Multi-task cooperative machining assembly of turn-milling hybrid numerical control machine tool

Abstract

The application discloses a turning-milling combined numerical control machine tool multi-task cooperative machining assembly and belongs to the technical field of turning-milling combined numerical control machine tools, which comprises a turning-milling combined machine tool table and a controller box, and further comprises the following placed in the controller box: a digital twin real-time monitoring module; a reinforcement learning scheduling engine, based on the deviation of the predicted state of the digital twin and the actual sensor data, the reinforcement learning algorithm is used to dynamically adjust the multi-channel task execution sequence, when the deviation of the predicted state of the digital twin and the actual sensor data exceeds a preset threshold, the reinforcement learning scheduling engine automatically triggers a task rearrangement algorithm, generates a new multi-channel cooperative execution sequence, and seamlessly switches to the new sequence execution. The application has the ability to autonomously cope with tool wear, material abnormalities or sudden vibration interference during the machining process, significantly improves the robustness and adaptability of the machining process, and effectively avoids downtime, waste products or equipment damage caused by the accumulation of small deviations.

Description

Technical Field

[0001] This invention belongs to the field of turning-milling composite CNC machine tool technology, and in particular to multi-task collaborative machining components for turning-milling composite CNC machine tools. Background Technology

[0002] A mill-turning composite CNC machine tool is a high-end manufacturing equipment that integrates multiple machining processes such as turning, milling, drilling, and tapping. Its core advantage lies in completing all or most of the machining operations on a workpiece in a single setup, effectively avoiding datum conversion errors caused by multiple setups and significantly improving the machining accuracy and efficiency of complex precision parts. These machine tools are typically equipped with multi-axis linkage systems (such as X / Y / Z / C / B axes), powered turrets, sub-spindles, and dual tool posts, enabling composite operations such as external turning, end milling, and eccentric feature machining on rotating parts. They are widely used in aerospace, medical devices, precision instruments, and other fields with extremely high requirements for part integration and precision.

[0003] However, in the process of fully leveraging its hardware composite capabilities, the problem of intelligent scheduling for multi-task collaborative processing has become increasingly prominent, becoming a key bottleneck restricting further improvement in its efficiency. In existing technologies, the multi-channel control of milling and turning composite machines mainly relies on preset G-code programs and fixed logical timing (e.g., simple synchronization between channels is achieved through M-codes or WAIT instructions).

[0004] This scheduling method has fundamental flaws, such as a disconnect between the scheduling logic and the physical world, and a lack of real-time status awareness and feedback mechanisms. The scheduling decisions of existing systems are entirely based on engineers' assumptions about ideal machining conditions during the offline programming phase. Once the program starts, it strictly follows the preset path and timing, becoming an "open-loop" system. The root cause of this shortcoming lies in the design philosophy of traditional CNC controllers—they are positioned as high-precision "motion command executors" rather than "intelligent decision-makers" capable of understanding the physical nature of the machining process. The controller lacks a virtual model that evolves synchronously with the physical machine tool, and it cannot access and process real-time data streams from multiple sensors, including those for force, vibration, and acoustic emissions. Failure to address this issue will have serious consequences: any minute, unforeseen dynamic disturbance will lead to loss of control during machining. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-task collaborative machining component for milling-turning composite CNC machine tools to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a multi-task collaborative machining component for a milling-turning composite CNC machine tool, comprising a milling-turning composite machine tool table and a controller box, and further comprising: The digital twin real-time monitoring module is used to construct digital twins of each machining unit of the milling and turning composite CNC machine tool, and to collect physical parameters such as spindle speed, tool position, cutting force, and vibration acceleration in real time, and generate a real-time mapping between the machine tool entity and the digital twin. The reinforcement learning scheduling engine dynamically adjusts the multi-channel task execution sequence based on the deviation between the predicted state of the digital twin and the actual sensor data using a reinforcement learning algorithm. The input of the reinforcement learning algorithm includes the task queue state of each channel, tool wear, spindle load rate, and machining accuracy deviation, and the output includes task priority adjustment, channel switching, and tool path compensation. When the deviation between the predicted state of the digital twin and the actual sensor data exceeds a preset threshold, the reinforcement learning scheduling engine automatically triggers a task reordering algorithm to generate a new multi-channel collaborative execution sequence and seamlessly switches to the execution of the new sequence, forming a real-time state feedback-task rescheduling closed-loop control process.

[0007] This preferred option also includes: The real-time data synchronization module enables millisecond-level synchronous acquisition of data from spindle speed, tool position, cutting force, and vibration acceleration sensors via the EtherCAT bus; The multi-dimensional state vector construction module standardizes the task queue status of each channel, tool wear, spindle load rate and machining accuracy deviation into feature vectors within the range of 0-1, and inputs them into the reinforcement learning scheduling engine. The reinforcement learning action space module includes task priority adjustment (pause / acceleration / delay), channel switching (main channel / sub-channel / backup channel), and toolpath compensation (X / Y / Z axis offset). The action space is optimized by the Q-learning algorithm with the objective function of minimizing the total machining time. The dynamic threshold adaptive logic automatically adjusts the vibration threshold (roughing ±0.5g, finishing ±0.1g) and load threshold (roughing ≤80% of rated value, finishing ≤60% of rated value) according to the processing stage (roughing / finishing). When the actual parameters exceed the dynamic threshold, the reinforcement learning scheduling engine is triggered to reorder tasks.

[0008] In a preferred embodiment of this scheme, the action space of the reinforcement learning scheduling engine includes: The task dependency reconstruction module dynamically adjusts the dependency between turning and milling processes based on the machining accuracy deviation predicted by the digital twin. The toolpath compensation calculation module calculates the toolpath compensation amount (ΔX, ΔY, ΔZ) based on the tool wear and spindle load rate predicted by the digital twin and outputs it to the machine tool entity for execution. The multi-channel task queue state encoding module uses a Long Short-Term Memory (LSTM) network to encode the execution time, completion rate, and priority of each channel task queue, forming the input state vector of the reinforcement learning algorithm. The reward function design module includes a processing efficiency reward (shortened task completion time), a processing quality reward (reduced accuracy deviation), and a safety constraint reward (avoiding vibration exceeding the threshold). The weights of the three are dynamically adjusted according to the processing stage (rough processing efficiency weight 0.6, quality weight 0.3, safety weight 0.1; fine processing efficiency weight 0.4, quality weight 0.5, safety weight 0.1).

[0009] In this preferred embodiment, when the processing accuracy deviation predicted by the digital twin exceeds the dynamic threshold, a risk level signal (high / medium / low) is sent to the reinforcement learning scheduling engine to trigger task reordering. At the same time, the new collaborative sequence generated by the reinforcement learning scheduling engine is synchronized to the digital twin to update the input conditions of the prediction model. The reinforcement learning scheduling engine uses the prediction deviation of the digital twin as part of the reward function (rewarding +0.5 for every 1% reduction in deviation), forming a dynamic optimization closed loop. The digital twin integrates the real-time states of both and jointly predicts the probability of multi-channel task conflicts to update the dynamic threshold. When the machining accuracy deviation exceeds the threshold, the execution order is automatically adjusted according to the task criticality weight (finishing > roughing) to ensure that high-precision processes are completed first. The weight coefficient is dynamically adjusted according to the machining stage (roughing weight = 1.0, finishing weight = 1.5).

[0010] In a preferred embodiment, a crossbeam is mounted on the milling-turning machine tool table. The inner wall of the crossbeam is provided with a three-jaw chuck for clamping workpieces. A milling cutter is provided at the free end of the crossbeam. The top surface of the crossbeam also has a moving module. A tool post is provided at the end of the moving module. When the three-jaw chuck clamps a shaft-type workpiece, the workpiece clamping mechanism located behind the milling-turning machine tool table assists in clamping, preventing tilting and shaking during processing.

[0011] In a preferred embodiment, the workpiece clamping mechanism includes an L-shaped fixture frame, a robotic arm mounted and connected to the back of the L-shaped fixture frame, and a first clamping assembly and a second clamping assembly respectively disposed on the inner walls of both sides of the L-shaped fixture frame. The first clamping assembly and the second clamping assembly are both connected to the L-shaped fixture frame via movable guide rails. An inverted U-shaped support column is mounted on the back of the milling and turning machine table. A guide rail is provided on the inner top wall of the U-shaped support column. A transverse moving block for sliding on the guide rail is provided at the end of the robotic arm away from the L-shaped fixture frame.

[0012] In a preferred embodiment, the L-shaped fixture frame includes an integrally formed upper offset frame and a lower offset frame, which are staggered. The inner walls of both the upper and lower offset frame are provided with mounting grooves for accommodating movable guide rail components. The movable guide rail components include guide rails mounted in the mounting grooves, slider seats slidably connected to the guide rails, and servo motors respectively mounted on the free ends of the upper and lower offset frame. The output shaft of the servo motor extends and is connected to a lead screw in the guide rail for driving. The slider seat is connected to the lead screw.

[0013] In a preferred embodiment of this solution, both the first clamping assembly and the second clamping assembly include a connecting housing that can be fixedly connected to an adjacent slider seat and two clamping jaws that can clamp inwards and move away from each other. The inner walls of the two adjacent clamping jaws are each fixed with a first clamping grinding block. The outer wall of the connecting housing near the clamping jaws is fixed with a second clamping grinding block. The side of the second clamping grinding block and the two first clamping grinding blocks that are in contact with the outer wall of the shaft workpiece are each provided with a grinding end face.

[0014] In this preferred embodiment, the tail ends of both grippers are hinged to a first connecting arm via pins, the ends of the two first connecting arms away from the grippers are hinged to a second connecting arm via pins, the ends of the two second connecting arms away from the first connecting arms are hinged to a connecting housing via pins, the outer walls of both sides of the connecting housing are symmetrically hinged to two driving arms via pins, and the ends of the four driving arms away from the connecting housing are respectively hinged to the ends of the adjacent grippers near the first connecting arm.

[0015] In a preferred embodiment, the drive arm is hinged to the periphery of two pins in the connecting housing and is located in the cavity of the connecting housing with an interference fit of drive gears that mesh with each other. A fixed seat is fixed to the outer wall of the connecting housing, and a drive motor is mounted on the fixed seat. The output shaft of the drive motor passes through the fixed seat and is fixedly connected to one of the pins of the drive arm. When the drive motor drives the pin shaft clockwise, it engages two drive gears, causing the two drive arms to separate and unfold in opposite directions. This pulls the jaws backward, disengaging them from the grinding end faces and preventing the second grinding block from contacting the shaft workpiece, thus opening the clamping of the shaft workpiece. When the drive motor drives the pin shaft counterclockwise, it engages two drive gears, causing the two drive arms to move closer together. This pushes the jaws forward, causing the jaws to clamp the first grinding block onto the outer circumference of the shaft workpiece. Simultaneously, the second grinding block abuts against the outer wall of the shaft workpiece. The three grinding end faces of the two first grinding blocks and one second grinding block enclose the outer wall of the shaft workpiece, allowing the grinding end faces to further grind the milled and turned outer surface of the shaft workpiece, removing burrs, when the three-jaw chuck rotates the shaft workpiece.

[0016] Compared with the prior art, the technical effects and advantages of the present invention are as follows: This multi-task collaborative machining component for milling and turning composite CNC machine tools constructs a closed-loop collaborative architecture of a "digital twin real-time monitoring module" and a "reinforcement learning scheduling engine." This enables the system to compare the real-time state of the physical machine tool with the predicted state of the virtual model at the millisecond level, achieving dynamic task rescheduling based on actual deviations rather than preset programs. Compared with existing technologies that rely on static G-code and fixed timing for multi-channel control, this component achieves the ability to autonomously cope with interferences such as tool wear, material abnormalities, or sudden vibrations during machining. This significantly improves the robustness and adaptability of the machining process and effectively avoids downtime, scrap, or equipment damage caused by the accumulation of small deviations.

[0017] By designing a data processing mechanism of "EtherCAT bus millisecond-level synchronous acquisition" and "multi-dimensional state vector Z-score standardization", and combining it with "adaptive dynamic threshold logic for the processing stage", the input features of the reinforcement learning algorithm have a unified dimension and process relevance. Different physical quantities are standardized into comparable vectors in the range of 0-1. At the same time, different safety thresholds are used in the roughing and finishing stages, making the scheduling decision more in line with the actual process requirements. Compared with the existing technology that uses a single fixed threshold or unprocessed raw data as the basis for judgment, it achieves more accurate task reordering trigger conditions and a lower false alarm rate, ensuring that the scheduling action is activated only when it is really necessary, thus balancing processing efficiency and system stability. Attached Figure Description

[0018] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the structure of the present invention; Figure 2 This is a schematic diagram of the U-shaped support column of the present invention; Figure 3 This is a schematic diagram of the inverted mounting structure of the robotic arm of the present invention; Figure 4 This is a schematic diagram of the connection structure between the robotic arm and the L-shaped fixture frame plate of the present invention; Figure 5 This is a schematic diagram of the connection structure between the upper biasing frame plate and the lower biasing frame plate of the present invention; Figure 6 This is a schematic diagram of the connection structure of the drive arm of the present invention; Figure 7 This is a schematic diagram of the installation structure of the drive motor of the present invention; Figure 8 This is a schematic diagram of the structure of the two drive gears in meshing state according to the present invention; Figure 9 This is a control architecture diagram of the multi-task collaborative machining component system for milling and turning composite CNC machine tools according to the present invention; Figure 10 This is a hardware composition diagram of the digital Li Sheng real-time monitoring module of the present invention; Figure 11 This is a flowchart of the reinforcement learning scheduling engine algorithm of the present invention; Figure 12 This is a diagram illustrating the architecture of the EtherCAT real-time data synchronization mechanism of the present invention. Figure 13 This is a flowchart illustrating the closed-loop collaborative control of the multi-task collaborative machining component of the milling-turning composite CNC machine tool according to the present invention.

[0020] Explanation of reference numerals in the attached figures: In the diagram: 1. Turning-milling composite machine table; 2. Controller box; 3. Crossbeam frame; 4. Three-jaw chuck; 5. Transverse moving guide rail; 6. Transverse moving table; 7. Longitudinal moving guide rail; 8. Longitudinal moving table; 9. Tool moving guide rail; 10. Tool post; 11. Milling cutter; 12. U-shaped support column; 13. Robotic arm; 14. Workpiece clamping mechanism; 15. L-shaped fixture frame; 16. Transverse moving block; 17. Upper offset frame; 18. Lower offset frame; 19. Wrist end. 20. First clamping assembly; 21. Second clamping assembly; 22. Moving guide rail; 23. Mounting groove; 24. Guide rail; 25. Slider seat; 26. Connecting seat housing; 27. Gripper; 28. Servo motor; 29. ​​First clamping grinding block; 30. Grinding end face; 31. Second clamping grinding block; 32. Drive arm; 33. First connecting arm; 34. Second connecting arm; 35. Pin; 36. Drive motor; 37. Drive gear; 38. Fixed seat. Detailed Implementation

[0021] In the following description, numerous specific details are set forth in order to provide a more thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention can be practiced without one or more of these details. In other instances, certain technical features well-known in the art have not been described in order to avoid obscuring the invention.

[0022] Unless otherwise defined, the directions mentioned herein, such as up, down, left, right, front, back, inside, and outside, are based on the directions shown in the figures of this invention, and are explained here together.

[0023] This embodiment provides, for example Figures 1 to 13 The multi-task collaborative machining component of the milling and turning CNC machine tool shown has the following overall architecture: Figure 9 As shown, it mainly includes four core components: a digital twin real-time monitoring module, a reinforcement learning scheduling engine, a real-time data synchronization mechanism, and a closed-loop collaborative control process. The digital twin real-time monitoring module is responsible for collecting the physical parameters of the machine tool and constructing a digital twin; the reinforcement learning scheduling engine dynamically adjusts the multi-channel task execution sequence based on the deviation between the predicted state of the digital twin and the actual sensor data; the real-time data synchronization mechanism achieves millisecond-level data interaction through the EtherCAT bus; and the closed-loop collaborative control process forms a complete control loop of state monitoring → deviation calculation → task rescheduling → execution feedback.

[0024] The digital twin real-time monitoring module and the reinforcement learning scheduling engine work together through bidirectional data flow. When the deviation between the digital twin's predicted state and the actual sensor data exceeds a preset threshold, the reinforcement learning scheduling engine automatically triggers the task reordering algorithm to generate a new multi-channel collaborative execution sequence and seamlessly switches to the new sequence for execution. At the same time, the new collaborative sequence is synchronized to the digital twin to update the input conditions of the prediction model, forming a real-time state feedback-task rescheduling closed-loop control process.

[0025] In this embodiment, the hardware composition of the digital twin real-time monitoring module is as follows: Figure 10 As shown, it mainly consists of three parts: sensor network, data acquisition circuit, and communication interface. I. Sensor Networks: (1) Spindle speed sensor: installed on the spindle motor shaft end, using photoelectric encoder, model E6B2C-M2100U (17-bit resolution), sampling frequency 100Hz. (2) Six-dimensional force sensor: integrated in the turret interface, model HPS-FT060E (EtherCAT interface, 24-bit ADC, 2000Hz sampling), can simultaneously measure force (Fx, Fy, Fz) in three directions and torque (Mx, My, Mz) in three directions. (3) Vibration sensor: installed on the spindle base, model DH5922N matching three-axis accelerometer, sampling frequency 500Hz. (4) Temperature sensor: arranged in the spindle bearing and tool clamping area, model DS18B20 digital temperature sensor, sampling frequency 10Hz.

[0026] II. Data Acquisition Circuit: (1) ESC (EtherCAT Slave Controller): AX58100 chip is used, integrating two Fast Ethernet PHYs that support 100Mbps full-duplex operation and HPAuto-MDIX function; (2) Microcontroller: STM32F407 chip is used, responsible for processing sensor data and communicating with ESC; (3) Analog Input Circuit: ADG5207 analog switch is used to select the signal source, LTC6362 buffer operational amplifier, LTC2380-24 analog-to-digital converter module, and LTC6655-4.096 voltage regulator chip to provide reference voltage; (4) Digital Input / Output Circuit: 6N137 high-speed optocoupler isolation chip is used to ensure the reliability of signal transmission.

[0027] III. Communication Interface: EtherCAT Interface: Supports high-speed communication between master and slave stations, with a data transmission speed of up to 100Mbps, a data refresh cycle of less than 100μs, and a synchronization accuracy of less than 1μs.

[0028] In this embodiment, the software implementation of the digital twin real-time monitoring module mainly includes three parts: multi-domain modeling, data acquisition and processing, and virtual-real interaction. I. Multi-domain modeling: (1) A multi-domain coupled model of CNC machine tool is constructed using the Modelica multi-domain unified modeling language, including three subsystem models: mechanical, electrical and control. (2) The mechanical subsystem model includes the kinematic and dynamic parameters of components such as spindle, turret and guide rail. (3) The electrical subsystem model includes the electromagnetic characteristics of components such as motor, driver and power supply. (4) The control subsystem model includes NC program, control algorithm and signal connection.

[0029] II. Data Acquisition and Processing: (1) Sensor bone data is periodically acquired via EtherCAT bus at a sampling frequency of 1ms. (2) Multidimensional state vectors are processed using the Z-score normalization formula: , where Z represents the normalized value; X represents the original data; μ and σ represent the mean and standard deviation of the original data, respectively. (3) The box plot algorithm is used to remove abnormal data, such as vibration mutation values, to ensure the reliability of the model input data.

[0030] III. Virtual-Real Interaction: (1) Real-time synchronization of sensor data to the digital twin to update model parameters. (2) Automatic adjustment of dynamic thresholds according to the processing stage (roughing / finishing): (3) Roughing stage: vibration threshold ±0.5 g, spindle load rate between 80%. (4) Finishing stage: vibration threshold ±0.1 g, spindle load rate between 60%.

[0031] In this embodiment, the reinforcement learning scheduling engine is implemented using the PPO (Proximal Policy Optimization) algorithm, and the algorithm flow is as follows: Figure 11 As shown: State space definition: state vector Where: c1(t), c2(t), and c3(t) represent the operating status (execution time, completion rate, load rate, etc.) of the lathe spindle, milling spindle, and turret, respectively; O1(t), O2(t), and O3(t) represent the external resource status of spindle load rate, tool wear, and vibration acceleration, respectively.

[0032] Action space definition: Action space A includes the following composite actions: (1) Task priority adjustment (pause / accelerate / delay): Adjust the order of the task queue by weighting coefficients, for example: (2) Channel switching (main channel / secondary channel / standby channel): When the spindle load rate exceeds the threshold, the channel is automatically switched. (3) Toolpath compensation (X / Y / Z axis offset): The compensation amount is calculated based on the Floyd algorithm. ; ; .

[0033] Reward function design: The reward function R includes three objectives: ; in: (Time-saving reward for tasks); (Reward for reduced accuracy deviation); (Safety constraint rewards); Dynamic weight adjustment: The weighting coefficients are dynamically adjusted according to the processing stage. ; ; ; Policy network parameters: learning rate α = 0.001; discount factor y = 0.99; policy network parameters are updated every 500 decisions.

[0034] The implementation details of the reinforcement learning scheduling engine include: State vector construction: Sensor data, including parameters such as spindle speed, tool position, cutting force, vibration acceleration and temperature, are acquired from the EtherCAT bus, and the multidimensional state vector is mapped to the 0-1 range using the Z-score normalization formula.

[0035] Action selection: Policy network input state vector S t Output action probability distribution The optimal action is selected using the Softmax activation function, for example, pausing the milling task and adjusting the toolpath compensation when vibration exceeds a threshold.

[0036] Reward Calculation: The reward value is calculated based on task completion time, machining accuracy, and safety constraints. For example, when vibration exceeds a threshold, the safety constraint reward r... maa If the value is negative, the action is penalized.

[0037] Strategy optimization: The PPO algorithm is used to update the policy network parameters to maximize long-term rewards. The policy network is updated every 500 decisions to ensure real-time performance.

[0038] In this embodiment, the real-time data synchronization mechanism is implemented using the EtherCAT bus, with the architecture as follows: Figure 12 As shown, I. Data Acquisition and Transmission: (1) TwinCAT software is used as the master station, located on the PC. The master station is responsible for importing and parsing XML files, creating slave devices, constructing initialization command frames, and performing communication initialization; (2) Each machining unit (spindle, turret, milling head) is an EtherCAT slave station. The slave station controller (ESC) uses the LAN9252 chip and interacts with the microcontroller through a parallel bus. The slave station software includes EtherCAT slave communication interface software, process data communication, and mailbox communication. (3) Data Synchronization Process: (a) The master station sends downlink data frames (output control data frames) to the slave station.

[0039] (b) The slave station ESC identifies whether there is a data unit corresponding to itself based on the logical address. If it exists, it reads the output data set by the master station into the internal register of the ESC.

[0040] (c) The slave station master controller obtains the output data set by the master station from the ESC register and controls the output module to execute the output.

[0041] (d) The slave station transmits the downlink data frame to the other slave stations connected in series, until the last slave station.

[0042] (e) Data loopback: The slave station puts the data obtained by the measurement module into the specified area of ​​the ESC internal register, and the ESC inserts the input data in the specified area of ​​the internal register into the corresponding position of the uplink data frame.

[0043] (f) The uplink data frame is returned to the master station, and after processing, it is displayed on the host computer human-machine interface.

[0044] II. Data Processing and Analysis (1) Data preprocessing: The multidimensional state vector was processed using the Z-score normalization formula. Outlier data, such as vibration mutation values, were removed using the box plot algorithm.

[0045] (2) State Vector Construction: Parameters such as spindle load rate, tool wear, vibration acceleration, and machining accuracy deviation are standardized into feature vectors within the range of 0-1. For example, the standardized formula for spindle load rate is: Where μ1 and σ1 are the historical load rate mean and standard deviation, respectively.

[0046] (3) Dynamic threshold adjustment: The vibration threshold and load threshold are automatically adjusted according to the processing stage. For example, the vibration threshold is ±0.5 g in the roughing stage and ±0.1 g in the finishing stage.

[0047] In this embodiment, the closed-loop collaborative control process is as follows: Figure 13 As shown, the main steps include: 1. Condition Monitoring and Data Acquisition: Real-time acquisition of machine tool physical parameters, including spindle speed, tool position, cutting force, vibration acceleration, and temperature, is achieved via a sensor network. Millisecond-level data synchronization is achieved using an EtherCAT bus, with a sampling frequency of 1ms.

[0048] 2. State Vector Construction and Standardization: The acquired sensor data is standardized into feature vectors within the range of 0-1. For example, the standardization formula for spindle load rate is: ; Automatically adjust dynamic thresholds based on processing stage: Rough machining stage: vibration threshold ±0.5g, spindle load rate threshold 80%; Finishing stage: Vibration threshold ±0.1g, spindle load rate between 60%.

[0049] 3. Deviation Calculation and Trigger Condition Judgment: Calculate the deviation between the digital twin's predicted state and the actual sensor data. When the deviation exceeds a preset threshold, the reinforcement learning scheduling engine is triggered to reorder tasks. For example, when the vibration deviation exceeds ±0.5g (rough machining stage) or ±0.1g (finishing stage), task reordering is triggered.

[0050] 4. The reinforcement learning scheduling engine dynamically adjusts the execution sequence of multi-channel tasks based on state vectors and biases using the PPO algorithm. The action space includes task priority adjustment, channel switching, and toolpath compensation. For example, when vibration exceeds a threshold, the milling task is paused and the toolpath compensation amount ΔZ = 0.02 mm is adjusted.

[0051] 5. Synchronize the scheduling results to the machine tool for execution, monitor new states during execution in real time, and form a complete closed-loop control process. For example, after performing toolpath compensation, the vibration is reduced to 0.45g (roughing stage), meeting safety constraints.

[0052] Example 1

[0053] Taking the machining of aircraft connectors by milling and turning as an example: an aircraft connector is a precision part with two mutually perpendicular long and short shafts, and the material is stainless steel.

[0054] The part has a diameter of Φ5 mm, a tolerance of 0.005 mm, a positional tolerance of 0.04 mm, and a coaxiality requirement of 0.02 mm.

[0055] Traditional machining method: It requires machining on a lathe and a milling machine respectively, with a total machining time of about 120 minutes.

[0056] Milling and turning combined machining method: Using this technology, the total machining time is reduced to 45 minutes.

[0057] The process flow includes: (1) In the machining mode, rough machining of the shape is performed. To ensure sufficient rigidity when finishing the right end, the minimum diameter of the left end is Φ10mm.

[0058] (2) Call the rear spindle to clamp the right end, keep the front and rear spindles synchronized, rotate the spindle, and rough machine the part shape in milling mode.

[0059] (3) After releasing the spindle and returning it to its original position, perform semi-finishing and finishing of the right end dimension in machining mode.

[0060] (4) Call the rear spindle to clamp the right end, keep the front and rear spindles synchronized, rotate the spindle, first use the back boring tool to finish the two Φ5mm short shafts in milling mode, and then semi-finish and finish all dimensions of the left end in turning mode.

[0061] Dynamic scheduling process: (1) During the turning stage, the spindle load rate reached 85% (exceeding the roughing threshold of 80%), and the vibration reached 0.55g (exceeding the roughing threshold of 0.5g).

[0062] (2) The vibration deviation predicted by the digital twin is 0.6g, while the actual sensor data is 0.55g.

[0063] (3) Trigger the reinforcement learning scheduling engine, select to pause the milling task and adjust the tool path compensation amount ΔZ=0.02mm.

[0064] (4) The vibration is reduced to 0.45g, which meets the safety constraints. The dynamic adjustment of task priority enables the finishing process to be executed in advance, reducing the risk of vibration exceeding the threshold.

[0065] Comparative example: Traditional process: Total processing time 120 minutes, vibration exceeding threshold 15 times.

[0066] This technical solution has a total processing time of 45 minutes, with only 1 instance of vibration exceeding the threshold, resulting in a 62.5% increase in efficiency and a 90% improvement in accuracy.

[0067] Example 2

[0068] In this embodiment, a crossbeam 3 is mounted on the milling-turning machine table 1. A three-jaw chuck 4 for clamping workpieces is installed on the inner wall of the crossbeam 3. A milling cutter 11 is installed at the free end of the crossbeam 3. The top surface of the crossbeam 3 also has a moving module, with a tool holder 10 installed at the end of the moving module. When the three-jaw chuck 4 clamps a shaft-like workpiece, a workpiece clamping mechanism 14 located behind the milling-turning machine table 1 assists in clamping and preventing tilting and shaking during machining. When the three-jaw chuck 4 clamps a shaft-like workpiece for high-speed rotational machining, the workpiece clamping mechanism 14 provides auxiliary support from the rear or middle of the workpiece. This auxiliary support effectively suppresses chatter and wobble during machining by counteracting cantilever vibration and bending deformation caused by cutting forces. Compared with existing technologies that rely solely on single-point clamping of the main chuck, which can easily lead to instability during the machining of slender shaft-type workpieces, this solution achieves a high-rigidity multi-point collaborative clamping function, significantly improving the dimensional accuracy and surface quality of milling and turning composite machining. It is especially suitable for shaft-type parts with a large length-to-diameter ratio, achieving a comprehensive effect of anti-tilting, anti-shaking, and vibration reduction.

[0069] In this embodiment, the workpiece clamping mechanism 14 includes an L-shaped fixture frame 15, a robotic arm 13 mounted and connected to the back of the L-shaped fixture frame 15, and a first clamping assembly 20 and a second clamping assembly 21 respectively disposed on the inner walls of both sides of the L-shaped fixture frame 15. Both the first clamping assembly 20 and the second clamping assembly 21 are connected to the L-shaped fixture frame 15 via movable guide rails 22. An inverted U-shaped support column 12 is mounted behind the turning-milling composite machine tool table 1. A guide rail is provided on the inner top wall of the U-shaped support column 12. A lateral moving block 16 for sliding on the guide rail is provided at the end of the robotic arm 13 away from the L-shaped fixture frame 15. The clamping mechanism not only has clamping capability but also overall lateral movement freedom. Specifically, the robotic arm 13 slides along the guide rail on the inner top wall of the U-shaped support column 12 via the lateral moving block 16, which can drive the entire clamping assembly to translate along the workpiece axis. This operating mechanism allows the clamping position to be dynamically adjusted according to the processing area, enabling support or grinding of different shaft segments without re-clamping the workpiece. Compared with existing fixed auxiliary support devices, this solution achieves stroke coverage of the clamping position, avoiding processing blind spots caused by fixed support points, and thus improving process adaptability, expanding the effective processing range, and reducing the number of clamping operations.

[0070] In this embodiment, the L-shaped fixture frame 15 includes an integrally formed upper offset frame 17 and a lower offset frame 18 arranged in an alternating manner. The inner walls of both the upper offset frame 17 and the lower offset frame 18 are provided with mounting grooves 23 for accommodating the movable guide rail component 22. The movable guide rail component 22 includes a guide rail 24 mounted in the mounting groove 23, a slider seat 25 slidably connected to the guide rail 24, and servo motors 28 respectively mounted on the free ends of the upper offset frame 17 and the lower offset frame 18. The output shaft of the servo motor 28 extends and is connected to a lead screw in the guide rail 24 for driving. The slider seat 25 is connected to the lead screw. The first clamping assembly 20 and the second clamping assembly 21 can independently and precisely perform fine-tuning of their positions along their respective guide rails. This operating principle is based on servo closed-loop control, realizing the staggered arrangement and independent feeding of the clamping assemblies in the vertical direction. Compared with existing clamping structures that use rigid connections or single-plane arrangements, this solution, through the design of offset + independent drive, avoids motion interference between the two clamping components and achieves adaptive fitting clamping of asymmetrical or stepped shaft workpieces, thus achieving the effects of optimized spatial layout, conflict-free movement, and high-precision clamping and positioning.

[0071] In this embodiment, both the first clamping assembly 20 and the second clamping assembly 21 include a connecting housing 26 that can be fixedly connected to an adjacent slider seat 25, and two grippers 27 that can clamp inwards and move away from each other. The inner walls of the two adjacent grippers 27 are each fixed with a first clamping grinding block 29. A second clamping grinding block 31 is fixed to the outer wall of the connecting housing 26 near the grippers 27. The second clamping grinding block 31 and the two first clamping grinding blocks 29 are each provided with a grinding end face 30 on the side that wraps around and contacts the outer wall of the shaft-like workpiece. The clamping action and surface grinding function are integrated. When the grippers 27 are closed, the three grinding end faces 30 form a three-point envelope around the outer circle of the workpiece, creating stable contact. This operating mechanism utilizes the clamping force to synchronously apply grinding pressure, causing the grinding end faces 30 to passively remove burrs during workpiece rotation through friction. Compared with existing technologies that require additional independent grinding units or rely on manual deburring, this solution achieves simultaneous processing and finishing through the design of a clamping-grinding composite structure. This not only eliminates subsequent processes but also avoids secondary positioning errors, thereby improving surface integrity, shortening the process chain, and reducing labor costs.

[0072] In this embodiment, the tail ends of the two grippers 27 are hinged to a first connecting arm 33 via a pin 35. The ends of the two first connecting arms 33 away from the grippers 27 are hinged to a second connecting arm 34 via a pin 35. The ends of the two second connecting arms 34 away from the first connecting arms 33 are hinged to the connecting housing 26 via a pin 35. The outer walls on both sides of the connecting housing 26 are symmetrically hinged to two driving arms 32 via pins 35. The ends of the four driving arms 32 away from the connecting housing 26 are respectively hinged to the ends of the adjacent grippers 27 near the first connecting arm 33.

[0073] In this embodiment, the drive arm 32 is hinged to the periphery of two pins 35 in the connecting housing 26 and is located in the cavity of the connecting housing 26 with an interference fit of drive gears 37 meshing together. A fixed seat 38 is fixed to the outer wall of the connecting housing 26, and a drive motor 36 is mounted on the fixed seat 38. The output shaft of the drive motor 36 passes through the fixed seat 38 and is fixedly connected to one of the pins 35 of the drive arm 32. When the drive motor 36 drives the pin 35 clockwise, it drives the two drive gears 37 to mesh and connect, causing the two drive arms 32 to separate and unfold in opposite directions. This causes the drive arms 32 to pull the grippers 27 backward, so that the two grippers 27 move away from the grinding end face 30 and the second clamping grinding block 31, preventing them from contacting the shaft workpiece, thus opening the clamping of the shaft workpiece. When the drive motor 36 drives the pin 35 counterclockwise, it drives the two drive gears 37 to mesh and connect, causing the two drive arms 32 to move closer together. The drive arm 32 pushes the jaws 27 forward, causing the two jaws 27 to clamp the first clamping and grinding block 29 on the outer periphery of the shaft workpiece. At the same time, the second clamping and grinding block 31 abuts against the outer wall of the shaft workpiece. The three grinding end faces 30 of the two first clamping and grinding blocks 29 and the one second clamping and grinding block 31 wrap around the outer wall of the shaft workpiece, so that when the three-jaw chuck 4 drives the shaft workpiece to rotate, the grinding end faces 30 can further grind the outer surface of the shaft workpiece after turning and milling, and remove burrs.

[0074] In this embodiment, the design of the movable guide rail 22 allows the first clamping component 20 and the second clamping component 21 to adjust their positions for clamping shaft-like workpieces under the action of the movable guide rail 22, enabling both the first clamping component 20 and the second clamping component 21 to clamp shaft-like workpieces. Through the staggered offset setting of the upper offset frame plate 17 and the lower offset frame plate 18, the first clamping component 20 and the second clamping component 21 are offset, thereby not only enabling the first clamping component 20 and the second clamping component 21 to clamp shaft-like workpieces simultaneously, but also preventing the first clamping component 20 and the second clamping component 21 from interfering with each other's movements, thus allowing for the grinding of different positions.

[0075] In this embodiment, the guide rail on the inner top wall of the U-shaped support column 12 and the lateral moving block 16 enable the robotic arm 13 to move laterally linearly along the inner top wall of the U-shaped support column 12, thereby driving the workpiece clamping mechanism 14 to move linearly. When the first clamping component 20 and the second clamping component 21 clamp the shaft-like workpiece, this linear movement of the robotic arm 13 can move laterally to grind the shaft-like workpiece. The three-jaw chuck 4 can be selected to drive the shaft-like workpiece to rotate, or it can be selected that the three-jaw chuck 4 does not need to drive the shaft-like workpiece to rotate; both methods can grind the shaft-like workpiece. The forward and reverse rotation of the drive motor 36 can precisely control the unfolding and retraction of the drive arm 32, thereby achieving reliable opening and closing of the gripper 27. In particular, the meshing transmission of the two drive gears 37 ensures that the movement of the left and right drive arms 32 is completely synchronized, eliminating the "jamming" or "asynchrony" problems that may occur with traditional single-sided drive. This operating principle, combined with the aforementioned three-point grinding structure, automatically activates the grinding function while clamping. Compared with existing technologies that separate clamping and grinding and require additional control system coordination, this solution achieves seamless switching between clamping / release / grinding states through mechatronic drive and gear synchronous transmission design. It achieves compact structure, rapid response, reliable operation and easy maintenance, and fundamentally solves the technical bottleneck of difficult online processing of surface burrs after milling and turning of slender shafts.

[0076] In this embodiment, the moving module includes a transverse moving guide rail 5 mounted on the top surface of the milling and turning machine tool table 1, a transverse moving stage 6 adapted to the top surface of the transverse moving guide rail 5, a longitudinal moving guide rail 7 mounted on the top surface of the transverse moving stage 6, a longitudinal moving stage 8 adapted to the top surface of the longitudinal moving guide rail 7, and a tool moving guide rail 9 mounted on the top surface of the longitudinal moving stage 8. The tool holder 10 is slidably connected to the top surface of the tool moving guide rail 9, and the tool holder 10 is used for mounting turning tools.

[0077] Working principle This multi-task collaborative machining component of the milling and turning composite CNC machine tool allows the operator to load one end of the shaft-type workpiece to be machined into the three-jaw chuck 4, which is initially clamped by the three-jaw chuck 4 on the crossbeam frame 3. At this time, the workpiece is in the main machining area of ​​the milling and turning composite machine tool table 1. The control system issues a command through the controller box 2 to drive the transverse moving block 16 on the inner top wall guide rail of the U-shaped support column 12 to move laterally along the guide rail, causing the robotic arm 13 and the L-shaped fixture frame plate 15 connected to it to move closer to the workpiece, so that the first clamping component 20 and the second clamping component 21 are aligned with the area of ​​the workpiece to be supported / ground.

[0078] Since the L-shaped fixture frame 15 is integrally formed by the upper offset frame 17 and the lower offset frame 18, the first clamping assembly 20 and the second clamping assembly 21 are offset in the vertical direction to avoid motion interference. The corresponding servo motors 28 are started, driving the lead screw in the guide rail 24 to rotate, which drives the slider seat 25 to slide back and forth along the guide rail 24, thereby accurately adjusting the clamping position of the first clamping assembly 20 and the second clamping assembly 21 in the axial direction to adapt to workpieces of different lengths or processing sections.

[0079] The drive motor 36 rotates counterclockwise, driving one of the pins 35 to rotate via the output shaft. This pin 35 drives the drive gear 37, which is in interference fit with it, and in turn drives another meshing drive gear 37 to rotate in the opposite direction. The two drive arms 32 move closer together due to gear transmission, pushing the jaws 27 forward. The jaws 27 rotate around the pin 35, causing the two first clamping grinding blocks 29 to close concentrically. At the same time, the second clamping grinding block 31 also abuts against the outer wall of the workpiece. The three grinding end faces 30 (two from the first clamping grinding block 29 and one from the second clamping grinding block 31) form a three-point envelope contact, stably clamping the outer cylindrical surface of the workpiece. At this time, the workpiece is not only mainly clamped by the three-jaw chuck 4, but also provided with auxiliary support in the middle or tail by the workpiece clamping mechanism 14, effectively preventing vibration, bending or chatter caused by excessive overhang during high-speed milling.

[0080] The tool holder 10 is precisely positioned to the workpiece to be turned via a moving module consisting of a transverse moving guide rail 5, a transverse moving table 6, a longitudinal moving guide rail 7, a longitudinal moving table 8, and a tool moving guide rail 9, and performs turning. The milling cutter 11 is mounted on the free end of the crossbeam 3 and performs milling on the end face or side of the workpiece. The three-jaw chuck 4 drives the workpiece to rotate at high speed to complete the turning and milling composite machining. During the rotation of the workpiece, the three grinding end faces 30 always maintain close contact with the machined surface of the workpiece. Utilizing the rotational kinetic energy of the workpiece itself, the grinding end faces 30 perform passive fine grinding on the burrs, small protrusions, or rough areas remaining after turning and milling. The grinding process does not require an additional power source, is energy-saving and efficient, and avoids secondary clamping errors.

[0081] If grinding is required on different axial sections of the workpiece, the control system can drive the transverse moving block 16 to move along the inner top wall guide rail of the U-shaped support column 12, thereby driving the entire workpiece clamping mechanism 14 to move laterally. During the translation, the clamping component can maintain the clamping state for continuous grinding or briefly release and then re-clamp in a new position. It supports static grinding (workpiece does not rotate, only the clamping mechanism moves) or dynamic grinding (workpiece rotates + clamping mechanism moves), adapting to various process requirements.

[0082] The drive motor 36 rotates clockwise, the drive arm 32 unfolds in opposite directions, pulling the gripper 27 outwards. The first gripping grinding block 29 and the second gripping grinding block 31 are released from the workpiece surface. The robotic arm 13 drives the L-shaped fixture frame 15 back to its initial position, avoiding the main machining area. The three-jaw chuck 4 is released, and the operator removes the shaft workpiece that has completed the integrated turning, milling and grinding machining.

[0083] It should be noted that, in this document, relational terms such as "one" and "two" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, the phrase "comprising an element defined as..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0084] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-task collaborative machining component for a milling-turning composite CNC machine tool, comprising a milling-turning composite machine tool table (1) and a controller box (2), characterized in that, Also included is the one located in the controller box (2): The digital twin real-time monitoring module is used to construct digital twins of each machining unit of the milling and turning composite CNC machine tool, and to collect physical parameters such as spindle speed, tool position, cutting force, and vibration acceleration in real time, and generate a real-time mapping between the machine tool entity and the digital twin. The reinforcement learning scheduling engine dynamically adjusts the multi-channel task execution sequence based on the deviation between the predicted state of the digital twin and the actual sensor data using a reinforcement learning algorithm. The input of the reinforcement learning algorithm includes the task queue state of each channel, tool wear, spindle load rate, and machining accuracy deviation, and the output includes task priority adjustment, channel switching, and tool path compensation. When the deviation between the predicted state of the digital twin and the actual sensor data exceeds a preset threshold, the reinforcement learning scheduling engine automatically triggers a task reordering algorithm to generate a new multi-channel collaborative execution sequence and seamlessly switches to the execution of the new sequence, forming a real-time state feedback-task rescheduling closed-loop control process.

2. The multi-task collaborative machining component for milling and turning composite CNC machine tools according to claim 1, characterized in that, Also includes: The real-time data synchronization module enables millisecond-level synchronous acquisition of data from spindle speed, tool position, cutting force, and vibration acceleration sensors via the EtherCAT bus; The multi-dimensional state vector construction module standardizes the task queue status of each channel, tool wear, spindle load rate and machining accuracy deviation into feature vectors within the range of 0-1, and inputs them into the reinforcement learning scheduling engine. The reinforcement learning action space module includes task priority adjustment, channel switching and toolpath compensation. The action space is optimized by the Q-learning algorithm with the objective function of minimizing the total machining time. The dynamic threshold adaptive logic automatically adjusts the vibration threshold and load threshold according to the processing stage. When the actual parameters exceed the dynamic threshold, the reinforcement learning scheduling engine is triggered to reorder tasks.

3. The multi-task collaborative machining component for milling and turning composite CNC machine tools according to claim 2, characterized in that: The action space of the reinforcement learning scheduling engine includes: The task dependency reconstruction module dynamically adjusts the dependency between turning and milling processes based on the machining accuracy deviation predicted by the digital twin. The toolpath compensation calculation module calculates the toolpath compensation amount based on the tool wear and spindle load rate predicted by the digital twin and outputs it to the machine tool entity for execution. The multi-channel task queue state encoding module uses a long short-term memory network to encode the execution time, completion rate, and priority of each channel task queue, forming the input state vector of the reinforcement learning algorithm. The reward function design module includes a processing efficiency reward, a processing quality reward, and a safety constraint reward, the weights of which are dynamically adjusted according to the processing stage.

4. The multi-task collaborative machining component for milling and turning composite CNC machine tools according to claim 3, characterized in that: When the processing accuracy deviation predicted by the digital twin exceeds the dynamic threshold, a risk level signal is sent to the reinforcement learning scheduling engine to trigger task reordering. At the same time, the new collaborative sequence generated by the reinforcement learning scheduling engine is synchronized to the digital twin to update the input conditions of the prediction model. The reinforcement learning scheduling engine uses the prediction deviation of the digital twin as part of the reward function to form a dynamic optimization closed loop. The digital twin integrates the real-time states of both and jointly predicts the probability of multi-channel task conflict to update the dynamic threshold. When the machining accuracy deviation exceeds the threshold, the execution order is automatically adjusted according to the task criticality weight to ensure that high-precision processes are completed first. The weight coefficient is dynamically adjusted according to the machining stage.

5. The multi-task collaborative machining component for milling and turning composite CNC machine tools according to claim 1, characterized in that, A crossbeam frame (3) is installed on the milling and turning machine tool table (1). A three-jaw chuck (4) for clamping workpieces is provided on the inner wall of the crossbeam frame (3). A milling cutter (11) is provided at the free end of the crossbeam frame (3). A moving module is also provided on the top surface of the crossbeam frame (3). A turning tool holder (10) is provided at the end of the moving module. When the three-jaw chuck (4) clamps a shaft-type workpiece, the workpiece clamping mechanism (14) provided behind the milling and turning machine tool table (1) assists in clamping and anti-tilting and anti-shaking processing.

6. The multi-task collaborative machining component for milling and turning composite CNC machine tools according to claim 5, characterized in that: The workpiece clamping mechanism (14) includes an L-shaped fixture frame (15), a robotic arm (13) mounted and connected to the back of the L-shaped fixture frame (15), and a first clamping assembly (20) and a second clamping assembly (21) respectively disposed on the inner walls of both sides of the L-shaped fixture frame (15). The first clamping assembly (20) and the second clamping assembly (21) are both connected to the L-shaped fixture frame (15) through a moving guide rail (22). An inverted U-shaped support column (12) is mounted on the back of the milling and turning machine tool table (1). A guide rail is provided on the inner top wall of the U-shaped support column (12). A transverse moving block (16) for sliding on the guide rail is provided at the end of the robotic arm (13) away from the L-shaped fixture frame (15).

7. The multi-task collaborative machining component for milling and turning composite CNC machine tools according to claim 6, characterized in that: The L-shaped fixture frame (15) includes an integrally formed upper offset frame (17) and a lower offset frame (18) arranged in an alternating manner. The inner walls of the upper offset frame (17) and the lower offset frame (18) are provided with mounting grooves (23) for accommodating the movable guide rail (22). The movable guide rail (22) includes a guide rail (24) mounted in the mounting groove (23), a slider seat (25) slidably connected to the guide rail (24), and servo motors (28) respectively mounted on the free ends of the upper offset frame (17) and the lower offset frame (18). The output shaft of the servo motor (28) extends and is connected to the lead screw in the guide rail (24) for driving. The slider seat (25) is connected to the lead screw.

8. The multi-task collaborative machining component for milling and turning composite CNC machine tools according to claim 7, characterized in that: The first clamping assembly (20) and the second clamping assembly (21) each include a connecting housing (26) that can be fixedly connected to an adjacent slider seat (25) and two clamping jaws (27) that can clamp inward and move away from each other. The inner walls of the two adjacent clamping jaws (27) are fixed with first clamping grinding blocks (29). The outer wall of the connecting housing (26) near the clamping jaws (27) is fixed with a second clamping grinding block (31). The second clamping grinding block (31) and the two first clamping grinding blocks (29) are all provided with grinding end faces (30) on the side that wraps and contacts the outer wall of the shaft workpiece.

9. The multi-task collaborative machining component for a milling-turning composite CNC machine tool according to claim 8, characterized in that: The tail ends of the two grippers (27) are hinged to a first connecting arm (33) by a pin (35). The ends of the two first connecting arms (33) away from the grippers (27) are hinged to a second connecting arm (34) by a pin (35). The ends of the two second connecting arms (34) away from the first connecting arms (33) are hinged to the connecting housing (26) by a pin (35). The outer walls of the two sides of the connecting housing (26) are symmetrically hinged to two driving arms (32) by pins (35). The ends of the four driving arms (32) away from the connecting housing (26) are respectively hinged to the ends of the adjacent grippers (27) near the first connecting arm (33).

10. The multi-task collaborative machining component for a milling-turning composite CNC machine tool according to claim 9, characterized in that: The drive arm (32) is hinged to the two pins (35) in the connecting housing (26) and is located in the cavity of the connecting housing (26) with an interference fit of a drive gear (37) meshing with each other. The outer wall of the connecting housing (26) is fixed with a fixed seat (38), and a drive motor (36) is installed on the fixed seat (38). The output shaft of the drive motor (36) passes through the fixed seat (38) and is fixedly connected to one of the pins (35) of the drive arm (32).

Images