Multi-tapping-station automatic tapping equipment and control system

Abstract

The invention relates to multi-tapping-station automatic tapping equipment and a control system. The multi-tapping-station automatic tapping equipment comprises a transposition base, a discharging sorting device, a switching cleaning device and at least two tapping devices. The control system comprises an intelligent interaction subsystem, the intelligent interaction subsystem comprises a model generation module, an action matching module and an instruction generation module, the tapping device can be provided with different tapping drilling tools to achieve different tapping effects, so that the equipment can adapt to machining of various parts, and through arrangement of at least two tapping stations, the machining efficiency is improved. According to the tapping device, tapping actions can be switched, the tapping device is prevented from being overheated, meanwhile, the scrap removing operation can be conducted on the tapping position in time, the to-be-machined part is driven to rotate in the scrap removing process by switching the work of the cleaning device, and it is guaranteed that the scrap removing action is smoothly completed; and meanwhile, the phenomenon that due to reuse of the rotating station and the tapping station, the rotating structure is subjected to large impact force, and displacement occurs is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of part processing, and more specifically, to an automatic tapping device with multiple tapping stations. Background Art

[0002] With the continuous improvement of the automation level in the manufacturing industry, the requirements for the efficiency and quality of precision tapping equipment in the field of metal part processing are increasing day by day. Traditional tapping equipment mostly adopts a single-station operation mode, and the tapping and chip cleaning processes are completed through a rotating clamping mechanism. Specifically, the following technical defects generally exist in the prior art:

[0003] Continuous operation at a single station results in a significant increase in the temperature of the cutting tool. Especially when processing high-hardness materials, the drill bit is prone to thermal deformation due to the lack of cooling intervals, directly affecting the tapping accuracy and the service life of the cutting tool. Although there are individual solutions that attempt to install a cooling mechanism, it increases the complexity of the equipment and is difficult to eradicate the problem of heat accumulation. And in this process, to solve the problem of heat accumulation, it is necessary to stop the machine for heat dissipation. At the same time, there is also the problem of chip cleaning, but this process requires the cooperation of multiple components. And the current solution is: in one station, tapping and chip cleaning are carried out, and tapping and chip cleaning are switched. However, in this way: the coupling degree between the chip cleaning function and the tapping station is too high. In the conventional design, the chip cleaning process depends on the rotation of the clamping mechanism. This makes the rotating parts of the clamping mechanism bear the axial impact force of tapping while also performing high-frequency rotation, accelerating the wear of the bearings and causing positioning deviation. According to statistics, the positioning accuracy of such a structure can drop by more than 0.05 mm after continuous operation for 200 hours.

[0004] On the other hand, the tapping system also has poor adaptability to multi-specification processing. A single tapping module is difficult to meet the processing requirements of different specifications of threads. When changing the cutting tool, it is necessary to stop the machine for adjustment, seriously affecting the equipment utilization rate. Although there are documents proposing a modular cutting tool group solution, the problem of positioning and calibration during the tool change process has not been solved. Summary of the Invention

[0005] In view of this, the object of the present invention is to provide a load resource data management system based on quality verification rules with intelligent switching, precise chip cleaning, and online detection functions.

[0006] To solve the above technical problems, the technical solution of the present invention is: an automatic tapping device with multiple tapping stations, including a transposition seat, a blanking and sorting device, a switching and cleaning device, and at least two tapping devices;

[0007] The transposition seat includes a station turntable and a rotation driving member, and the rotation driving member is used to drive the rotation on the station turntable;

[0008] The transposition seat is provided with a loading station, a tapping station and a unloading station, wherein the tapping station is arranged corresponding to the tapping device, and the unloading and sorting device is arranged at the unloading station;

[0009] The tapping device includes a tapping robot arm and a tapping module, wherein the tapping robot arm is used to drive the tapping module to move at the tapping station, and the tapping module is used to perform the tapping action;

[0010] The switching cleaning device is arranged between the tapping devices, and includes a reversing auxiliary part and a cleaning execution module. When the part to be processed moves between two tapping stations driven by the station turntable, the reversing auxiliary part drives the part to be processed to rotate on the station turntable, and the cleaning execution module is used to perform the cleaning action;

[0011] The transposition seat is provided with a plurality of fixed modules, and the fixed modules are used to fix the parts to be processed at the tapping station;

[0012] The material unloading and sorting device includes an image recognition module and a sorting module. The image recognition module is configured with an image recognition strategy. The image recognition strategy is used to recognize the image of the processed parts to generate quality data, and generate sorting instructions based on the quality data and send them to the sorting module. The sorting module drives the processed parts to move to the corresponding unloading area according to the sorting instructions.

[0013] Furthermore: the transposition seat is provided with a fixed groove on the tapping station. When the part to be processed is rotated to the tapping station, it falls into the fixed groove to limit the rotation of the part to be processed. The tapping device is provided with a suction head unit. When the suction head unit is working, it sucks the part to be processed to make it leave the fixed groove.

[0014] Furthermore: the image recognition module includes an image acquisition unit and a focus adjustment unit, the image acquisition unit is arranged on the focus adjustment unit, and the focus adjustment unit is used to adjust the distance between the image acquisition unit and the processed part.

[0015] There is also provided a multi-tapping station automatic tapping control system configured in the above-mentioned multi-tapping station automatic tapping equipment, which further includes an intelligent interaction subsystem. The intelligent interaction subsystem includes a model generation module, an action matching module, and an instruction generation module. The model generation module is used to generate a part simulation model according to the part information to be processed input by the user, and process the part simulation model through the processing target information to obtain the processing difference model. The action matching module is configured with a tapping configuration parameter library, and the tapping configuration parameter library stores a number of tapping constraint features, and the tapping constraint features are indexed by tapping device parameters. The action matching module obtains the corresponding tapping constraint features by substituting the tapping device parameters of different tapping devices, and generates a tapping form constraint according to the tapping constraint features, and matches different tapping form constraints with the processing difference model respectively to generate a tapping trajectory sequence. The tapping trajectory sequence includes a number of tapping sub-trajectories. The instruction generation module is configured with a static cost algorithm, and the static cost algorithm is used to calculate the static cost of each tapping sub-trajectory, and generate a tapping trajectory set according to the adjacent tapping sub-trajectories with the total static cost as a constraint, and generate corresponding tapping action instructions according to the tapping trajectory set to control the corresponding tapping device to work.

[0016] Further: The intelligent interaction subsystem further includes a cleaning association module. The cleaning association module includes a cleaning instruction set and an action association table. The cleaning instruction set includes a number of cleaning instructions, and the cleaning instructions are indexed by action association marks. The action association table stores a number of action association marks, and the action association marks are indexed by associated action number information. The associated action number information is the previous action number and the subsequent action number. The cleaning association module obtains the corresponding previous action number and subsequent action number according to the adjacent tapping action instructions to retrieve the cleaning instructions, and the cleaning instructions are used to control the cleaning execution module to work.

[0017] Further: There is also included a dynamic detection module. The dynamic detection module includes a device detection unit and a workpiece detection unit. The device detection unit is used to detect the state feedback data when the tapping device is working, and the workpiece detection unit is used to detect the dynamic feedback data when the part to be processed is being processed. The intelligent interaction subsystem further includes a dynamic correction module. The dynamic correction module is configured with dynamic correction conditions. When the dynamic correction conditions are triggered, the dynamic correction module generates a dynamic correction instruction according to the state feedback data and the dynamic feedback data, and divides the target tapping sub-trajectory from the current tapping trajectory set to the adjacent target tapping trajectory set according to the dynamic correction instruction.

[0018] Further: The intelligent interaction subsystem further includes a macro correction module. The macro correction module is configured with a quality matching strategy. The quality matching strategy generates macro correction parameters according to quality data, and corrects the static cost algorithm according to the macro correction parameters.

[0019] Further: The dynamic correction conditions include device temperature sub-conditions, part temperature sub-conditions, and part chip quantity sub-conditions.

[0020] Further: The blanking and sorting device is configured with an abnormal feature database, which stores a number of quality abnormal items, and each quality abnormal item is indexed by a quality graphic feature; the image recognition strategy generates a shooting instruction according to the processing difference model, and controls the image recognition module to shoot a part image according to the shooting instruction, and matches the quality graphic feature with the part image to obtain the corresponding quality difference item, and forms the quality data according to the quality difference item.

[0021] Further: The macro correction module is configured with a macro correction database, and the macro correction database is configured with a number of abnormal association sub-items and corresponding abnormal association values. For the abnormal association sub-items and abnormal association values, the quality matching strategy includes calculating the total abnormal association value under each abnormal association item, and the total abnormal association value is the weighted result of the abnormal association values corresponding to the abnormal association sub-items included in the abnormal association item. When the total abnormal association value is greater than a preset abnormal trigger value, the macro correction parameter is generated according to the total abnormal association value.

[0022] The technical effects of the present invention are mainly reflected in the following aspects: By setting like this, the tapping device can be set with different tapping drills to achieve different tapping effects, so as to enable the equipment to adapt to the processing of various parts. Moreover, through the setting of at least two tapping stations, it is possible to switch between tapping operations, avoid overheating of the tapping device, and at the same time can timely perform chip cleaning operations on the tapping positions. By switching the operation of the cleaning device, during the chip cleaning process, the part to be processed is driven to rotate, ensuring the smooth completion of the chip cleaning operation, and at the same time avoiding the phenomenon that the rotation structure is subjected to a large impact force due to the reuse of the rotation station and the tapping station and thus being displaced. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 : Axonometric structure schematic of a multi-tapping station automatic tapping device of the present invention Figure 1 ;

[0024] Figure 2 : Axonometric structure schematic of a multi-tapping station automatic tapping device of the present invention Figure 2 ;

[0025] Figure 3 : Side view of a multi-tapping station automatic tapping device of the present invention;

[0026] Figure 4 : Top view of a multi-tapping station automatic tapping device of the present invention;

[0027] Figure 5:Schematic diagram of the automatic tapping control system architecture with multiple tapping stations of the present invention.

[0028] Reference numerals: 100, transfer seat; 110, station turntable; 120, rotation driving member; 130, fixing module; 131, workpiece detection unit; 200, blanking and sorting device; 210, image recognition module; 211, image acquisition unit; 212, focal length adjustment unit; 220, sorting module; 300, tapping device; 310, tapping robotic arm; 320, tapping module; 321, device detection unit; 322, tapping driving motor; 323, tapping linkage structure; 324, tapping cutter head; 325, tool aligner; 326, suction head unit; 400, switching and cleaning device; 410, commutation auxiliary part; 420, cleaning execution module; 421, cleaning driving mechanism; 422, air outlet driving member; 423, air nozzle; 510, model generation module; 520, motion matching module; 530, instruction generation module. Detailed implementation manners

[0029] The following further details the specific implementation manners of the present invention in conjunction with the accompanying drawings, so that the technical solutions of the present invention are easier to understand and master.

[0030] Refer to Figures 1 - 5 , an automatic tapping device with multiple tapping stations: including a transfer seat 100, a blanking and sorting device 200, a switching and cleaning device 400, and at least two tapping devices 300;

[0031] Refer to Figures 1 - 5 , the transfer seat 100 includes a station turntable 110 and a rotation driving member 120, and the rotation driving member 120 is used to drive the rotation of the station turntable 110;

[0032] A loading station, a tapping station, and a blanking station are formed on the transfer seat 100. The tapping station is arranged corresponding to the tapping device 300, and the blanking and sorting device 200 is arranged at the blanking station;

[0033] The tapping device 300 includes a tapping robotic arm 310 and a tapping module 320. The tapping robotic arm 310 is used to drive the tapping module 320 to move at the tapping station, and the tapping module 320 is used to perform tapping operations. The tapping robotic arm 310 can be set as a three-axis robotic arm driven by three driving devices respectively to ensure the moving range space. The tapping module 320 includes a tapping driving motor 322, a tapping linkage structure 323, and a tapping cutter head 324. The tapping driving motor 322 drives the tapping linkage structure 323 to act. The tapping linkage structure 323 includes a quick chuck for clamping the tapping cutter head 324. The structures of the tapping cutter heads 324 set in different tapping devices 300 can be different to adapt to different scenarios. An oil supply and cooling device is also arranged on the side of the tapping cutter head 324 to assist in cooling the tapping cutter head 324. Preferably, the tapping device 300 further includes a tool setter 325, which is used to judge whether the installation position of the tapping cutter head 324 is correct and whether there is any deviation. The tool setter 325 can be set as an infrared sensor. During installation, the tapping cutter head 324 is moved to the corresponding position, and the tool setter 325 collects data once as the initial data. Then, the tool setter 325 judges whether the cutter head has any deviation according to the distance of the actual cutter head.

[0034] The switching and cleaning device 400 is arranged between the tapping devices 300. The switching and cleaning device 400 includes a commutation assisting part 410 and a cleaning execution module 420. When the part to be processed moves between the two tapping stations driven by the station turntable 110, the commutation assisting part 410 drives the part to be processed to rotate on the station turntable 110, and the cleaning execution module 420 is used to perform cleaning operations. It should be noted that the switching and cleaning device 400 is arranged between the tapping devices 300. As shown in the figure, there are two tapping devices 300. If the number of tapping devices 300 is set to three, the number of switching and cleaning devices 400 should be set to two. The commutation assisting part 410 can be set as a friction track under the commutation seat 100. When the commutation seat 100 rotates, it drives the part to move. When the part passes through the friction track, it will rotate, so that the switching and cleaning device 400 can clean the part to be processed in all directions. The cleaning execution module 420 is specifically set as a cleaning driving mechanism 421, an air outlet driving part 422, and an air nozzle 423. The cleaning driving mechanism 421 is used to drive the air nozzle 423 to move along with the air outlet driving part 422. The air outlet driving part 422 is used to generate a gas source. The air nozzle 423 is used to perform chip cleaning operations when working.

[0035] The transposition seat 100 is provided with a plurality of fixed modules 130, and the fixed modules 130 are used to fix the part to be processed at the tapping station; the transposition seat 100 is provided with a fixed groove at the tapping station, and when the part to be processed rotates to the tapping station, it falls into the fixed groove to limit the rotation of the part to be processed, and the tapping device 300 is provided with a suction head unit 326, and when the suction head unit 326 is working, it sucks the part to be processed to make it out of the fixed groove. Preferably, a clamping mechanism is provided in the fixed groove to clamp different types of processing parts, and when the processing is completed, the clamping mechanism is released, the processing part is sucked out of the fixed groove, and leaves the fixed groove, and the transposition seat 100 is taken away from the corresponding tapping station when working.

[0036] The material unloading and sorting device 200 includes an image recognition module 210 and a sorting module 220. The image recognition module 210 is configured with an image recognition strategy, which is used to recognize the image of the processed parts to generate quality data, and generate a sorting instruction based on the quality data and send it to the sorting module 220. The sorting module 220 drives the processed parts to move to the corresponding unloading area according to the sorting instruction. The sorting module 220 includes a support frame, on which a transplanting drive is arranged. The transplanting drive drives the magnetic clamp to move through the transplanting structure to drive the product to different output ports according to the quality data. The image recognition module 210 includes an image acquisition unit 211 and a focal length adjustment unit 212. The image acquisition unit 211 is arranged on the focal length adjustment unit 212. The focal length adjustment unit 212 is used to adjust the distance between the image acquisition unit 211 and the processed parts.

[0037] Reference Figure 4 The working principle is as follows: loading is carried out through the loading station, and the rotating driving member 120 drives the rotating disk to rotate to the first tapping station. After completing the first part of the tapping action, it is driven to the second tapping station. During the process, the tapping device 300 is cooled, and the cleaning device 400 is switched to clean the parts to be processed. If there are multiple actions, it will continue to go back and forth between the first tapping station and the second tapping station until all actions are completed, and then go to the unloading station for quality identification and unloading.

[0038] Reference Figure 5 , intelligent interaction subsystem, the intelligent interaction subsystem includes a model generation module 510, an action matching module 520 and an instruction generation module 530,

[0039] The model generation module is used to generate a part simulation model based on the information of the part to be processed input by the user, and process the part simulation model through the machining target information to obtain the machining difference model. The purpose of the model generation module is to perform customized processing on the model so that the machining action is based on the model. Since the information after ideal machining is known, as long as the input information is constructed, the position to be machined can be obtained by the method of model subtraction. The input information can be obtained by three-dimensional scanning of the part to be processed, and then the relative position relationship is unified. By importing the machining target information into the model, the corresponding machining difference model can be obtained. Specifically, the simulation processing of the scanned part model is as follows: , where is the model coupling operator, is the part simulation model, is the part model obtained by scanning, is the simulation parameter, and there is *K shape / E, where is the tapping stress, is the material thickness, E is the elastic modulus, and K shape is the shape correction coefficient. The generation formula of the machining difference model is , where is the target machining information, is the machining difference model.

[0040] The action matching module is configured with a tapping configuration parameter library, which stores a number of tapping constraint features. The tapping constraint features are indexed by tapping device parameters, and the tapping device parameters are as follows: all static parameters related to the tapping device (such as the movement range of the robotic arm, the type of tool head, the motor performance, the torque limit, etc.), and the dynamic constraint conditions derived from these parameters (such as the maximum feed speed, the machining angle range, the cutting force threshold). Indexed by the unique identifier of the tapping device (such as the device ID), all its parameters are associated.‌ Mechanical parameters‌: The number of axes of the robotic arm (such as three axes), the stroke range (the maximum displacement of the X / Y / Z axes), the repeat positioning accuracy (±0.01 mm).‌ Power parameters‌: The rated power of the tapping drive motor (such as 500 W), the maximum speed (3000 rpm), the torque curve (such as the maximum torque of 5 N·m).‌ Tool parameters‌: The type of tool head (such as an M6 tap), the tip angle (118°), the cooling requirement (oil cooling / gas cooling).‌ Sensor parameters‌: The detection accuracy of the tool setter (±0.005 mm), the sampling frequency of the infrared sensor (100 Hz). In this way, the tapping constraint features can be output according to the device parameters, and the constraint features determine how to split each tapping action. The constraint conditions extracted from the tapping configuration parameter library are used to limit the physical boundaries and dynamic performance of the tapping action, ensuring machining safety and accuracy. For example, the tapping constraint features can constrain the maximum feed speed and torque limit‌, tool life, cooling efficiency constraint, etc.

[0041] The action matching module obtains the corresponding tapping constraint features by substituting the tapping device parameters of different tapping devices, and generates tapping form constraints according to the tapping constraint features. The different tapping form constraints are respectively matched with the machining difference model to generate a tapping trajectory sequence. Based on the machining difference model (the area to be machined), combined with the machining path rules generated by the tapping constraint features, it is used to describe the geometric shape (such as the thread depth, helix angle) and dynamic characteristics (such as the feed speed, acceleration) of the tapping action.‌ Generation steps‌: First, perform‌ model segmentation‌: decompose the machining difference model into multiple sub-regions (such as thread hole area 1, thread hole area 2). Then perform‌ parameter matching‌: match the constraint features of the tapping device for each sub-region. For example: When machining deep hole positions, the rotation speed needs to be reduced (to avoid poor chip evacuation).‌ Form rule generation‌: For example‌ helix angle calculation‌: , where, is the helix angle constraint, is the feed speed, n is the rotation speed, and d is the thread diameter.

[0042] The tapping trajectory sequence includes several tapping sub-trajectories. The instruction generation module is configured with a static cost algorithm. The static cost algorithm is used to calculate the static cost of each tapping sub-trajectory, and generate a set of tapping trajectories based on adjacent tapping sub-trajectories with the total static cost as a constraint, and generate corresponding tapping action instructions according to the set of tapping trajectories to control the corresponding tapping device to work. The tapping form constraint is transformed into a specific motion path, which is composed of multiple tapping sub-trajectories (such as feed, cutting, retraction) in sequence. The generation method is as follows: Trajectory planning: A* algorithm: Search for the optimal path within the motion range of the robotic arm and avoid singular points. Spline interpolation: Generate a smooth acceleration curve to reduce mechanical vibration. Sub-trajectory division: Feed trajectory: The tool head quickly moves from the safe position to the machining starting point. Cutting trajectory: Thread machining is performed according to the helix angle β and the layer depth. Retraction trajectory: Rotate in the reverse direction to withdraw and return to the safe position. The motion equation of a single sub-trajectory: , where is the displacement with respect to time change, is the tool head rotation angle, t is the time,[[]] is the motion speed, is the starting position, is the acceleration change,[[]] is the starting angle, is the angle change. The cost function is designed as follows: , where,[[]] is the sub-trajectory time consumption, which reflects the time required for the motion to the next tapping device,[[]] is the energy consumption cost, which reflects the power consumption cost corresponding to the tapping equipment and cooling equipment required for this work,[[]] is the tool wear coefficient, which reflects the loss value of the tool. The longer the single work, the greater the loss cost of the tool,[[]] is the precision cost,[[]] is the corresponding configuration weight value, which can be adjusted according to the actual cost preference and reflects the influence cost value of the current tool work on the precision. Trajectory set generation: Traverse all possible permutations and combinations of sub-trajectories. Dynamic programming (DP) selects the sequence with the minimum total cost. Constraint optimization: Use the Lagrange multiplier method to handle multi-constraint problems (such as total time ≤ T_max). Example: Assume two candidate trajectory sets: Trajectory set A: Total cost = 1.2. Trajectory set B: Total cost = 1.5. The algorithm selects trajectory set A as the optimal solution.[[]]

[0043] The intelligent interaction subsystem further includes a cleaning association module, which includes a cleaning instruction set and an action association table. The cleaning instruction set includes several cleaning instructions. The cleaning instructions are indexed by action association tags. The action association table stores several action association tags. The action association tags are indexed by associated action number information. The associated action number information is the previous action number and the subsequent action number. The cleaning association module obtains the corresponding previous action number and subsequent action number according to adjacent tapping action instructions to retrieve cleaning instructions, and the cleaning instructions are used to control the cleaning execution module to work. The cleaning instruction set is a structured database that stores all executable cleaning action instructions (such as high-pressure air blowing, oil mist spraying, air nozzle movement paths, etc.). Each instruction is uniquely indexed by an action association tag. Instruction types: Basic cleaning instructions: Short-time air blowing (such as 0.5 seconds), long-time air blowing (such as 2 seconds), reciprocating cleaning. Composite cleaning instructions: Air blowing + oil mist cooling (for high-temperature debris), multi-angle cleaning (air nozzle swinging). Parameter configuration: Air pressure value (unit: MPa): Adjusted according to the debris type (such as 0.3 MPa for aluminum debris and 0.5 MPa for steel debris). Duration (unit: seconds): Dynamically calculated based on the amount of debris in the tapping action. Air nozzle path: Defines the movement trajectory of the air nozzle (such as straight line, spiral, fan shape). Definition: The action association table is a relational database that stores the mapping relationship between action association tags and associated action number information, and is used to quickly match the tapping action sequence and cleaning instructions. Data structure: Primary key: Associated action number information (previous action number + subsequent action number). Value: Action association tag (uniquely identifies the cleaning instruction). Construction algorithm: Tapping action classification: Classify tapping actions by type number (such as T1 = rough tapping, T2 = finish tapping, T3 = chamfering). Association rule definition: If the types of the previous and subsequent actions are different (such as T1→T2), trigger the tool change cleaning instruction (tag T1→T2). If the types of the previous and subsequent actions are the same (such as T2→T2), trigger the repeated machining cleaning instruction (tag T2→T2). Table storage optimization: Use a hash table to achieve O(1) time complexity query. Definition of action association tag: The action association tag is a string that uniquely identifies the combination of the previous and subsequent actions and is used to locate the corresponding cleaning instruction in the cleaning instruction set. Coding rule: Tag generation: Connect the previous action number and the subsequent action number with a separator (such as "T1→T2"). Dynamic expansion: Support automatically generating new tags when new tapping action types are added (such as when adding T4, the tag T3→T4 is automatically added to the association table). Example: The tapping action sequence is rough tapping (T1) → finish tapping (T2), then the tag is T1→T2. If finish tapping is performed twice in a row (T2→T2), then the tag is T2→T2, triggering the deep cleaning instruction.Definition of associated action number information: The associated action number information consists of a previous action number and a subsequent action number, representing the sequential relationship between two adjacent tapping actions and serving as the query basis for the action association table. Numbering rule: Previous action number: The identification of the currently completed tapping action type (such as T1). Subsequent action number: The identification of the tapping action type to be executed next (such as T2). Application scenario: If the tapping sequence is T1→T3→T2, the associated action number information is as follows: The first group: Previous = T1, Subsequent = T3 → Mark T1→T3. The second group: Previous = T3, Subsequent = T2 → Mark T3→T2. Scenario description: Tapping sequence: Rough tapping (T1) → Finish tapping (T2) → Chamfering (T3). Cleaning requirements: T1→T2: After rough machining, it is necessary to thoroughly remove aluminum chips to prevent affecting finish machining. T2→T3: After finish machining, oil mist cooling is required to avoid burrs during chamfering. Execution process: After T1 is completed, the module detects that the next action is T2 and generates the mark T1→T2. The query finds the instruction to be High-pressure blowing (0.5MPa, 2 seconds), and the air nozzle sweeps along a spiral path. After T2 is completed, it is detected that the next action is T3 and the mark T2→T3 is generated. The query finds the instruction to be Oil mist cooling (8ml / s) + Fixed-point blowing (0.3 seconds).

[0044] It also includes a dynamic detection module, which includes a device detection unit 321 and a workpiece detection unit 131. The device detection unit 321 is used to detect the status feedback data of the tapping device 300 during operation, and the workpiece detection unit 131 is used to detect the dynamic feedback data of the part to be machined during machining. The intelligent interaction subsystem also includes a dynamic correction module, which is configured with dynamic correction conditions. When the dynamic correction conditions are triggered, the dynamic correction module generates a dynamic correction instruction based on the status feedback data and the dynamic feedback data, and divides the target tapping sub-trajectory from the current tapping trajectory set to an adjacent target tapping trajectory set according to the dynamic correction instruction. The dynamic correction conditions include a device temperature sub-condition, a part temperature sub-condition, and a part chip sub-condition. The dynamic detection module consists of the device detection unit 321 and the workpiece detection unit 131, and collects the machining status data of the tapping device 300 and the workpiece in real time, providing input for dynamic correction. Device detection unit 321, function:

[0045] Monitor the physical status (vibration, temperature, current, etc.) of the tapping device 300 during operation to ensure machining stability. Core components:

[0046] Vibration sensor: Detect the vibration amplitude at the end of the robotic arm (unit: mm / s²), and set the threshold according to the ISO 10816 standard (e.g., ≤4.5 mm / s²). Temperature sensor: Monitor the temperature of the motor winding (unit: °C), and set the over-temperature threshold to 80% of the rated value (e.g., if the motor rated temperature is 80 °C, the alarm threshold is 64 °C). Current sensor: Collect the current of the tapping drive motor 322 (unit: A), and judge the load abnormality through the current fluctuation (e.g., a sudden increase in current by 20% is regarded as a risk of jamming). Encoder: Feedback the position accuracy of the robotic arm (unit: μm), and trigger calibration when the error exceeds the limit (e.g., ±5 μm). Vibration spectrum analysis: Extract the characteristic frequency through FFT (Fast Fourier Transform) to identify the resonance point. Motor health assessment: Deduce the tool wear degree based on the relationship between current and torque. Workpiece detection unit 131, Function: Monitor the dynamic deformation, force and position offset during the workpiece processing to ensure the processing accuracy. Core components: Force sensor: Measure the axial force of the tapping tool head 324 (unit: N), and trigger protection when it exceeds the limit (e.g., the processing force for steel parts > 500 N). Laser displacement sensor: Real-time detect the position offset of the workpiece (unit: μm) to compensate for mechanical errors. Strain gauge: Affix to the surface of the workpiece to detect local deformation (unit: με), and pause the processing when the deformation exceeds the limit (e.g., > 200 με). Dynamic correction module, The dynamic correction module judges whether to trigger the correction condition according to the real-time detection data, and re-plans the tapping trajectory to ensure the processing safety and accuracy. Trigger logic: Threshold trigger: Any detection parameter exceeds the preset safety range (e.g., vibration > 4.5 mm / s²). Trend trigger: The continuous change rate of the parameter is abnormal (e.g., the temperature rises > 5 °C per minute). Composite trigger: Abnormal correlation of multiple parameters (e.g., high vibration + high current). is the static cost of the i-th trajectory set, is the deviation penalty term from the original trajectory, and is the weight of the corresponding deviation penalty term. That is, the more abnormal conditions there are in the corresponding trajectory set, the higher the weighting of the deviation penalty term. When the deviation of the trajectory set is small and the other cost is small, correction can be performed, so that it can be changed from the first set of execution instructions to the second set of execution instructions. Target tapping trajectory set division:

[0047] Implementation method: Trajectory library pre-storage: Define multiple sets of tapping trajectory sets in advance (e.g., high-speed rough machining set, low-speed finish machining set).

[0048] Dynamic matching: Select the optimal trajectory set according to the correction conditions. Example:

[0049] Current trajectory set: High-speed rough machining (feed rate 0.3 mm / rev, rotational speed 3000 rpm).

[0050] Vibration over - standard detected → Trigger correction → Target trajectory set: Low - speed finish machining (feed rate 0.1mm / rev, rotational speed 1500rpm).

[0051] The intelligent interaction subsystem further includes a macro - correction module. The macro - correction module is configured with a quality - matching strategy. The quality - matching strategy generates macro - correction parameters based on quality data and corrects the static cost algorithm according to the macro - correction parameters. The macro - correction module is configured with a macro - correction database. The macro - correction database is configured with several anomaly - associated sub - items and corresponding anomaly - associated values. For the anomaly - associated sub - items and anomaly - associated values, the quality - matching strategy includes calculating the total anomaly - associated value under each anomaly - associated item. The total anomaly - associated value is the weighted result of the anomaly - associated values corresponding to the anomaly - associated sub - items included in the anomaly - associated item. When the total anomaly - associated value is greater than a preset anomaly - trigger value, the macro - correction parameters are generated according to the total anomaly - associated value. Macro - correction database, definition: A macro - correction database is a structured database that stores all the associated sub - items related to machining anomalies and their quantization parameters (anomaly - associated values), which is used to systematically analyze the root causes of quality problems and deduce correction parameters. Anomaly - associated sub - item, definition: Defines the independent factor categories that may cause quality anomalies. For example: Tool wear (sub - item ID: W01), Temperature anomaly (sub - item ID: T02), Vibration over - standard (sub - item ID: V03), Workpiece offset (sub - item ID: P04), Part flash (sub - item ID: P02), Part breakage (sub - item ID: P15). Anomaly - associated value: The quantization parameter of each sub - item, indicating the degree of influence of the anomaly on quality. Value range: 0 - 1 (0 = no influence, 1 = severe failure). Assignment rule: Based on historical data statistics (such as tool wear value = cumulative usage time / life cycle). Real - time detection data normalization (such as vibration over - standard value = current vibration amplitude / threshold). Definition: By analyzing quality data (such as machining dimension error, surface roughness), calculate the total anomaly - associated value, and generate correction parameters accordingly to optimize the static cost algorithm. Execution steps: Data mapping: Associate quality data with anomaly - associated sub - items. For example: Oversize aperture → Associate tool wear (W01) and workpiece offset (P04). Total anomaly - associated value calculation:

[0052] : The total value of the j - th anomaly - associated item (such as the "dimension anomaly" item).

[0053] : Sub - item weight (defined by the expert system, such as W01 weight = 0.6, P04 weight = 0.4).

[0054] : Sub - item associated value.

[0055] Trigger judgment: If > (such as 0.7), a macro correction parameter is generated.

[0056] For example, when the aperture tolerance is detected, calculate the associated total value: W01 associated value = 0.8 (tool wear 80%), weight 0.6 → contribution value = 0.48; P04 associated value = 0.5 (offset 50%), weight 0.4 → contribution value = 0.2. Total value = 0.48 + 0.20 = 0. If the threshold = 0.65, correction is triggered. Regarding the macro correction parameter: A parameter used to adjust the weights or constraint conditions of the static cost algorithm to specifically suppress high-frequency anomalies. Generation rule: Linear correction: Adjust the algorithm parameters proportionally according to the total anomaly association value. - ); where α is the correction coefficient (such as α = 0.5), is the adjustment amount of the k-th weight in the static cost algorithm. Non-linear correction: For severe anomalies (such as > 0.9), directly disable the high-risk trajectory set. Parameter type: Weight correction: Adjust the time, energy consumption, and wear weights in the static cost algorithm. Constraint correction: Tighten the vibration or temperature threshold (such as reducing the vibration threshold from 4.5 mm / s² to 3.5 mm / s²). For example: If correction is triggered due to tool wear ( = 0.68), the original weight of the static cost algorithm is = 0.4, = 0.3 After correction: = + Δw = 0.3 + 0.5(0.68 - 0.65) = 0.315 The tool wear weight is increased, and the algorithm is more inclined to select low-wear trajectories. Static cost algorithm correction:

[0057] Definition: Embed the macro correction parameter into the original static cost algorithm to dynamically optimize the machining trajectory selection strategy. Correction formula: Original cost function:

[0058]

[0059] Corrected cost function:

[0060] Threshold fine-tuning to avoid vibration risks in advance. Module linkage process Steps: Quality data input: The image recognition module of the blanking and sorting device provides quality data (such as aperture error 0.05 mm). Abnormal association matching: The macro correction module matches the aperture error to tool wear (W01) and workpiece offset (P04). Total value calculation and triggering: =0.68 > 0.65 → Trigger correction. Parameter generation and injection: Increase wear weight = +0.015 Reduce vibration threshold = -0.006 mm / s². Static cost algorithm update: The optimized algorithm preferentially selects low-wear and low-vibration trajectories.

[0061] The blanking and sorting device is configured with an abnormal feature database, and the abnormal feature database stores a number of quality abnormal items, each quality abnormal item is indexed by quality graphic features; the image recognition strategy generates a shooting instruction according to the processing difference model, and controls the image recognition module to shoot a part image according to the shooting instruction, and matches the quality graphic features with the part image to obtain the corresponding quality difference item, and forms the quality data according to the quality difference item. The abnormal feature database is a structured storage system that records all known quality abnormal items and their corresponding visual features (such as shape, texture, color, etc.) for quickly identifying part processing defects. Quality abnormal item: Defines the defect type in part processing, for example: Oversize hole diameter (ID: D01): The hole diameter exceeds the tolerance range (such as the standard Φ10 ± 0.05 mm, the measured Φ10.12 mm). Surface crack (ID: D02): There are visible linear cracks on the material surface. Thread burr (ID: D03): There are unremoved metal protrusions on the thread edge. Quality graphic feature: The image feature description corresponding to each abnormal item, including: Geometric feature: The shape and size of the defect area (such as crack length > 2 mm). Texture feature: Roughness and contrast extracted by the gray-level co-occurrence matrix (GLCM). Color feature: The color difference between the abnormal area and the background (such as ΔH > 10° in the HSV color space). It is a mathematical model that judges whether to trigger the image shooting and analysis action by comparing the difference between the design parameters and the measured data. Input parameter. Design drawing size (such as hole diameter Φ10 ± 0.05 mm). Theoretical accuracy of processing equipment (such as positioning error ± 3 μm). Difference calculation: If the processing parameter deviation exceeds the theoretical accuracy range, a shooting instruction is triggered. Image recognition strategy: Function: According to the output of the processing difference model, control the image recognition module to shoot a part image, and identify the quality abnormal item through feature matching.

[0062] Execution Steps: Generation of shooting instructions. Triggering timing: After processing is completed and when the difference exceeds the limit. Shooting parameters: Light source intensity (adjusted according to the material's reflectivity, e.g., 5000K cold light for aluminum alloy). Camera resolution (≥5 million pixels, with a precision corresponding to 0.01 mm / pixel). Multi-angle shooting (front view, side view, local close-up). Image feature extraction: Preprocessing: Noise reduction, edge enhancement, binarization. Feature calculation: Detect the circular aperture contour using the Hough transform. Extract the crack edge using the Canny algorithm. Analyze the thread texture using the Local Binary Pattern (LBP). Feature matching: Compare the extracted features with the indices in the abnormal feature database for similarity. Module linkage process Steps: Processing monitoring: The sensor collects the processing dimension data in real time and inputs it into the processing difference model. Difference exceeding limit judgment: The model calculates Δx. If the limit is exceeded, a shooting instruction is sent to the image recognition module. Image acquisition and analysis: Shoot the part image and extract features. Match with the abnormal feature database and generate a quality difference item report. Data feedback: The quality data is uploaded to the macro correction module to trigger dynamic adjustment of the processing parameters. For example, when batch processing flange parts, continuous thread burrs (D03) occur. Process: Difference model trigger: After processing, the deviation of the thread pitch diameter is detected to be 0.1 mm (> the threshold of 0.05 mm) → Trigger shooting. Image analysis: Extract the LBP texture feature of the thread edge and compare it with the D03 index, SSIM = 0.72. It is determined to be a D03 abnormality with a severity level of 2. Quality data generation: Mark that 10% of the parts in this batch have D03 defects. Macro correction response: The macro correction module increases the tool compensation weight and adds a deburring process dwell time of 0.2 seconds. Result: The D03 incidence rate in subsequent batches drops to 2%.

[0063] Of course, the above are only typical examples of the present invention. In addition, the present invention can also have many other specific implementation manners. Any technical solutions formed by equivalent replacement or equivalent transformation fall within the scope of protection required by the present invention.

Claims

1. An automatic tapping device with multiple tapping stations, characterized in that: It includes a transposition seat, a material sorting device, a switching cleaning device, and at least two tapping devices; The transposition seat includes a station turntable and a rotation driving member, and the rotation driving member is used to drive the station turntable to rotate; The transposition seat is provided with a loading station, a tapping station and a unloading station, wherein the tapping station is arranged corresponding to the tapping device, and the unloading and sorting device is arranged at the unloading station; The tapping device includes a tapping robot arm and a tapping module, wherein the tapping robot arm is used to drive the tapping module to move at the tapping station, and the tapping module is used to perform the tapping action; The switching cleaning device is arranged between the tapping devices, and includes a reversing auxiliary part and a cleaning execution module. When the part to be processed moves between two tapping stations driven by the station turntable, the reversing auxiliary part drives the part to be processed to rotate on the station turntable, and the cleaning execution module is used to perform the cleaning action; The transposition seat is provided with a plurality of fixed modules, and the fixed modules are used to fix the parts to be processed at the tapping station; The material unloading and sorting device includes an image recognition module and a sorting module. The image recognition module is configured with an image recognition strategy. The image recognition strategy is used to recognize the image of the processed parts to generate quality data, and generate sorting instructions based on the quality data and send them to the sorting module. The sorting module drives the processed parts to move to the corresponding unloading area according to the sorting instructions.

2. The multi-tapping station automatic tapping equipment according to claim 1, characterized in that: The transposition seat is provided with a fixed groove on the tapping station. When the part to be processed is rotated to the tapping station, it falls into the fixed groove to limit the rotation of the part to be processed. The tapping device is provided with a suction head unit. When the suction head unit is working, it sucks the part to be processed to make it leave the fixed groove.

3. The multi-tapping station automatic tapping equipment according to claim 1, characterized in that: The image recognition module includes an image acquisition unit and a focus adjustment unit. The image acquisition unit is arranged on the focus adjustment unit. The focus adjustment unit is used to adjust the distance between the image acquisition unit and the processed part.

4. A multi-tapping station automatic tapping control system, characterized in that: An automatic tapping device with multiple tapping stations configured as described in any one of claims 1 to 3, further comprising an intelligent interaction subsystem, wherein the intelligent interaction subsystem comprises a model generation module, an action matching module and an instruction generation module, wherein the model generation module is used to generate a part simulation model according to the part information to be processed input by a user, and to process the part simulation model through processing target information to obtain the processing difference model, wherein the action matching module is configured with a tapping configuration parameter library, wherein the tapping configuration parameter library stores a plurality of tapping constraint features, wherein the tapping constraint features are indexed by tapping device parameters, and the action matching module is used to generate a part simulation model according to the part information to be processed input by a user, and to process the part simulation model through processing target information to obtain the processing difference model, wherein the action matching module is configured with a tapping configuration parameter library, wherein the tapping configuration parameter library stores a plurality of tapping constraint features, wherein the tapping constraint features are indexed by tapping device parameters, and the action matching module is used to generate a part simulation model according to the part information to be processed input by a user, and to process the part simulation model according to the processing target information to obtain the processing difference model, and ... used to generate a part simulation model according to the part information to be processed The matching module obtains the corresponding tapping constraint features by substituting the tapping device parameters of different tapping devices, and generates tapping morphological constraints according to the tapping constraint features, and matches the different tapping morphological constraints with the processing difference models respectively to generate a tapping trajectory sequence, wherein the tapping trajectory sequence includes a plurality of tapping sub-trajectories, and the instruction generation module is configured with a static cost algorithm, which is used to calculate the static cost of each tapping sub-trajectory, and generate a tapping trajectory set according to adjacent tapping sub-trajectories with the total static cost as a constraint, and generate corresponding tapping action instructions according to the tapping trajectory set to control the operation of the corresponding tapping device.

5. The multi-tapping station automatic tapping control system according to claim 4, characterized in that: The intelligent interactive subsystem also includes a cleaning association module, which includes a cleaning instruction set and an action association table. The cleaning instruction set includes a plurality of cleaning instructions, and the cleaning instructions are indexed by action association tags. The action association table stores a plurality of action association tags, and the action association tags are indexed by associated action number information. The associated action number information is a front action number and a rear action number. The cleaning association module obtains the corresponding front action number and rear action number according to adjacent tapping action instructions to retrieve the cleaning instruction, and the cleaning instruction is used to control the operation of the cleaning execution module.

6. The multi-tapping station automatic tapping control system according to claim 4, characterized in that: It also includes a dynamic detection module, which includes a device detection unit and a workpiece detection unit. The device detection unit is used to detect the state feedback data of the tapping device when it is working, and the workpiece detection unit is used to detect the dynamic feedback data of the part to be processed when it is being processed. The intelligent interactive subsystem also includes a dynamic correction module, and the dynamic correction module is configured with a dynamic correction condition. When the dynamic correction condition is triggered, the dynamic correction module generates a dynamic correction instruction based on the state feedback data and the dynamic feedback data, and divides the target tapping sub-trajectory from the current tapping trajectory set to the adjacent target tapping trajectory set according to the dynamic correction instruction.

7. The multi-tapping station automatic tapping control system according to claim 4, characterized in that: The intelligent interaction subsystem also includes a macro correction module, which is configured with a quality matching strategy. The quality matching strategy generates macro correction parameters according to quality data and corrects the static cost algorithm according to the macro correction parameters.

8. The multi-tapping station automatic tapping control system according to claim 4, characterized in that: The dynamic correction conditions include device temperature sub-conditions, component temperature sub-conditions, and component chip sub-conditions.

9. The multi-tapping station automatic tapping control system according to claim 5, characterized in that: The material sorting device is equipped with an abnormal feature database, which stores a number of quality abnormality items, each of which is indexed by a quality graphic feature; The image recognition strategy generates a shooting instruction according to the processing difference model, and controls the image recognition module to shoot the part image according to the shooting instruction, matches the quality graphic features and the part image to obtain the corresponding quality difference items, and forms the quality data according to the quality difference items.

10. The multi-tapping station automatic tapping control system according to claim 9, characterized in that: The macro correction module is configured with a macro correction database, and the macro correction database is configured with a number of abnormal association sub-items and corresponding abnormal association values. The abnormal association sub-items and abnormal association values, the quality matching strategy includes calculating the total abnormal association value under each abnormal association item, the total abnormal association value is the weighted result of the abnormal association values ​​corresponding to the abnormal association sub-items contained in the abnormal association item, when the total abnormal association value is greater than the preset abnormal trigger value, the macro correction parameter is generated according to the total abnormal association value.

Images