Intelligent self-adaptive screw locking control method for motor

Abstract

The invention relates to the technical field of automobile part optimization, in particular to an intelligent self-adaptive screw locking control method for a motor, which can acquire an image of an assembly area in real time and establish a three-dimensional coordinate model of a screw hole pose through a camera integrated at the front end of the motor and a three-dimensional visual recognition algorithm. According to the method, the position and the posture of the screw hole can be determined with high precision, so that the vertical alignment relation between the screwdriver blade and the target screw hole is ensured, and the consistency and the accuracy of locking operation are improved.

Description

Technical Field

[0001] This invention relates to the field of automotive parts optimization technology, and in particular to an intelligent adaptive screw-locking control method for motors. Background Technology

[0002] Currently, with the development of industrial automation, motor-driven screw tightening technology has become a key process in the assembly field. In existing technologies, motor-driven screw tightening operations typically rely on an external PLC (Programmable Logic Controller) system to implement the control logic. The operation is completed collaboratively through independently configured vision modules, torque sensors, and motion control modules. Specifically, in the motor-driven screw tightening process, traditional methods usually rely on manual operation or simple automated equipment to complete the screw tightening. These methods mainly include:

[0003] 1. Manual tools: Using manual tools such as screwdrivers to tighten screws is inefficient and can easily lead to screws being overtightened or undertightened.

[0004] 2. Semi-automatic mechanical devices: These devices use electric or pneumatic screwdrivers to initially tighten screws by setting pre-defined speed and torque limits. However, these devices lack intelligent adjustment functions and cannot dynamically adjust operating parameters according to actual working conditions.

[0005] 3. Basic Automation Systems: These systems use basic sensors (such as torque sensors) to monitor torque changes during screw tightening and make simple adjustments accordingly. However, these systems have limited adaptability and accuracy.

[0006] However, existing technologies have the following drawbacks:

[0007] 1. Accuracy and consistency issues: Traditional methods have difficulty guaranteeing the consistency of screw tightening each time, especially in applications that require high precision, such as electronic equipment assembly and precision machinery manufacturing.

[0008] 2. Lack of intelligent adjustment mechanism: Many existing systems cannot monitor and respond to various variables in the screw tightening process in real time, such as screw hole position deviation and axial pressure change, thus affecting the tightening quality.

[0009] 3. Weak ability to handle abnormal situations: For common abnormal situations such as stripped threads, thread damage, and foreign object jamming, existing technologies often lack effective identification and response strategies, which may lead to assembly failure or a decline in product quality.

[0010] Therefore, there is a need for an intelligent adaptive screw-locking control method for motors that can solve the above problems. Summary of the Invention

[0011] This invention provides an intelligent adaptive screw-locking control method for motors. By integrating a camera and a 3D vision recognition algorithm at the front end of the motor, this invention can acquire images of the assembly area in real time and establish a 3D coordinate model of the screw hole's position and orientation. This method can accurately determine the position and orientation of the screw hole, thereby ensuring the vertical alignment between the screwdriver tip and the target screw hole, improving the consistency and accuracy of the tightening operation.

[0012] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows:

[0013] A smart adaptive screw-locking control method for electric motors includes the following steps:

[0014] Step S1: Obtain real-time images of the assembly area using a camera integrated at the front end of the motor, and establish a three-dimensional coordinate model of the screw hole pose based on a three-dimensional vision recognition algorithm.

[0015] Step S2: The motor performs pose calibration according to the three-dimensional coordinate model to establish a vertical alignment between the screwdriver head and the target screw hole, and then starts the motor to rotate and press down.

[0016] Step S3: When starting the motor rotation and pressing operation, simultaneously collect the real-time current value of the motor winding, the rotation angle of the output shaft, and the axial pressure data;

[0017] Step S4: Establish a dynamic torque model, compare the real-time current value of the motor winding with the preset current-torque mapping relationship, and calculate the actual output torque by combining the axial pressure value fed back by the pressure sensor.

[0018] Step S5: When the actual output torque is detected to exceed the preset torque threshold range, immediately reduce the motor speed to below the safety threshold and start the reverse rotation retraction program;

[0019] Step S6: When the screwdriver head rotates to 90%-95% of the preset standard angle value, switch to precision torque control mode and complete the final tightening at a speed 20%-30% lower than the initial motor speed.

[0020] Step S7: After completing the locking operation, record the characteristic parameter set of this operation, including the maximum torque value, the final stable torque, the total time, and the abnormal event marker.

[0021] Furthermore, the three-dimensional visual recognition algorithm in step S1 includes:

[0022] S1-1. Collect stereo vision images through a dual-camera module and generate depth point cloud data;

[0023] S1-2. Extract screw hole contour feature points based on edge detection algorithm;

[0024] S1-3. The iterative nearest point algorithm is used to register the real-time point cloud data with the preset standard hole position model;

[0025] S1-3. Calculate the spatial angle deviation between the screw hole center axis and the motor axis.

[0026] Furthermore, the motor rotation and pressing operation in step S2 includes:

[0027] a. Set the initial rotational speed and axial pressure value of the motor output shaft;

[0028] b. The motor rotates at the set initial rotational speed and simultaneously applies a preset axial pressure to the screwdriver head.

[0029] Furthermore, the method for establishing the dynamic torque model in step S4 includes:

[0030] S4-1. Establish a current-torque reference curve under no-load conditions of the motor;

[0031] S4-2. Dynamically compensate the reference curve based on the real-time detected axial pressure value of the motor output shaft. For every 1N increase in pressure, the torque compensation coefficient increases by 0.5%-0.8%.

[0032] Furthermore, the reverse rotation and retraction procedure in step S5 specifically includes:

[0033] S5-1. Record the rotation angle position of the motor output shaft when the abnormal torque occurs;

[0034] S5-2. Rotate the motor output shaft in the reverse direction at 50% of the normal speed to the position 2-3 revolutions before the abnormality occurred;

[0035] S5-3, When re-attempting locking, use an initial torque threshold that is reduced by 20%.

[0036] Furthermore, the anomaly diagnosis procedure in step S7 includes:

[0037] S7-1. Establish a current fluctuation spectrum characteristic database, including typical patterns of stripped threads, thread damage, and foreign object jamming;

[0038] S7-2. Perform wavelet transform on the real-time current signal to extract the energy distribution characteristics of 3-5 key frequency bands;

[0039] S7-3. Use a pattern matching algorithm to compare the real-time features with the database to determine the anomaly type.

[0040] Furthermore, in step S6, if the standard angle value is not reached within the predetermined time window, an abnormality diagnosis program is triggered. The program analyzes the frequency characteristics of current fluctuations to determine whether stripping or thread misalignment has occurred. When thread misalignment is detected, a multi-level compensation strategy is automatically executed.

[0041] Furthermore, the multi-level compensation strategy includes:

[0042] Level 1 compensation: Keep the downforce constant and adjust the rotation angle of the motor output shaft by ±2°-5°;

[0043] Secondary compensation: Change the downforce to 80%-120% of the initial value;

[0044] Level 3 compensation: Switch to pulse rotation mode to lock the device by intermittent rotation.

[0045] Furthermore, after completing the final locking at a speed 20%-30% lower than the initial motor speed in step S6, an anti-loosening detection step is also included:

[0046] a. After locking, maintain the preset holding torque for 10-15 seconds and monitor the amount of retraction of the motor shaft rotation angle;

[0047] b. When the retraction amount exceeds 0.5°, a secondary locking procedure is triggered until the retraction amount is less than 0.2°.

[0048] Furthermore, in step S7, after recording the feature parameter set of this operation, including the maximum torque value, the final stable torque, the total time consumption, and the abnormal event markers, the historical feature parameter set is analyzed by machine learning through the adaptive learning module to dynamically optimize the torque threshold and motor speed curve for subsequent operations.

[0049] The advantages of this invention are:

[0050] 1. Improve accuracy and consistency:

[0051] This invention utilizes a camera integrated into the front end of the motor and a 3D vision recognition algorithm to acquire images of the assembly area in real time and establish a 3D coordinate model of the screw hole's pose. This method can accurately determine the position and orientation of the screw hole, thereby ensuring the vertical alignment between the screwdriver tip and the target screw hole, and improving the consistency and accuracy of the tightening operation.

[0052] When the screwdriver head rotates to 90%-95% of the preset standard angle value, the present invention switches to precision torque control mode to complete the final tightening at a speed 20%-30% lower than the initial speed of the motor, thereby further improving the tightening quality.

[0053] 2. Implement an intelligent adjustment mechanism:

[0054] This invention establishes a dynamic torque model, compares the real-time current value of the motor winding with a preset current-torque mapping relationship, and combines this with the axial pressure value fed back by the pressure sensor to calculate the actual output torque. This model can dynamically adjust operating parameters according to real-time operating conditions to ensure the locking quality of each tightening.

[0055] This invention records the feature parameter set of each operation and performs machine learning analysis on the historical feature parameter set through an adaptive learning module, dynamically optimizing the torque threshold and motor speed curve of subsequent operations, thereby achieving self-optimization and improvement of the system.

[0056] 3. Enhance the ability to handle abnormal situations:

[0057] When the actual output torque is detected to exceed the preset torque threshold range, the motor speed is immediately reduced to below the safety threshold, and the reverse rotation retraction program is started, effectively avoiding equipment damage or product quality problems caused by overload.

[0058] When the standard angle value is not reached within the predetermined time window, the abnormal diagnosis program is triggered and a multi-level compensation strategy is automatically executed, including adjusting the rotational entry angle of the motor output shaft, changing the downforce, and switching to pulse rotation mode, which can effectively deal with problems such as stripped teeth and thread misalignment.

[0059] Anti-loosening detection steps: After tightening, maintain the preset holding torque for 10-15 seconds and monitor the amount of back movement of the motor shaft. If the back movement exceeds the set threshold, a secondary tightening procedure is triggered to ensure that the screw is securely locked.

[0060] 4. Improve overall efficiency and security

[0061] This invention's screw-locking control method employs multiple intelligent techniques, such as 3D vision recognition, dynamic torque models, and reverse rotation / retraction programs. This not only improves the accuracy and reliability of the locking operation but also reduces the need for manual intervention, thereby increasing production efficiency. By recording and analyzing data from each operation, the system can continuously learn and optimize its operating parameters, thus improving long-term operational stability and reliability. Detailed Implementation

[0062] The technical solution of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] Example 1:

[0064] A smart adaptive screw-locking control method for a motor includes the following steps: Step S1: Acquire real-time images of the assembly area using a camera integrated at the front end of the motor, and establish a three-dimensional coordinate model of the screw hole pose based on a three-dimensional visual recognition algorithm; Step S2: Calibrate the motor pose according to the three-dimensional coordinate model, and start the motor rotation and pressing operation after establishing a vertical alignment between the screwdriver head and the target screw hole; Step S3: When the motor rotation and pressing operation is started, simultaneously collect the real-time current value of the motor winding, the output shaft rotation angle, and the axial pressure data; Step S4: Establish a dynamic torque model, and compare the real-time current value of the motor winding with a preset current-torque model. The torque mapping relationship is compared, and the actual output torque is calculated based on the axial pressure value fed back by the pressure sensor. Step S5: When the actual output torque exceeds the preset torque threshold range, the motor speed is immediately reduced to below the safety threshold, and the reverse rotation retraction program is started. Step S6: When the screwdriver head rotation angle reaches 90%-95% of the preset standard angle value, the precision torque control mode is switched to complete the final tightening at a speed 20%-30% lower than the initial motor speed. Step S7: After the tightening operation is completed, the characteristic parameter set of this operation is recorded, including the maximum torque value, the final stable torque, the total time, and the abnormal event marker. The dynamic torque model refers to a mathematical model that dynamically calculates the output torque based on the real-time collected motor parameters. Specifically, it can be implemented using a current-torque mapping relationship combined with axial pressure compensation. The precision torque control mode refers to a high-precision low-speed control method used in the final stage of tightening. Specifically, it can be implemented using a closed-loop PID control algorithm.

[0065] The core innovation of this application lies in proposing an integrated and adaptive screw-locking control method. By directly integrating the vision module into the front end of the motor, a tight coupling between image acquisition and motor control is achieved, significantly improving the system's real-time performance. Simultaneously, the adoption of a dynamic torque model and a multi-level adaptive control strategy greatly enhances torque control accuracy and anomaly handling capabilities. Furthermore, the introduction of feature parameter recording and an adaptive learning mechanism enables dynamic optimization of control parameters, giving the system self-optimization capabilities.

[0066] The working principle of this application is as follows: First, a camera integrated at the front end of the motor acquires real-time images of the assembly area. Using a 3D vision recognition algorithm, the system establishes a 3D coordinate model of the screw hole's pose. This step achieves high-precision target positioning, laying the foundation for subsequent operations. The motor performs pose calibration based on the established 3D coordinate model. Through precise adjustment, the screwdriver head is vertically aligned with the target screw hole, ensuring the accuracy of the starting position of the tightening operation. When the motor rotation and pressing operation is initiated, the system simultaneously collects multiple real-time data, including the current value of the motor windings, the output shaft rotation angle, and axial pressure data. The motor rotation and pressing operation includes: a. setting the initial rotation speed and axial pressure value of the motor output shaft; b. the motor rotates at the set initial rotation speed while simultaneously applying a preset axial pressure to the screwdriver head. This method enables precise control of the motor rotation and pressing operation.

[0067] Specifically, the improved method of this application achieves refined management of the screw-locking process by preset initial parameters and synchronously controlling rotation and pressure. The initial rotation speed directly affects the initial tightening effect of the screw, while the axial pressure value determines the contact stability between the screwdriver bit and the screw hole. Proper configuration of these two parameters is crucial for preventing thread stripping and ensuring perpendicular screw insertion. For example, different initial rotation speeds can be set for screws of different sizes. Small screws require higher initial speeds to quickly complete the initial tightening, while large screws require lower initial speeds to ensure stability. Similarly, the axial pressure value also needs to be adjusted according to the screw material and hole characteristics. Soft materials require lower pressure to avoid deformation, while hard materials require higher pressure to ensure good contact. By synchronously controlling rotation and pressure, dynamic adjustment of the entire screw-locking process is achieved. Therefore, the rotation speed and pressure can be adjusted based on real-time feedback information to adapt to the tightening requirements at different stages. For example, when the screw first contacts the thread, lower pressure and speed are needed to ensure correct alignment; while after the thread is fully engaged, the speed and pressure can be appropriately increased to accelerate the tightening process.

[0068] In a preferred embodiment, the initial rotational speed of the motor output shaft can be set to 300-500 rpm, a range that typically meets the requirements of most common screw specifications. The axial pressure value can be set to 5-10 N, a range that provides sufficient contact pressure in most cases without causing material deformation. By implementing the improved method of this application, the accuracy and efficiency of screw fastening can be significantly improved. Compared with traditional fixed-parameter screw fastening methods, the method of this application can reduce thread damage or insufficient tightening caused by improper parameters, thereby improving assembly quality and production efficiency. In addition, since the parameters can be flexibly adjusted according to different screw types and material characteristics, this method also has strong adaptability and can meet diverse assembly needs. Compared with the prior art, the improved method of this application has significant advantages in the following aspects: First, by preset initial parameters, the "trial and error" process common in traditional methods is avoided, reducing thread damage caused by improper parameters; by synchronously controlling rotation and pressure, refined management of the screw fastening process is achieved, improving the tightening quality.

[0069] Furthermore, the method for establishing the dynamic torque model in step S4 includes: S4-1, establishing a current-torque reference curve under motor no-load conditions; S4-2, dynamically compensating the reference curve based on the real-time detected axial pressure value of the motor output shaft, with the torque compensation coefficient increasing by 0.5%-0.8% for every 1N increase in pressure. The core of this dynamic torque model establishment method lies in considering the influence of axial pressure on torque. In the actual screw-tightening process, changes in axial pressure significantly affect the actual torque output by the motor. By introducing axial pressure as a dynamic compensation factor, the actual output torque can be estimated more accurately, thereby improving the accuracy and reliability of the screw-tightening process.

[0070] Specifically, the dynamic torque model establishment method of this application can be implemented according to the following steps: Under no-load conditions, by gradually increasing the motor input current while simultaneously measuring the output shaft torque, an initial current-torque reference curve is established. This reference curve reflects the relationship between the motor input current and output torque under no-load conditions. During the actual screw tightening process, the axial pressure on the motor output shaft is monitored in real time (achieved by installing a pressure sensor on the motor output shaft). Based on the detected axial pressure value of the motor output shaft, the reference curve is dynamically compensated. The compensation method is as follows: for every 1N increase in axial pressure, the corresponding torque compensation coefficient increases by 0.5%-0.8%. For example, if the detected axial pressure is 10N, then the actual torque will be 5%-8% higher than the value predicted by the reference curve.

[0071] As a specific embodiment, suppose a miniature screw needs to be tightened into a plastic housing on a precision electronic product assembly line. First, an experimental test was conducted to establish a no-load current-torque reference curve for the motor, revealing a no-load torque of 0.5 Nm at a 2A current. During the actual screw-tightening process, when the motor current was 2A, the pressure sensor detected an axial pressure of 5 N. According to the method of this application, a torque compensation coefficient of 1.025-1.04 (i.e., 5*0.5% to 5*0.8%) was calculated. Therefore, the estimated actual output torque is 0.5125-0.52 Nm, which is 2.5%-4% higher than under no-load conditions. The control system adjusts the motor output based on this more accurate torque estimate to ensure the screw is tightened to the appropriate force, preventing loosening due to insufficient force or damage to the threads due to excessive force.

[0072] Furthermore, the reverse rotation and retraction procedure in step S5 specifically includes:

[0073] S5-1. Record the rotation angle position of the motor output shaft when the abnormal torque occurs;

[0074] S5-2. Rotate the motor output shaft in the reverse direction at 50% of the normal speed to the position 2-3 revolutions before the abnormality occurred;

[0075] S5-3, When re-attempting locking, use an initial torque threshold that is reduced by 20%.

[0076] This reverse rotation retraction procedure is designed to address the issue of rapid recovery in the event of abnormal torque. By recording the angular position at the time of the anomaly, the system can accurately pinpoint the location of the problem. Reversing at 50% of the normal rotation speed allows for a safe and rapid return to the appropriate position. Choosing to retract 2-3 revolutions before the anomaly occurs ensures complete escape from the potential anomaly area without excessive retraction affecting efficiency. Implementing this procedure first requires real-time monitoring of the motor output shaft's rotation angle using an encoder or other angle sensor. When an abnormal torque is detected, the control system immediately records the current angle value. This step can be achieved by setting an angle variable in the controller, which is updated with each detection cycle and locked when an anomaly is triggered. Next, the system calculates the required reverse rotation angle. Assuming the motor speed during normal locking is 1000 rpm, the reverse rotation speed is set to 500 rpm. This speed adjustment can be achieved by modifying the motor driver's speed parameters. The system also needs to calculate the angle value corresponding to 2-3 revolutions; for example, if a retraction of 2.5 revolutions is chosen, the reverse rotation angle will be 900°. During reverse rotation, the system continuously monitors angle changes. A counter can be set up to increment the count whenever the angle change reaches a certain value (e.g., 1°). When the count reaches a preset reverse rotation angle value, the system stops rotating in the reverse direction.

[0077] When re-attempting tightening, the system reduces the initial torque threshold by 20%. For example, if the original torque threshold is 10 Nm, the new threshold will be set to 8 Nm. This can be achieved by modifying the torque limit parameter in the control algorithm. This reverse rotation retraction procedure effectively handles torque anomalies encountered during tightening. Through precise angle control and appropriate speed adjustment, the system can quickly return to a safe position. Reducing the initial torque threshold increases the likelihood of successful re-tightening. Compared to existing technologies, the reverse rotation retraction procedure of this application has significant advantages. Traditional methods typically use a fixed retraction distance, which cannot be flexibly adjusted according to actual anomalies. This application, by recording the specific location of the anomaly, can achieve more precise retraction control. Furthermore, traditional methods often use the same parameters when re-attempting tightening, easily leading to repeated failures. This application, by reducing the initial torque threshold, increases adaptability and improves the success rate of re-tightening. By implementing this reverse rotation retraction procedure, this application can significantly improve the reliability and efficiency of screw tightening operations. When encountering torque anomalies, the system can quickly resume normal operation, reducing downtime and the need for manual intervention. This is especially important for automated production lines that require high precision and efficiency, as it can effectively reduce production interruptions and quality problems caused by screw tightening failures.

[0078] Furthermore, the anomaly diagnosis procedure in step S7 includes: S7-1, establishing a current fluctuation spectrum feature database, containing typical patterns of stripped threads, thread damage, and foreign object jamming; S7-2, performing wavelet transform on the real-time current signal to extract energy distribution features of 3-5 key frequency bands; S7-3, using a pattern matching algorithm to compare the real-time features with the database to determine the anomaly type. The advantage of this anomaly diagnosis method is that it can analyze the current signal during the screw-locking process in real-time and dynamically, and quickly identify the anomaly type by comparing it with a pre-established feature database. Compared to traditional methods based on simple threshold judgment, the solution in this application has higher recognition accuracy and stronger adaptability.

[0079] For example, when a stripped thread occurs, the current signal will exhibit a specific periodic fluctuation pattern; while thread damage may cause sudden spikes in the current signal. By analyzing these characteristics, the system can accurately distinguish between different types of anomalies and take corresponding remedial measures.

[0080] As a preferred implementation, the current fluctuation spectrum feature database includes the following: 1. Slippage feature: periodic low-amplitude current fluctuations, with frequency related to thread pitch. 2. Thread damage feature: irregular high-amplitude current fluctuations, usually accompanied by significant torque abrupt changes. 3. Foreign object jamming feature: continuous high current values, accompanied by a significant decrease in rotational speed. During wavelet transform, commonly used wavelet basis functions such as Daubechies wavelet or Morlet wavelet can be selected to perform multi-scale decomposition of the current signal. Typically, 3-5 frequency bands are selected for analysis, such as 0-10Hz, 10-50Hz, 50-200Hz, 200-500Hz, and above 500Hz. These frequency bands correspond to different types of anomaly features. The pattern matching algorithm can employ methods such as Euclidean distance, cosine similarity, or dynamic time warping (DTW). For example, when using Euclidean distance, a threshold of 0.8 can be set; when the distance between a real-time feature and a certain pattern in the database is less than 0.8, it is determined to be an anomaly of that type. This method enables the rapid and accurate identification of abnormal situations during screw fastening, providing a reliable basis for subsequent compensation strategies. This not only improves the success rate of screw fastening operations but also effectively reduces equipment damage and production interruptions caused by abnormal situations.

[0081] Furthermore, in step S6, if the standard angle value is not reached within the predetermined time window, an abnormality diagnosis program is triggered. This program analyzes the frequency characteristics of current fluctuations to determine if stripping or thread misalignment has occurred. By introducing an abnormality diagnosis program and a multi-level compensation strategy, this invention effectively solves the problem of abnormalities such as stripping or thread misalignment that may occur during the tightening process. By analyzing the frequency characteristics of current fluctuations, the system can accurately identify the type of abnormality and take corresponding compensation measures according to the specific situation, thereby improving the success rate and reliability of screw tightening operations. Specifically, the abnormality diagnosis program first analyzes the real-time current value of the motor windings to determine if an abnormality has occurred. When the system detects that the frequency characteristics of current fluctuations match the preset stripping or thread misalignment pattern, it triggers the corresponding diagnostic process. For example, stripping is usually characterized by a high frequency and small amplitude of current fluctuations, while thread misalignment may be characterized by sudden large current fluctuations. Furthermore, when the system identifies thread misalignment, it automatically executes a multi-level compensation strategy. This strategy can include multiple levels of adjustment measures, such as: 1. Fine-tuning the rotation angle of the motor output shaft to attempt to realign the threads; 2. Changing the rotation mode, such as using pulsed rotation to overcome thread jamming. This dynamic adjustment method allows the system to respond flexibly to different abnormal situations, greatly improving the adaptability and success rate of screw-tightening operations.

[0082] As a preferred implementation, a dedicated anomaly diagnosis module can be integrated into the motor control system. This module can acquire and analyze the motor's current signal in real time, and identify different types of anomaly patterns through pre-trained machine learning algorithms. For example, signal processing techniques such as wavelet transform or Fourier transform can be used to extract the frequency domain features of the current signal, and then classification algorithms such as support vector machine (SVM) or convolutional neural network (CNN) can be used to determine the anomaly type. In specific implementation, the system can set a predetermined time window, such as 5 seconds. If the screw rotation angle fails to reach the expected standard angle value within these 5 seconds (e.g., expected to rotate 360° but actually only rotated 300°), the anomaly diagnosis program is triggered. The anomaly diagnosis program immediately analyzes the current fluctuation data within the most recent second and extracts its frequency features. If a high-frequency (e.g., 50-100Hz) small fluctuation (peak-to-peak value less than 10% of the rated current) is detected, the system will determine that a stripped thread has occurred; if a sudden large fluctuation (peak value exceeding 50% of the rated current) is detected, it will determine that the thread is misaligned. When thread misalignment is detected, the system will attempt to tighten the screw step by step according to a preset multi-level compensation strategy. First, the motor output shaft is retracted 2-3°, and then the screw is re-attempted at a lower speed (e.g., 50% of the initial speed). If this fine-tuning is ineffective, the system will proceed to the next level of compensation, for example, increasing the downward pressure to 120% of the initial value. If the problem still cannot be resolved, the system switches to a pulse rotation mode, pausing for 0.1 seconds after each 45° rotation, completing the remaining rotation in this rhythm. Through this multi-level compensation strategy, this application can effectively handle various complex thread misalignment situations, significantly improving the success rate of screw tightening operations. Compared to traditional methods, the technical solution of this application has the following advantages:

[0083] 1. Strong real-time response capability: By analyzing the current fluctuation characteristics in real time, the system can react quickly as soon as a problem occurs, avoiding thread damage caused by delayed processing.

[0084] 2. High diagnostic accuracy: By utilizing the frequency characteristics of current fluctuations, the system can accurately distinguish between different types of anomalies such as stripped teeth and thread misalignment, providing a precise basis for subsequent compensation strategies.

[0085] 3. Flexible compensation strategies: Multi-level compensation strategies can take corresponding measures according to different abnormal situations, which greatly improves the system's ability to handle complex situations.

[0086] 4. Reduced human intervention: Automated anomaly diagnosis and handling processes greatly reduce the need for human intervention, improving production efficiency and consistency.

[0087] Compared to existing technologies, the solution presented in this application has significant advantages in handling abnormal situations during screw fastening. Traditional methods typically rely on simple torque or angle thresholds to determine abnormalities, failing to accurately identify the type of abnormality and lacking flexible compensation strategies. In contrast, this application, by analyzing the frequency characteristics of current fluctuations, can more accurately identify the type of abnormality and employs a multi-level compensation strategy to handle complex situations, greatly improving the system's adaptability and reliability. This method not only effectively reduces the probability of screw fastening failure but also maximizes the protection of threads and workpieces, avoiding damage caused by repeated attempts.

[0088] When thread misalignment is detected, a multi-level compensation strategy is automatically executed, which includes:

[0089] Level 1 compensation: Keep the downforce constant and adjust the rotation angle of the motor output shaft by ±2°-5°;

[0090] Secondary compensation: Change the downforce to 80%-120% of the initial value;

[0091] Level 3 compensation: Switch to pulse rotation mode to lock the device by intermittent rotation.

[0092] To better understand these compensation mechanisms, each compensation level is explained in detail below:

[0093] Level 1 compensation: Keep the downforce constant, adjust the rotation angle by ±2°-5°

[0094] The downward pressure remains constant, maintaining the axial pressure currently applied to the screw.

[0095] Adjust the rotational approach angle of the motor output shaft (i.e., the initial contact angle of the screwdriver tip relative to the screw). By fine-tuning this angle, the angle at which the screw enters the screw hole can be optimized, ensuring that the screw can enter more smoothly and reducing the possibility of slippage or misalignment.

[0096] Secondary compensation: Change the downforce to 80%-120% of the initial value.

[0097] If primary compensation is insufficient to resolve the issue, adjust the axial pressure applied to the screw. The new downward pressure range is 80% to 120% of the initial setting. This can be achieved by increasing or decreasing the axial pressure to accommodate different locking requirements and material properties.

[0098] By adjusting the axial pressure, the screw insertion process can be better controlled, especially when encountering harder materials or requiring greater stability.

[0099] Level 3 compensation: Switch to pulse rotation mode to lock the device by intermittent rotation.

[0100] Pulse Rotation Mode: When primary and secondary compensation fail to achieve the desired effect, the system switches to pulse rotation mode. In this mode, the motor output shaft rotates intermittently instead of continuously.

[0101] Intermittent rotation method: By using short rotation and pause cycles, the screw can be gradually driven into the screw hole. This method helps to provide better control under complex assembly conditions and avoid excessive torque or damage to the screw and workpiece.

[0102] The rotational engagement angle of the motor output shaft refers to the initial contact angle of the motor output shaft (the part connected to the screwdriver head) relative to the screw. Adjusting this angle optimizes the screw's entry into the screw hole. The downward pressure refers to the axial force applied to the screw to ensure its stable entry into the screw hole.

[0103] The multi-level compensation strategy in this application aims to solve the thread misalignment problem and improve the locking success rate. This strategy adapts to different degrees of thread misalignment by progressively adjusting the rotational approach angle, downforce, and rotation mode of the motor output shaft. Specifically, the first-level compensation keeps the downforce constant while adjusting the rotational approach angle of the motor output shaft by ±2°-5°. This fine-tuning can resolve minor thread misalignment issues and increases the success rate of thread engagement by changing the contact angle between the screwdriver bit and the thread. The second-level compensation changes the downforce to 80%-120% of the initial value. When the first-level compensation cannot solve the problem, adjusting the downforce can change the contact pressure between the screw and the threaded hole, further increasing the likelihood of thread engagement. The adjustment range of the downforce ensures sufficient contact force while avoiding thread damage caused by excessive pressure.

[0104] The three-level compensation switches to a pulsed rotation mode, using intermittent rotation to complete the locking process. This mode, through periodic rotation and pauses, effectively reduces thread jamming or stripping problems that may occur with continuous rotation, making it particularly suitable for severe misalignment or situations involving foreign objects. This three-level compensation strategy is designed to account for different degrees of thread misalignment, ranging from fine-tuning to significant adjustments, and then to changing the rotation mode, forming a progressive solution. Each level of compensation targets a specific misalignment situation, increasing the likelihood of successful locking by adjusting different parameters. In practical applications, this three-level compensation strategy can be automatically executed based on real-time feedback. For example, when thread misalignment is detected, the system first attempts level one compensation. If level one compensation is unsuccessful, it automatically enters level two compensation, and so on. This automated multi-level compensation mechanism greatly improves the system's adaptability and reliability.

[0105] As a preferred implementation, a priority and time threshold for executing a compensation strategy can be set. For example, Level 1 compensation can be set to an execution time of 2 seconds. If locking is not successful within 2 seconds, it automatically enters Level 2 compensation. Level 2 compensation can be set to an execution time of 3 seconds. If it is still unsuccessful, it enters Level 3 compensation. This setting ensures that the system can try different compensation strategies within a reasonable time, improving the locking success rate while avoiding excessively long single operation times. In a screw-locking operation, the system detected thread misalignment. First, Level 1 compensation is executed, adjusting the rotation angle of the motor output shaft to +3°. The system monitors for 1.5 seconds and finds that the thread is still not properly engaged. Subsequently, the system automatically enters Level 2 compensation, increasing the downward pressure to 110% of the initial value. The system monitors again for 2 seconds and finds that the thread engagement has improved but is still not completely successful. Finally, the system initiates Level 3 compensation, switching to a pulse rotation mode, pausing for 0.2 seconds every 0.5 seconds of rotation. After three such cycles, the thread successfully engages, completing the locking operation.

[0106] This multi-level compensation strategy has significant advantages over traditional single-level compensation methods. Traditional methods typically employ only a single compensation means, such as simple retraction and retry or increasing downward pressure, which often fails to effectively address complex thread misalignment situations. In contrast, the multi-level compensation strategy of this application, by gradually adjusting different parameters, can adapt to various misalignment conditions, greatly improving the locking success rate. For example, in some test scenarios, after adopting this multi-level compensation strategy, the locking failure rate caused by thread misalignment decreased from 30% to less than 5%, significantly improving the system's reliability and efficiency.

[0107] Furthermore, after completing the final locking at a speed 20%-30% lower than the initial motor speed in step S6, an anti-loosening detection step is also included:

[0108] a. After locking, maintain the preset holding torque for 10-15 seconds and monitor the amount of retraction of the motor shaft rotation angle;

[0109] b. When the retraction amount exceeds 0.5°, a secondary locking procedure is triggered until the retraction amount is less than 0.2°.

[0110] The anti-loosening detection steps proposed in this application aim to address the problem of loosening that may occur if screws are not fully tightened. By maintaining a certain torque after the tightening operation and monitoring the amount of retraction, potential loosening risks can be detected and addressed in a timely manner, thereby improving the reliability and stability of the tightening process.

[0111] Specifically, the anti-loosening detection steps of this application include the following key features: First, after the tightening operation is completed, the system maintains a preset holding torque for 10-15 seconds. This time period is chosen after careful consideration, allowing sufficient observation of the screw's stability without excessively prolonging the entire operation cycle. The holding torque can be adjusted according to specific application scenarios; for example, it can be set to 80%-90% of the final tightening torque.

[0112] During torque holding, the system continuously monitors the rotational angle backlash of the motor shaft. This backlash refers to the minute rotation angle of the screw under the action of a counterforce. Monitoring can be achieved using a high-precision angle sensor, such as a photoelectric encoder with a resolution of 0.01°. The system has a threshold criterion for judging the backlash. When the detected backlash exceeds 0.5°, a secondary tightening procedure is triggered. This threshold is set based on empirical data and actual testing, effectively capturing potential loosening risks while avoiding misjudgments due to oversensitivity. The secondary tightening procedure continues until the backlash decreases to below 0.2°, ensuring a high degree of stability in the screw connection. Secondary tightening can be performed gradually, such as increasing the torque slightly each time until the target backlash is reached.

[0113] In practical applications, this anti-loosening detection step can be seamlessly integrated with the aforementioned intelligent adaptive screw-locking control method. For example, after final tightening is completed in step S6, the anti-loosening detection stage begins immediately. The system can utilize existing motor control and sensor systems, requiring only the addition of corresponding software algorithms to achieve this function. By introducing this anti-loosening detection step, this application significantly improves the reliability of screw connections. Compared to traditional methods that rely solely on a single tightening operation, this solution can monitor and respond quickly to potential loosening risks in real time. This is particularly important for applications involving vibration environments, drastic temperature changes, or alternating loads. In specific implementations, relevant parameters can be adjusted according to different application requirements. For example, in a specific embodiment:

[0114] 1. Set the holding torque time to 12 seconds, and the holding torque value to 85% of the final locking torque.

[0115] 2. Use a photoelectric encoder with a resolution of 0.005° to monitor the rotation angle.

[0116] 3. When the retraction exceeds 0.5°, a secondary locking is triggered, using a step-by-step torque increase method, increasing by 2% each time, up to a maximum of 110% of the original torque.

[0117] 4. After each adjustment, wait 2 seconds to observe the amount of retraction until the amount of retraction drops below 0.15° or reaches the maximum allowable torque.

[0118] 5. If the requirements still cannot be met after multiple adjustments, the system will issue an alarm, prompting manual intervention and inspection.

[0119] This implementation method can effectively prevent screws from loosening in most industrial applications, while ensuring operational safety and efficiency.

[0120] Compared with existing technologies, the anti-loosening detection step of this application has significant advantages. Traditional methods typically only focus on the tightening process itself, neglecting potential risks after tightening. Even some systems employ fixed secondary tightening procedures, lacking the ability for real-time monitoring and intelligent adjustment. In contrast, this application provides a dynamic, adaptive anti-loosening mechanism that can adjust promptly according to actual conditions, greatly improving the long-term reliability of screw connections.

[0121] Furthermore, after recording the feature parameter set of this operation in step S7, including the maximum torque value, final stable torque, total time consumption, and abnormal event markers, the adaptive learning module performs machine learning analysis on the historical feature parameter set to dynamically optimize the torque threshold and motor speed curve for subsequent operations. By introducing the adaptive learning module, this application can dynamically adjust key parameters based on historical operation data, improving the system's adaptability and efficiency. Specifically, the adaptive learning module of this application first collects and stores the feature parameter set of each screw-locking operation, including the maximum torque value, final stable torque, total time consumption, and abnormal event markers. This data constitutes a rich historical database, reflecting the operational characteristics under different working conditions. Furthermore, the adaptive learning module uses machine learning algorithms to analyze this historical data. For example, algorithms such as Support Vector Machine (SVM) or Random Forest can be used to extract key features and patterns from historical data. Through learning from a large amount of data, the system can identify key factors affecting screw-locking quality and efficiency. Based on the machine learning analysis results, the system can dynamically optimize two key parameters: the torque threshold and the motor speed curve. Optimizing the torque threshold can improve locking accuracy and avoid thread damage or insufficient locking caused by excessive or insufficient torque. Optimizing the motor speed curve can maximize operational efficiency while ensuring locking quality. As a preferred implementation, the adaptive learning module can establish multiple parameter models based on different types of screws (such as material and size). When the system identifies a new screw type, it can quickly switch to the corresponding optimized model, achieving rapid adaptation. Furthermore, the adaptive learning module can also combine anomaly event tagging to optimize the system's anomaly handling strategy. For example, if historical data shows that a certain type of screw is prone to stripping, the system can automatically adjust the initial torque and speed to reduce the risk of stripping.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart adaptive screw-locking control method for a motor, characterized in that, Includes the following steps: Step S1: Obtain real-time images of the assembly area using a camera integrated at the front end of the motor, and establish a three-dimensional coordinate model of the screw hole pose based on a three-dimensional vision recognition algorithm. Step S2: The motor performs pose calibration according to the three-dimensional coordinate model to establish a vertical alignment between the screwdriver head and the target screw hole, and then starts the motor to rotate and press down. Step S3: When starting the motor rotation and pressing operation, simultaneously collect the real-time current value of the motor winding, the rotation angle of the output shaft, and the axial pressure data; Step S4: Establish a dynamic torque model, compare the real-time current value of the motor winding with the preset current-torque mapping relationship, and calculate the actual output torque by combining the axial pressure value fed back by the pressure sensor. Step S5: When the actual output torque is detected to exceed the preset torque threshold range, immediately reduce the motor speed to below the safety threshold and start the reverse rotation retraction program; Step S6: When the screwdriver head rotates to 90%-95% of the preset standard angle value, switch to precision torque control mode and complete the final tightening at a speed 20%-30% lower than the initial motor speed. Step S7: After completing the locking operation, record the characteristic parameter set of this operation, including the maximum torque value, the final stable torque, the total time, and the abnormal event marker.

2. The intelligent adaptive screw-locking control method for a motor according to claim 1, characterized in that, The three-dimensional visual recognition algorithm in step S1 includes: S1-1. Collect stereo vision images through a dual-camera module and generate depth point cloud data; S1-2. Extract screw hole contour feature points based on edge detection algorithm; S1-3. The iterative nearest point algorithm is used to register the real-time point cloud data with the preset standard hole position model; S1-3. Calculate the spatial angle deviation between the screw hole center axis and the motor axis.

3. The intelligent adaptive screw-locking control method for a motor according to claim 1, characterized in that, The motor rotation and downward pressing operation in step S2 includes: a. Set the initial rotational speed and axial pressure value of the motor output shaft; b. The motor rotates at the set initial rotational speed and simultaneously applies a preset axial pressure to the screwdriver head.

4. The intelligent adaptive screw-locking control method for a motor according to claim 1, characterized in that, The method for establishing the dynamic torque model in step S4 includes: S4-1. Establish a current-torque reference curve under no-load conditions of the motor; S4-2. Dynamically compensate the reference curve based on the real-time detected axial pressure value of the motor output shaft. For every 1N increase in pressure, the torque compensation coefficient increases by 0.5%-0.8%.

5. The intelligent adaptive screw-locking control method for a motor according to claim 1, characterized in that, The reverse rotation and retraction procedure in step S5 specifically includes: S5-1. Record the rotation angle position of the motor output shaft when the abnormal torque occurs; S5-2. Rotate the motor output shaft in the reverse direction at 50% of the normal speed to the position 2-3 revolutions before the abnormality occurred; S5-3, When re-attempting locking, use an initial torque threshold that is reduced by 20%.

6. The intelligent adaptive screw-locking control method for a motor according to claim 1, characterized in that, The abnormality diagnosis procedure in step S7 includes: S7-1. Establish a current fluctuation spectrum characteristic database, including typical patterns of stripped threads, thread damage, and foreign object jamming; S7-2. Perform wavelet transform on the real-time current signal to extract the energy distribution characteristics of 3-5 key frequency bands; S7-3. Use a pattern matching algorithm to compare the real-time features with the database to determine the anomaly type.

7. The intelligent adaptive screw-locking control method for a motor according to claim 1, characterized in that, If the standard angle value is not reached within the predetermined time window in step S6, an abnormality diagnosis program is triggered. The program analyzes the frequency characteristics of current fluctuations to determine whether stripping or thread misalignment has occurred. When thread misalignment is detected, a multi-level compensation strategy is automatically executed.

8. The intelligent adaptive screw-locking control method for a motor according to claim 7, characterized in that, The multi-level compensation strategy includes: Level 1 compensation: Keep the downforce constant and adjust the rotation angle of the motor output shaft by ±2°-5°; Secondary compensation: Change the downforce to 80%-120% of the initial value; Level 3 compensation: Switch to pulse rotation mode to lock the device by intermittent rotation.

9. The intelligent adaptive screw-locking control method for a motor according to claim 1, characterized in that, After completing the final locking at a speed 20%-30% lower than the initial motor speed in step S6, an anti-loosening detection step is also included: a. After locking, maintain the preset holding torque for 10-15 seconds and monitor the amount of retraction of the motor shaft rotation angle; b. When the retraction amount exceeds 0.5°, a secondary locking procedure is triggered until the retraction amount is less than 0.2°.

10. The intelligent adaptive screw-locking control method for a motor according to claim 1, characterized in that, Step S7 records the feature parameter set of this operation, including the maximum torque value, the final stable torque, the total time, and the abnormal event markers. Then, the historical feature parameter set is analyzed by machine learning through the adaptive learning module to dynamically optimize the torque threshold and motor speed curve for subsequent operations.