalar error; the central control module performs a weighted moving average calculation on the historical workpiece size deviation sequence to obtain the systematic drift trend, and adds the systematic drift trend to the tool compensation parameters.

[0013] The milling execution module collects the load current of the spindle servo motor and calculates the load current variance. When the load current variance exceeds a preset stability threshold, it triggers a spindle speed variation strategy to control the spindle speed to fluctuate in a sinusoidal shape. The milling execution module has an embedded load adaptive control algorithm, which estimates the instantaneous cutting load rate through a load observer and compares it with the preset optimal spindle load rate. The feed rate is adjusted according to the comparison result.

[0014] The central control module performs the following process control: in response to the clamping signal, it controls the milling execution module to perform semi-finishing; after the semi-finishing is completed, it controls the environmental pretreatment module to start airflow purging; after the airflow is turned off and a preset resting time has elapsed, it controls the vision inspection module and the laser inspection module to perform inspection; based on the inspection results, it generates the tool compensation parameters for the finishing process and controls the milling execution module to perform finishing.

[0015] In summary, the present invention has at least one of the following beneficial technical effects:

[0016] 1. This invention monitors the static pressure and pressure fluctuation rate in the air pipeline of the fixture in real time through the positioning detection module. It generates a clamping signal only when the static pressure reaches a preset threshold and the fluctuation rate is lower than a stable threshold. It identifies the fit between the workpiece and the positioning surface through dual judgment logic, preventing machining reference deviation caused by chip padding or clamping tilt, and ensuring the positioning accuracy of milling operations.

[0017] 2. This invention employs a process control logic that involves semi-finishing followed by directional air knife purging, then in-machine testing and feedback compensation before finishing. Combined with the sealed housing of the sensor protection component and the fast-opening and closing shutter, it isolates the interference of cutting fluid mist and chips on the vision and laser measurement units, ensuring the accuracy of in-machine measurement data. Furthermore, it automatically corrects the residual machining errors of the previous process through closed-loop feedback between processes.

[0018] 3. This invention utilizes the spindle servo load current variance to trigger a sinusoidal wave fluctuation strategy for rotational speed, and combines a thermal error model with a historical machining error trend table for multidimensional compensation; it actively disrupts the phase condition of regenerative chatter, suppressing the impact of cutting vibration on surface quality; on the one hand, it incorporates material deformation caused by temperature and systematic drift caused by tool wear into the compensation range, improving the dimensional consistency of batch processing. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the overall architecture of the new energy motor rear cover milling control system of the present invention;

[0020] Figure 2 This is a flowchart illustrating the error analysis logic and dynamic compensation strategy of the present invention.

[0021] Figure 3 This is a comparison chart of the closed-loop compensation effect for machining accuracy of the present invention;

[0022] Figure 4 This is a schematic diagram of the spindle speed sinusoidal fluctuation control strategy of the present invention;

[0023] Figure 5 This is a time-domain signal comparison diagram of the cutting chatter suppression effect of the present invention.

[0024] The module comprises: 1. Central control module; 2. Positioning and detection module; 3. Environmental preprocessing module; 4. Vision inspection module; 5. Laser inspection module; 6. Milling execution module; and 7. Data storage module. Detailed Implementation

[0025] The following is in conjunction with the appendix Figure 1 -Appendix Figure 5 The present invention will be further described in detail below.

[0026] This invention provides a milling control system for the rear end cover of a new energy motor.

[0027] like Figure 1 As shown, a milling control system for the rear end cover of a new energy motor is applied to the precision manufacturing process of the rear end cover of a new energy vehicle motor. It can realize closed-loop control of the workpiece from clamping and positioning, multi-stage milling, environmental cleaning, in-situ detection and error compensation.

[0028] The milling control system for the rear end cover of the new energy motor includes a central control module 1, a positioning detection module 2, an environmental preprocessing module 3, a vision detection module 4, a laser detection module 5, a milling execution module 6, and a data storage module 7. The central control module 1 is electrically connected to the positioning detection module 2, the environmental preprocessing module 3, the vision detection module 4, the laser detection module 5, the milling execution module 6, and the data storage module 7, respectively, for data acquisition, processing, and the issuance of control commands.

[0029] The positioning detection module 2 is located within the workpiece clamping system and is used to monitor the physical clamping status of the workpiece in real time. After detecting that the workpiece is correctly positioned and clamped, the positioning detection module 2 generates a clamping position signal and transmits it to the central control module 1. This signal serves as a prerequisite for the system to start subsequent processing flows.

[0030] The milling execution module 6 is connected to the drive system of the CNC machining center and is used to receive machining commands from the central control module 1. The milling execution module 6 can adjust the spindle speed, feed rate and tool path according to the commands, and perform cutting operations on the workpiece according to the process logic of roughing, semi-finishing and finishing.

[0031] The environmental pretreatment module 3 is located inside the machining area to improve the inspection environment. Before the milling process is paused and the inspection process begins, the environmental pretreatment module 3 performs a cleaning action on the surface of the workpiece to be tested, removing residual cutting fluid and metal debris, and controls the protective devices of the relevant optical sensors to be activated to ensure that the optical inspection path is unobstructed.

[0032] The vision inspection module 4 is used to acquire image data of the workpiece surface and identify casting defects and rough feature contours. The laser inspection module 5 is used to acquire depth and position data of key geometric features of the workpiece. The laser inspection module 5 includes a laser displacement sensor and a coordinate adjustment submodule. The coordinate adjustment submodule is used to adjust the spatial position and measurement angle of the laser displacement sensor, enabling the laser displacement sensor to perform non-contact measurement on the surface of the semi-finished workpiece.

[0033] Data storage module 7 is used to store the workpiece's three-dimensional CAD model data, preset machining accuracy standards, semi-finishing allowance parameters, tool compensation algorithm model, and real-time detection data during the machining process.

[0034] In the specific workflow, after receiving the clamping signal from the positioning detection module 2, the central control module 1 first controls the milling execution module 6 to perform roughing and semi-finishing operations. During the semi-finishing stage, the central control module 1 controls the milling execution module 6 to retain a preset unilateral safety margin, so that the workpiece size is in an intermediate state close to the final design size.

[0035] After the semi-finishing process is completed, the central control module 1 activates the environmental preprocessing module 3 to purge the key areas of the workpiece with high-pressure airflow. After cleaning, the central control module 1 controls the vision inspection module 4 and the laser inspection module 5 to perform in-situ inspection of the workpiece and obtain the actual geometric data of the current semi-finished product.

[0036] The central control module 1 compares the collected actual geometric data of the semi-finished product with the final design standard in the data storage module 7. Based on the comparison results, the central control module 1 calculates the actual remaining machining amount for each part of the workpiece, and generates tool compensation parameters and corrected tool paths for the finishing process based on the actual remaining machining amount.

[0037] The central control module 1 sends the generated tool compensation parameters and the corrected toolpath to the milling execution module 6. The milling execution module 6 performs the finishing operation according to the received parameters, removes the remaining material, and completes the final dimensional correction of the workpiece.

[0038] In this embodiment, the environmental preprocessing module 3 is constructed as a physical barrier and conversion mechanism between the milling machining environment and the precision optical inspection environment. It mainly consists of a high-pressure pneumatic purging assembly and a sensor protection assembly, aiming to solve the problem of interference from cutting fluid oil film and metal chips on non-contact measurement.

[0039] The high-pressure pneumatic purging assembly is configured to generate a high-energy air curtain with a specific fluid profile. The assembly's air source interface unit connects to the workshop's air supply network and uses a precision pressure regulating valve to stabilize the input air pressure within a preset operating pressure range. Typically, this pressure range is set between 0.4 MPa and 0.6 MPa to ensure the airflow has sufficient kinetic energy to peel off the cutting fluid film adhering to the workpiece surface, while avoiding excessive pressure that could cause slight workpiece displacement. The electromagnetic control valve assembly acts as the actuator, responding to commands from the central control module 1 to control the opening and closing of the air path. The directional air knife is rigidly fixed to the side of the spindle, and its internal flow channel design follows the Coanda effect. The outlet slit width is set between 0.05 mm and 0.1 mm, allowing the compressed air to be rectified into a laminar, sheet-like high-speed air curtain. The outlet angle of the directional air knife is set to an angle of 30° to 45° with the workpiece surface to be measured, using shearing force rather than vertical impact force to remove surface deposits. This angle setting minimizes the risk of liquid splashing and backflow.

[0040] The core of the sensor protection component lies in providing physical isolation for precision optical instruments. The sealed protective housing is integrally machined from aluminum alloy, boasting an IP67 protection rating, and completely encloses the vision inspection module 4 and the laser inspection module 5. Corresponding to the direction of light projection, the front end of the sealed protective housing has a light-transmitting window. A rapid-opening and closing shutter is mounted on the outside of this window, driven by a miniature cylinder, and, in conjunction with an embedded fluororubber sealing strip, forms a gas-liquid seal when closed.

[0041] Regarding the collaborative control logic, the central control module 1 does not simply trigger each component sequentially, but rather executes a timing control strategy based on state monitoring. When the semi-finishing process terminates and the spindle completely stops, the central control module 1 first activates the electromagnetic control valve group, driving the directional air knife to perform a process on the workpiece surface for a duration of [duration missing]. High-pressure purging. Here, The value of is not fixed, but is based on the surface area of ​​the area to be measured on the workpiece. With air knife effective coverage width It is determined that the following relationship is satisfied:

[0042] ;

[0043] In the formula, This is the scanning speed of the air knife relative to the workpiece (if the air knife is fixed, then the effective diffusion speed of the fluid is taken). As a safety redundancy factor, the value is typically between 1.2 and 1.5 to ensure that the residual cutting fluid rate is below the allowable threshold for optical detection.

[0044] It is worth noting that, in order to prevent secondary contamination of the optical window by splashing droplets during the purging process, the central control module 1 is configured to execute delayed opening logic: only opening the window after the purging time reaches a certain threshold. And the settling time after the airflow is completely shut off Only then is an opening command sent to the sensor protection component to drive the rapid opening / closing door 322 to open. The aerosol settling time is typically set to 1 to 2 seconds. This logic ensures that the visual inspection module 4 and the laser inspection module 5 are only exposed to and perform data acquisition in a clean and stable air environment, thereby guaranteeing the signal-to-noise ratio and accuracy of subsequent measurement data.

[0045] In this embodiment, the positioning detection module 2 is constructed as an embedded airtightness monitoring subsystem. Based on the orifice throttling principle in fluid dynamics, it quantitatively evaluates the microscopic fit between the bottom surface of the rear cover of the new energy motor and the positioning reference surface of the fixture by monitoring the change of air back pressure.

[0046] Specifically, the positioning detection module 2 mainly consists of a pneumatic pressure sensing nozzle, a precision air passage, and a digital pressure sensor unit integrated within the fixture's positioning support surface. Several vertical air vents are precision-machined on the main positioning surface of the fixture, which supports the workpiece end face. These air vents form the end openings of the pneumatic pressure sensing nozzle, with a diameter... The setting is typically 1.0mm to 1.5mm to ensure smooth airflow without significant back pressure when no workpiece is placed. Each pressure-sensing nozzle converges through an internal flow channel to a digital pressure sensor unit, which is configured to acquire the static pressure value within the air path in real time at a high-frequency sampling rate of at least 50Hz. The data is transmitted to the central control module 1 via an industrial fieldbus.

[0047] Its detection logic is based on the following physical mechanism: when the workpiece is correctly clamped, the workpiece's reference surface will cover the outlet of the pneumatic sensing nozzle. If there is a tiny gap between the workpiece and the positioning surface caused by cutting chips or burrs... When airflow leaks through this gap, it creates flow resistance. According to fluid mechanics principles, the back pressure within the air passage... and gap It exhibits a non-linear inverse proportional relationship. When the gap... When the pressure approaches zero (i.e., when the fit is tight), the flow resistance is at its maximum, and the back pressure in the air passage is at its highest. It rose rapidly and approached the supply pressure. Conversely, if there is a positioning error, the amount of air leakage will increase, resulting in a significant decrease in back pressure.

[0048] To eliminate misjudgments caused by gas source fluctuations and instantaneous pipeline impacts, and to ensure the robustness of clamping status determination, the central control module 1 internally runs a dual verification algorithm based on threshold and time-domain stability. The system only monitors pressure in real time. The clamping position signal is generated and the start lockout of the milling execution module 6 is released only when the following steady-state determination conditions are met simultaneously:

[0049] ;

[0050] ;

[0051] In the formula, Standard gas supply pressure; This is the positioning determination coefficient, and its value range is set from 0.85 to 0.95. The determination is based on pre-calibrated pressure and clearance characteristic curves, corresponding to the maximum positioning clearance allowed by the process. The critical pressure ratio (typically 0.01 mm to 0.02 mm) is required. The time window for determination is typically set between 200ms and 500ms. For integration variables; This is the pressure fluctuation rate threshold, used to ensure that the gas path pressure has been established in equilibrium and without dynamic fluctuations, preventing false triggering caused by transient values ​​during the pressure establishment process.

[0052] Furthermore, as a comprehensive safety mechanism, the positioning detection module 2 remains continuously active during the milling process. The central control module 1 is configured to execute dynamic monitoring logic: if any abnormality is detected during the cutting process... If a sudden change occurs and the workpiece falls below the safety threshold, it indicates that the workpiece has become loose or displaced under the cutting force. The system will immediately trigger an emergency stop command, forcing the spindle to stop and retract the tool, thereby avoiding batch scrapping or equipment collision accidents caused by workpiece displacement.

[0053] In this embodiment, the visual inspection module 4 is constructed as the macroscopic feature perception unit of the system. It uses photoelectric conversion and image processing algorithms to achieve transient screening of casting defects on the surface of the rear cover of the new energy motor and global coordinate guidance for subsequent precision measurement.

[0054] The hardware architecture of the vision inspection module 4 is specifically configured based on the imaging characteristics of highly reflective aluminum alloy surfaces. Its core components include a high-resolution industrial area scan camera, dual telecentric optical lenses, and a composite illumination unit. The industrial area scan camera preferably uses a global shutter CMOS sensor with a pixel resolution of no less than 20 megapixels. This ensures that while covering the entire workpiece end face in a wide field of view, the single-pixel accuracy remains at the 50μm to 100μm level, thus capturing minute surface defects. The dual telecentric optical lenses feature principal rays that are parallel to the optical axis on both the object and image sides. This design eliminates parallax distortion, ensuring a constant magnification across different height steps on the workpiece, thereby avoiding measurement errors caused by workpiece thickness tolerances. The composite illumination unit employs a combination topology of a coaxial light source and a low-angle ring light source. The coaxial light source illuminates smooth surfaces, making recessed sand holes appear as high-contrast dark spots; the low-angle ring light source highlights edge contours and surface scratches. Both are mixed and controlled through independent PWM dimming channels to adapt to the imaging requirements of different surface roughnesses.

[0055] Regarding defect identification logic, the central control module 1 does not rely solely on a single grayscale threshold, but instead runs a Blob analysis algorithm based on the fusion of local contrast and geometric features. After acquiring the original image and performing Gaussian filtering for noise reduction, the system defines the mathematical model for defect determination as follows: For any connected component in the image... Calculate its average gray value grayscale values ​​of local background areas Weber contrast :

[0056] ;

[0057] In the formula, To prevent tiny positive numbers with a denominator of zero (e.g.) ), used to handle computational stability against a pure black background.

[0058] Simultaneously, the system calculates the geometric area of ​​the connected component. and roundness characteristics ( ,in, (Circumference). The system sets a contrast threshold. (Value range 0.2 to 0.4) Area threshold (Value range 0.5mm) 2 Up to 1.0mm 2） and roundness threshold Only if the condition is met. At that time, the region is marked as a candidate defect; furthermore, if If the result matches the porosity characteristics (i.e., close to 1.0), it is determined to be a casting porosity defect. This multi-dimensional judgment logic effectively distinguishes between real defects and surface residual oil stains (which usually have low contrast and irregular shapes).

[0059] In terms of coordinate guidance logic, the visual inspection module 4 uses a sub-pixel edge extraction algorithm (such as the Canny operator combined with least squares fitting) to obtain the center pixel coordinates of the main reference hole of the workpiece in the image coordinate system. To convert these two-dimensional pixel coordinates into three-dimensional physical coordinates in the machine tool machining coordinate system. The system pre-established the affine transformation relationship from the image plane to the machine tool plane using a nine-point calibration method. Considering the relatively small linear distortion of the dual telecentric lenses, its coordinate mapping model is simplified as follows:

[0060] ;

[0061] In the formula, the matrix The vector represents the rotation and scaling relationship between the camera coordinate system and the machine tool coordinate system. This characterizes the translational deviation between the two coordinate origins. All of the above parameters are fixed in the data storage module 7 through a calibration process. Based on this mapping model, the central control module 1 calculates the actual center position of the workpiece and its rotational deviation relative to the theoretical clamping angle. The corrected coordinate data is then sent to the laser detection module 5, guiding it to move directly into the small neighborhood of the feature to be measured.

[0062] In this embodiment, the laser detection module 5 is configured as the system's micro-precision measurement unit. Based on the optical triangulation principle or the spectral confocal principle, it performs non-contact, high-precision sampling of the key features of the workpiece (such as the inner diameter of the bearing chamber, the depth of the stop, and the flatness of the end face) after coarse positioning by the vision module.

[0063] The hardware architecture of the laser detection module 5 includes a laser displacement sensor, a coordinate adjustment submodule, and a signal conditioning circuit. To overcome the interference of high reflectivity and anisotropic textures on the aluminum alloy surface commonly found on the rear end cover of new energy motors on optical measurements, a blue-violet dot laser displacement sensor with a center wavelength of 405nm is preferred. Compared to traditional red light sensors, short-wavelength lasers exhibit better diffuse reflection characteristics on metal surfaces, significantly reducing spot noise caused by specular reflection, achieving a Z-axis repeatability accuracy of 0.5μm to 1.0μm. The coordinate adjustment submodule, as a key actuator for enabling flexible sensor intervention in the machining area, is specifically constructed as a servo-driven telescopic slide mounted on the side of the machine tool spindle box. This slide has two physical states: a measurement position and a clearance position. During milling, the slide retracts to the clearance position, completely removing the laser displacement sensor from the cutting fluid splash range and the tool's motion envelope.

[0064] During the testing phase, the slide table drives the sensor to extend to the preset measurement position, sending the sensor probe into the travel range of the machine tool spindle.

[0065] To achieve precise alignment between the measurement data and the machining coordinate system, the central control module 1 incorporates a unified coordinate mathematical model. Once the coordinate adjustment submodule moves the sensor to a fixed position, the raw data acquired by the laser displacement sensor is only a local one-dimensional distance value relative to the sensor probe itself. To convert it into absolute spatial coordinates in the machine tool machining coordinate system. The system performs real-time calculations based on the following rigid body kinematic transformation equations:

[0066] ;

[0067] In the formula, The real-time position vector of the current machine tool spindle center in the machine tool coordinate system. This data is directly provided by the grating ruler feedback of the CNC system, and the system eliminates the time lag between machine tool movement and sensor sampling through a timestamp alignment algorithm; The pre-calibrated rotation matrix (3×3) characterizes the angular deviation of the sensor's optical axis relative to the main axis. This is the offset vector of the robotic arm length in the extended state for the coordinate adjustment submodule; The local measurement vector output by the sensor This model enables the system to eliminate geometric errors caused by the sensor's installation location.

[0068] For measuring the bearing chamber diameter of the rear cover of a new energy motor, the central control module 1 controls the machine tool spindle to drive the sensor along a circular trajectory, executing a multi-point sampling strategy. Since residual cutting fluid droplets or tiny metal dust are unavoidable at the machining site, direct fitting may lead to errors. Therefore, this embodiment employs a two-stage data processing logic of coarse filtering and fine fitting. The system first collects a series of discrete spatial point cloud data. ,in, The total number of sampling points (usually) Subsequently, outlier removal was performed using a random sample consensus algorithm: a subset was randomly selected to estimate the circle parameters, and the Euclidean distance from all points to the circle was calculated. ,like If the value is not specified, then the point is marked as a noise point. The threshold value is... The value is set according to the sensor's nominal accuracy and allowable surface roughness, and is usually between 0.005mm and 0.01mm.

[0069] After removing noise points, the system uses the least squares method to fit the actual circular contour to the remaining set of interior points. The objective function for fitting is... Defined as minimizing the sum of squared distances from all valid sampling points to the fitted circle:

[0070] ;

[0071] In the formula, The number of interior points, Let the coordinates of the actual center of the circle be the coordinates to be determined. Let be the actual aperture radius to be determined. By solving for the extrema of this objective function (i.e., when the partial derivatives are zero), the system obtains high-confidence geometric parameters. It should be noted that, to ensure numerical stability of the calculation, the spatial distribution of the sampling points must cover an arc of at least 180 degrees to avoid matrix singularities causing solution divergence.

[0072] Finally, the central control module 1 will calculate the actual radius. Theoretical design values ​​in data storage module 7 Perform a comparison. If the deviation value... Within the preset compensable range (e.g., ±0.05mm), the system will automatically generate the corresponding tool compensation value and proceed to the next process; if the deviation exceeds the safety threshold, the workpiece is determined to be scrapped in the semi-finishing stage, the system outputs an abnormal signal and terminates the process.

[0073] In this embodiment, the data storage module 7 is constructed as the core knowledge base and parameter reference source of the entire system, aiming to provide standardized data support for high-frequency read and write operations for closed-loop control. Its physical layer preferably adopts industrial-grade non-volatile memory with shock-resistant characteristics to adapt to the high-frequency vibration environment of the workshop; the logic layer is built based on a real-time relational database, which not only stores static geometric parameters, but also pre-sets a digital process twin model covering the physical properties of materials, the definition of process intermediate states, and dynamic compensation strategies.

[0074] Specifically, data storage module 7 is internally configured with a 3D model feature mapping library. This library parses the CAD source file (such as STEP format) of the motor rear end cover, converting the geometric features of the physical workpiece into a machine-readable set of vector parameters. For the key mating features of the motor rear end cover, namely the bearing housing and the stop, the feature mapping library stores their theoretical center coordinate vectors in the machine tool coordinate system. Theoretical nominal radius and the design tolerance range The system defines the theoretical target radius for the semi-finishing stage using the following formula. :

[0075] ;

[0076] In the formula, The preset allowance for semi-finished single-sided solid is typically set to 0.05mm to 0.15mm.

[0077] This parameter The thickness is not arbitrarily chosen, but is set based on the cutting characteristics of aluminum alloy (such as ADC12) materials: the thickness must be greater than the minimum cutting thickness of the finishing tool to avoid slippage or tool deflection, while it must be small enough to limit the cutting heat generated during the finishing process and prevent secondary thermal deformation of the workpiece.

[0078] Considering that the rear end cover of new energy motors is generally made of die-cast aluminum alloy, its coefficient of linear expansion is relatively large (approximately 2.3 × 10⁻⁶). −5 (°C), ambient temperature fluctuations have a significant impact on micron-level processing. Therefore, the data storage module 7 has a built-in thermal error normalization model.

[0079] When the central control module 1 receives laser detection data, it calls the model to restore the measured value at the current temperature to the equivalent size at the standard reference temperature (20℃). The normalization calculation formula is as follows:

[0080] ;

[0081] In the formula, This is the normalized standard diameter; This represents the measured diameter of the sensor under the current environment. The coefficient of linear expansion of the material is stored in the database. For ADC12 aluminum alloy, this value is taken as [value missing]. ; The real-time workpiece temperature is fed back by temperature sensors arranged on the surface of the fixture; This serves as the standard reference temperature. The introduction of this model ensures that the system's evaluation criteria for processing errors remain consistent even under conditions where workshop temperatures are not constant.

[0082] Furthermore, to achieve adaptive tool wear compensation, data storage module 7 maintains a dynamically updated historical machining error trend table. This table records the most recent... One (e.g.) Final dimensional deviation sequence of the workpiece after finishing The central control module 1 uses a weighted moving average algorithm to calculate the current systemic drift trend. :

[0083] ;

[0084] In the formula, For the first Dimensional deviation of each historical workpiece (measured value minus theoretical value); Let be the weight coefficient, and satisfy... That is, samples closer to the current moment are given greater weight in order to keenly capture the accelerating trend of tool wear.

[0085] Calculated by this algorithm The tool compensation parameters are directly fed back to the next machining cycle. After finishing (or based on the statistical data of semi-finishing), the tool life database is updated for the initial tool compensation prediction of the next piece. This realizes the upgrade of the control strategy from post-out-of-tolerance alarm to pre-event trend compensation, effectively improving the process capability index in the mass production process.

[0086] like Figure 1 As shown, in this embodiment, the central control module 1 integrates an error analysis and dynamic compensation unit. This unit, as the decision-making core of the closed-loop control system, is configured to receive normalized measurement data from the laser detection module 5, perform differential analysis based on the process standards in the data storage module 7, and finally generate executable compensation instructions for the machine tool CNC system.

[0087] Specifically, the error analysis and dynamic compensation unit performs multi-dimensional error decoupling operations. Considering that the bearing housing of the rear end cover of the new energy motor has a high-precision inner contour feature, the system decomposes the measurement deviation into two independent components: dimensional scalar error and position vector error. Let the measured diameter after temperature normalization be... The semi-finishing target diameter defined in data storage module 7 is Then the scalar error of the dimension Defined as:

[0088] ;

[0089] Meanwhile, let the measured coordinates of the center of the circle be... The theoretical reference coordinates are Then the position vector error Defined as:

[0090] ;

[0091] Based on the above decoupling, the system will Mapped to the tool wear position register of the CNC system to compensate for the radius reduction caused by tool wear;

[0092] Will Mapped to the workpiece coordinate system origin offset register (such as G54 or G55) to eliminate coaxiality errors caused by fixture thermal expansion or repeated positioning deviations.

[0093] To suppress high-frequency random noise caused by residual cutting fluid or microscopic surface roughness, the error analysis and dynamic compensation unit uses an exponentially weighted moving average algorithm to calculate the smoothed trend error. This algorithm, by assigning decreasing weights to historical data, can accurately capture tool wear trends while filtering out abrupt changes in a single measurement. Its discretization calculation formula is as follows:

[0094] ;

[0095] In the formula, Indicates the serial number of the workpiece currently being processed; As a smoothing factor, its value range is set from 0.2 to 0.4. This value is based on the following: If The system's response to sudden anomalies such as tool chipping is delayed; if If this is not done, the noise cannot be effectively smoothed, which may cause the servo system to oscillate.

[0096] In obtaining smoothing error Then, the system generates the final tool compensation increment based on the three-level logic of dead zone suppression, proportional adjustment, and ultimate circuit breaking. For internal hole machining scenarios (i.e., increasing the cutting amount is required if the hole diameter is too small, and triggering an alarm if the hole diameter is too large), the control model is described as follows:

[0097] ;

[0098] In the formula, The insensitive dead zone threshold is usually set to the repeatability of the machine tool feed axis (e.g., 0.003 mm). When the error is within this range, the system maintains the current state to avoid system lag caused by frequent fine-tuning. This is the proportional gain coefficient, ranging from 0.6 to 0.8. A coefficient less than 1 is intended to achieve undercompensation control, preventing aperture deviation due to overcompensation in a gradual approximation manner. The safety melting threshold is set (e.g., 0.05mm). When the error exceeds this value, it is judged as abnormal wear (e.g., tool breakage or clamping failure), and the system immediately triggers a shutdown alarm.

[0099] Ultimately, the generated compensation increment and position correction amount The parameters are written directly to the corresponding registers of the CNC system via industrial Ethernet. Before the start of the next workpiece machining cycle, the machine tool controller automatically loads the updated parameters, thereby achieving closed-loop self-healing of machining accuracy at the physical level.

[0100] In this embodiment, the milling execution module 6 is constructed as a terminal execution unit that directly drives the physical motion of the machine tool. It establishes bidirectional real-time communication with the machine tool CNC system through a high-speed industrial fieldbus and a dedicated API interface. This module is not only responsible for executing the static dimension compensation instructions generated by the error analysis unit, but also has embedded load adaptive control and chatter suppression algorithms, which can proactively cope with the machining instability caused by the thin-walled and weakly rigid structure of the rear cover of the new energy motor.

[0101] Specifically, the closed-loop control logic of milling execution module 6 includes two levels: static geometry correction and dynamic mechanical optimization. In the static geometry correction level, the system fine-tunes the toolpath by modifying macro variables of the CNC system. When it receives tool compensation increments from the error analysis unit... and position vector error Subsequently, the milling execution module 6 does not directly modify the compiled G-code program, but instead uses synchronous read / write instructions to... Accumulate to the tool radius offset register (the storage area corresponding to the D code), and then... The component values ​​are written to the currently active workpiece coordinate system offset register (such as the X / Y axis offset of G54). This processing method ensures that the next toolpath automatically includes the correction amount without interrupting the machining cycle, thereby eliminating the time loss caused by traditional stop-and-set tooling.

[0102] In the dynamic mechanics optimization layer, considering that there may be pores or hard spots inside the motor rear end cover blank (usually ADC12 die-cast aluminum), which may cause nonlinear fluctuations in material removal rate and cutting force, the milling execution module 6 collects the load current of the spindle servo motor in real time. The system estimates the instantaneous cutting load rate through a built-in load observer and dynamically adjusts the feed rate based on the constant power cutting principle. Its adaptive control algorithm is as follows:

[0103] ;

[0104] In the formula, This is the feed rate command adjusted in real time; The baseline feed rate generated for CAM software; The feed rate factor is calculated using a reciprocal mapping model with dead zone and amplitude limit:

[0105] ;

[0106] In the formula, The preset optimal spindle load rate is typically set to 40% to 60% of the rated load to keep the tool in the optimal cutting efficiency range; The real-time spindle load rate is processed by moving average filtering. The filtering is introduced to eliminate high-frequency electromagnetic noise in the current signal. To prevent small positive numbers with a denominator of zero (such as 10) -3） ; (e.g., 0.5) and (e.g., 1.2) represent the lower and upper limits of the scaling factor adjustment, respectively. The physical meaning of this algorithm is: when the cutting load... When the set value is exceeded (e.g., when encountering hard points), the system automatically reduces the feed rate to prevent chipping; when the load is lower than the set value (e.g., due to cutting air or insufficient allowance), the system automatically increases the feed rate to shorten the machining cycle.

[0107] Furthermore, to address the regenerative chatter that is prone to occur during the machining of thin-walled parts, the milling execution module 6 integrates vibration suppression logic based on frequency domain analysis. The system monitors the AC component of the spindle load current at a high frequency (e.g., 1kHz) and within the sliding time window... Calculate its variance .when Exceeding the preset stability threshold When this occurs, it indicates that the system has entered an unstable cutting region. At this point, the module automatically triggers the spindle speed variation (SSV) strategy, modulating the spindle speed with a sine wave. :

[0108] ;

[0109] In the formula, This is the nominal spindle speed; To modulate the amplitude by rotational speed, the value is usually taken as... 5% to 10%; The modulation frequency (e.g., 0.5 Hz to 2 Hz).

[0110] The physical mechanism of this strategy is as follows: by periodically changing the spindle speed, the phase lag effect of regenerative chatter is continuously destroyed, so that the cutting parameter trajectory of the machining process can never stay in the unstable region of the stability lobe diagram, thereby effectively avoiding the formation of fish scale defects on the workpiece surface caused by self-excited vibration, and extending tool life while ensuring surface quality.

[0111] To verify the effectiveness of the present invention, a specific application embodiment is given below. This embodiment selects the rear end cover of a certain model of new energy vehicle drive motor as the processing object. The workpiece material is ADC12 die-cast aluminum alloy (linear expansion coefficient 2.3 × 10⁻⁶). −5 / ℃), its key machining feature is the central bearing chamber, with a designed diameter of The diameter is 47.000 mm, the tolerance zone is 0 to +0.015 mm, and the surface roughness requirement is... The experiment was conducted on a vertical machining center with a maximum speed of 12,000 rpm. A control group and an experimental group were set up to compare and verify the continuous production of 50 workpieces. The control group adopted a traditional open-loop machining process, with manual tool setting only during the first trial cut, and subsequent machining using a fixed program, with manual sampling every 10 pieces. The experimental group used a fully closed-loop control system, which included environmental pretreatment, laser in-situ detection, thermal error normalization, and SSV chatter suppression functions.

[0112] In the specific experimental process, the experimental group system executed the following closed-loop logic: After the semi-finishing was completed, the environmental pretreatment module first performed a high-pressure purging at 0.5MPa for 2 seconds to remove the surface oil film; then data acquisition and processing were performed. For example, when inspecting the 25th workpiece, the laser sensor measured the actual diameter as 46.955mm. At this time, the sensor reported the workpiece temperature as 35℃. The system used the thermal error normalization model to calculate the equivalent size at the standard temperature (20℃) as 46.939mm; based on this data, the system calculated the dimensional error. The tool compensation value is set at -0.061mm (meaning there is still 0.061mm of material left to be removed), and is automatically generated based on historical trends for the finishing process. Write to the CNC system register. Furthermore, during the machining of the 40th workpiece, due to tool wear causing a change in cutting conditions, the system detected spindle load variance. If the preset threshold is exceeded, the milling execution module immediately triggers the SSV strategy, controlling the spindle speed to fluctuate sinusoidally at a frequency of 1.0Hz within the range of 8000±400rpm, in order to disrupt regenerative chatter.

[0113] like Figure 3The graph shows a comparison of the closed-loop compensation effect in machining accuracy. The dashed line represents the control group (traditional open-loop machining), and the solid line represents the experimental group (closed-loop control of this invention). As can be seen from the graph, in the control group (traditional open-loop machining), the workpiece hole diameter shows a monotonically decreasing trend as the number of machining operations increases, affected by tool wear and machine tool thermal expansion. Starting from the 21st piece, the size has fallen below the lower tolerance (47.000 mm), leading to the scrapping of subsequent batches of products. In contrast, the data curve of the experimental group (closed-loop control of this invention) shows obvious sawtooth fluctuations, indicating that the system detects errors accumulating close to the dead zone threshold (…). When this occurs, the compensation logic is triggered to pull the dimensions back to the target center value. Final statistical results show that the dimensions of all 50 workpieces in the experimental group (closed-loop control of this invention) are stably controlled within the golden tolerance zone of 47.004-47.012 mm, and the process capability index significantly improved from 0.63 in the control group to 1.85, effectively verifying the... Figure 3 The stability of the error compensation strategy in mass production.

[0114] Figure 4 The time-domain variation of the spindle speed is shown. Figure 5 The corresponding cutting vibration amplitudes are shown. Analysis reveals that during the constant rotational speed (8000 rpm) phase from 0 to 5 seconds, the vibration energy trend and real-time vibration signal increase sharply, indicating regenerative chatter in the thin-walled workpiece. After activating the SSV mode at 5 seconds, the rotational speed begins to exhibit sinusoidal modulation, and the vibration energy trend and real-time vibration signal rapidly decay and remain within a low, safe range. This result confirms that the SSV strategy effectively suppresses the formation of fish-scale defects by breaking the phase lag effect caused by constant rotational speed, thus improving the surface roughness of the workpiece. Optimized from 1.2μm to 0.4μm.

Claims

1. A milling control system for the rear end cover of a new energy motor, characterized in that, include: The positioning detection module (2) is used to monitor the physical clamping status of the workpiece in real time and generate a clamping position signal transmission. The environmental pretreatment module (3) is used to perform a cleaning action on the surface of the workpiece to be tested to remove cutting fluid residue and metal chips, and to activate the protective device of the optical sensor. The milling execution module (6) is used to receive machining instructions and adjust the spindle speed, feed rate and tool path based on the machining instructions to perform cutting operations; Data storage module (7) is used to preset machining accuracy standards, semi-finishing allowance parameters and tool compensation algorithm models; The central control module (1) is used to receive the clamping signal and control the milling execution module (6) to complete the semi-finishing, trigger the environmental preprocessing module (3) to perform the cleaning action, and control the vision inspection module (4) and the laser inspection module (5) to obtain the actual geometric data of the workpiece semi-finished product. Based on the comparison result between the actual geometric data and the preset standard in the data storage module (7), the tool compensation parameters and the corrected tool path are generated to control the milling execution module (6) to complete the finishing.

2. The milling control system for the rear end cover of a new energy motor according to claim 1, characterized in that, The central control module (1) is used to execute the following process control logic: Upon receiving the clamping signal, the milling execution module (6) is controlled to perform roughing and semi-finishing operations, and during the semi-finishing stage, the workpiece is controlled to retain a preset single-sided safety margin. After the semi-finishing process is completed, the environmental pretreatment module (3) is started to purge the key test area of ​​the workpiece with high-pressure airflow. After the high-pressure airflow is cleaned, the visual inspection module (4) and the laser inspection module (5) are controlled to perform in-situ inspection on the workpiece to obtain the actual geometric data of the current workpiece semi-finished product. The central control module (1) compares the collected actual geometric data with the final design standard in the data storage module (7), calculates the actual remaining machining amount of each part of the workpiece based on the comparison result, and generates the tool compensation parameters and machining instructions for the finishing process based on the actual remaining machining amount, and transmits them to the milling execution module (6) to drive it to adjust the tool path. The generated tool compensation parameters and the corrected tool path are sent as finishing instructions to the milling execution module (6), which controls the milling execution module (6) to drive the spindle and feed axis to perform the finishing process according to the finishing instructions, thereby completing the final dimension correction of the workpiece.

3. The milling control system for the rear end cover of a new energy motor according to claim 1, characterized in that, The positioning detection module (2) includes a pneumatic pressure sensing nozzle, a pneumatic pipeline, a temperature sensor and a digital pressure sensor unit integrated in the positioning support surface of the fixture; The digital pressure sensor unit is used to collect the static pressure value inside the gas pipeline; When the positioning detection module (2) detects that the static pressure value reaches the preset pressure threshold and the pressure fluctuation rate is lower than the preset stability threshold, it generates the clamping signal.

4. The milling control system for the rear end cover of a new energy motor according to claim 1, characterized in that, The environmental pretreatment module (3) includes a high-pressure pneumatic purging assembly and a sensor protection assembly; The high-pressure pneumatic purging assembly includes a directional air knife fixed to the side of the spindle, and the air outlet angle of the directional air knife forms a preset angle with the surface of the workpiece to be tested; The sensor protection assembly includes a sealed protective housing and a fast-opening and closing shutter. The visual detection module (4) and the laser detection module (5) are encapsulated in the sealed protective housing. The fast-opening and closing shutter is located at the light-transmitting window of the sealed protective housing.

5. The milling control system for the rear end cover of a new energy motor according to claim 1, characterized in that, The visual inspection module (4) includes an industrial area array camera, a dual telecentric optical lens, and a composite illumination unit; The central control module (1) uses the image data collected by the vision inspection module (4) to identify the defects on the surface of the workpiece to be tested, and extracts the center pixel coordinates of the main reference hole of the workpiece in the image coordinate system, and converts the center pixel coordinates into physical coordinates in the machine tool machining coordinate system.

6. The milling control system for the rear end cover of a new energy motor according to claim 1, characterized in that, The laser detection module (5) includes a laser displacement sensor and a coordinate adjustment submodule; The coordinate adjustment submodule includes a servo-driven telescopic slide, which is used to drive the laser displacement sensor to switch between the avoidance position and the measurement position. The central control module (1) converts the local distance values ​​collected by the laser displacement sensor into absolute spatial coordinates in the machine tool machining coordinate system based on coordinate transformation calculation.

7. The milling control system for the rear end cover of a new energy motor according to claim 3, characterized in that, The data storage module (7) stores thermal error calculation model, material linear expansion coefficient, and three-dimensional CAD model data of the workpiece; The central control module (1) receives the real-time temperature data of the workpiece transmitted by the temperature sensor, uses the thermal error calculation model to restore the measured size collected by the laser detection module (5) to the equivalent size at the standard reference temperature, and calculates the scalar error of the size based on the equivalent size.

8. The milling control system for the rear end cover of a new energy motor according to claim 6, characterized in that, The data storage module (7) also stores a historical processing error trend table; The central control module (1) performs a weighted moving average calculation on the historical workpiece size deviation sequence to obtain a systematic drift trend, and then adds the systematic drift trend to the tool compensation parameters.

9. The milling control system for the rear end cover of a new energy motor according to claim 1, characterized in that, The milling execution module (6) collects the load current of the spindle servo motor in real time and calculates the variance of the load current; When the load current variance exceeds the preset stability threshold, the milling execution module (6) triggers the spindle speed variation strategy to control the spindle speed to fluctuate periodically on the basis of the nominal speed in a sinusoidal pattern.

10. The milling control system for the rear end cover of a new energy motor according to claim 1, characterized in that, The milling execution module (6) is embedded with a load adaptive control algorithm; The milling execution module (6) estimates the instantaneous cutting load rate through the load observer and compares the instantaneous cutting load rate with the preset optimal spindle load rate, and dynamically adjusts the feed rate according to the comparison result.

Images